Edith Cowan University Research Online
Theses: Doctorates and Masters Theses
2000
Factors affecting the recruitment of riparian vegetation on the Ord and Blackwood Rivers in Western Australia
Neil Pettit Edith Cowan University
Follow this and additional works at: https://ro.ecu.edu.au/theses
Part of the Other Ecology and Evolutionary Biology Commons
Recommended Citation Pettit, N. (2000). Factors affecting the recruitment of riparian vegetation on the Ord and Blackwood Rivers in Western Australia. https://ro.ecu.edu.au/theses/1625
This Thesis is posted at Research Online. https://ro.ecu.edu.au/theses/1625 Edith Cowan University
Copyright Warning
You may print or download ONE copy of this document for the purpose of your own research or study.
The University does not authorize you to copy, communicate or otherwise make available electronically to any other person any copyright material contained on this site.
You are reminded of the following:
Copyright owners are entitled to take legal action against persons who infringe their copyright.
A reproduction of material that is protected by copyright may be a copyright infringement. Where the reproduction of such material is done without attribution of authorship, with false attribution of authorship or the authorship is treated in a derogatory manner, this may be a breach of the author’s moral rights contained in Part IX of the Copyright Act 1968 (Cth).
Courts have the power to impose a wide range of civil and criminal sanctions for infringement of copyright, infringement of moral rights and other offences under the Copyright Act 1968 (Cth). Higher penalties may apply, and higher damages may be awarded, for offences and infringements involving the conversion of material into digital or electronic form. USE OF THESIS
The Use of Thesis statement is not included in this version of the thesis. FACTORS AFFECTING THE RECRUITMENT OF RIPARIAN VEGETATION ON THE ORD AND BLACKWOOD RIVERS IN WESTERN AUSTRALIA.
Neil Pettit (Bachelor of Applied Science, Master of Philosophy) School of Natural Sciences Edith Cowan University 25 May 2000
This Thesis is presented for the degree of Doctor of Philosophy at Edith Cowan University. ABSTRACT
This thesis provides baseline information on the ecological processes involved in the recruitment and regeneration of riparian vegetation. As there has been a paucity of basic ecological studies on riparian vegetation in Australia, the project is broad in scope, and gives a general picture of the factors influencing the regeneration of riparian vegetation and provides a starting point for more detailed work. The project focuses on factors determining recruitment events and the life history traits of particular species on a river in the cool temperate zone of south western Australia (Blackwood River) and on a river in the dry tropics of the Kimberley region of north western Australia (Ord River). By studying two contrasting river systems in different climatic zones, the influence of the physical environment and biotic factors on vegetation persistence and recruitment in the riparian zone can be· distinguished. This will contribute to providing an ecological basis for the rehabilitation and management of riparian vegetation in these types of systems.
The structure of the vegetation on the Blackwood River consists of an overstorey dominated by Eucalyptus rudis, with a shrub understorey at ungrazed sites and with annual species dominant in areas grazed by livestock. On the Ord River there is a much more diverse overstorey and a species-poor understorey dominated by perennial grasses. Fencing to exclude stock, or to at least manage grazing, is a fundamental step towards achieving rehabilitation of degraded riparian sites where recruitment may be severely limited. Exclosure experiments on the Blackwood River show little improvement in recruitment after three years, with only minor increases in the occurrence and cover of native species. Establishment of these species may be difficult with the increase in abundance of exotic grasses and annual herbs which has occurred as a result of the absence of grazing. These results are however for the short-term and a much longer period is required to look at the vegetation dynamics and successional processes of these sites. In the riparian zone, regeneration of the vegetation from soil seedbanks is important for annual species of herbs and grasses but of only minor significance for perennial species. For perennial species, particularly the overstorey, direct seedfall from existing vegetation occurs, and enhanced dispersal by floating downstream with flood debris is a consequential recruitment mechanism. Hence, reproductive phenology of the four species monitored in this study appears to be well adapted to the hydrological regime on the respective rivers. Position in the riparian landscape where seedlings manage to establish is strongly related to environmental conditions that provide adequate moisture but protect seedlings from flooding.
Historical flow records can be used to develop an understanding of the natural flow regime for a particular river which··can then be related to patterns of vegetation development in terms of i i
reproductive phenology, seedling establishment and population structure, as well as plant community patterns in the riparian zone. Variability in natural flow regimes, as a disturbance, can therefore be used in conjunction with other abiotic and biotic factors in developing a model of vegetation dynamics for the riparian zone. For example, the regime of intermittent high frequency large flood disturbances on the Ord River prevents the establishment of stable states of the vegetation and the ecosystem is characterised by long periods of transition between short lived stable states. This riparian ecosystem is thus driven by physical (allogenic) processes rather than by vegetation successional (autogenic) processes. In contrast, lower energy seasonal flooding on the Blackwood River allows mature stands of trees to develop throughout the river profile. Recruitment is continual, although species can also respond to large flood events. This disturbance regime results in long periods of stable states with short periods of transition. The vegetation is thus subject to longer periods of autogenic processes and, because of lower frequency flooding disturbance, shorter periods of allogenic processes. These results highlight the effect of the different fluvial regimes of the two rivers on the respective vegetation dynamics. Management of the riparian vegetation should therefore take into account the frequency and rate of change in the vegetation and that disturbed states and long periods of transition between states, particularly on the Ord River, are part of the natural process. This would suggest that altering the natural flow regimes, such as through river regulation, would have significant effects on riparian vegetation dynamics.
This work has relevance to all aspects of riparian zone vegetation, including management of natural systems unaffected by man-made disturbances, for systems affected by stock grazing, for areas requiring rehabilitation and on regulated rivers. It highlights the importance of fluvial processes to riparian vegetation and indicates that understanding the natural flow regime of a target river is a critical first step in the management of riparian vegetation and in the planning of riparian zone rehabilitation. Where the riparian zone is highly modified, through, for example, livestock grazing and/or weed invasion, natural regeneration of the riparian vegetation may be a long term process. If intervention, such as replanting, is appropriate, care should be taken that species selected are adapted to particular site conditions, such as flooding regime, landscape position and river geomorphology. iii
Declaration I certify that this thesis does not, to the best of my knowledge and belief: 1. Incorporate without acknowledgment any material previously submitted for a degree or diploma in any institution of higher education. 2. Contain any material previously published or written by another person except where due reference is made in the text; or 3. Contain any defamatory material.
Neil Pettit 25 May 2000 iv
ACKNOWLEDGMENTS
I wish to thank Dr Ray Froend for supervising this project and for his ideas and sagacity. Ray has been the inspiration to undertake this academic endeavour and it has been my great luck to be associated with him. Dr Peter Davies (University of Western Australia) as my eo-supervisor and for being my link with LWRRDC and providing support and advice and valuable comments on a draft of this thesis. Dr Pierre Horwitz for his encouragement and providing a home at ECU. Dr Luke Pen (Water and Rivers Coinmission) who was always positive about this work and provided valuable comments on a thesis draft. Professor Stuart Bunn (Griffith University) for providing the opportunity to work in such an interesting area as riparian ecology. I would also like to thank Robert Doupe, an excellent guide and companion in the field, for passing on his knowledge and enthusiasm of the Kimberley region and for many interesting and lively philosophical discussions around the campfire. Mark Angeloni for his invaluable assistance in the field. Jim Roe and Richard Finke for their generosity, good hmnour ;;md friendliness in providing accommodation and the loan of equipment while in Kununurra. Alice Karafalis (Blackwood Catchment Coordinating Group) for her continued support throughout this project. Gary Ogden for his assistance in the field and many fruitful discussions and for putting up with my surliness in the office. Dr Michael Douglas (Northern Territory University) for providing information on the effects of fire on riparian vegetation from his PhD thesis. Jeff Kite (Water and Rivers Commission of Western Australia) for providing financial assistance for this project. \ Water and Rivers Commission staff in Kununurra for use of a vehicle and equipment in the field and WRC in Perth for providing hydrological data. The scholarship for this work provided by
the Land and Water R~sources Research and Development Corporation (LWRRDC) through the National Riparian Program is also gratefully acknowledged.
To Anne Pettit the best technical editor there is. For her patience, forbearance and support and for taking on the bulk of the responsibilities of our family for the last four years. V
TABLE OF CONTENTS
ABSTRACT ...... : ...... I
ACKNOWLEDGMENTS ...... IV
TABLE OF CONTENTS ...... V
CHAPTER 1 ...... 1
INTRODUCTION ...... 1 Definition ...... 1 Importance of Riparian Vegetation ...... ,.', ...... 3 Impacts on Riparian Vegetation ...... 6 Significance of Study ...... 9 Thesis Outline ...... : ...... l2
CHAPTER 2 ...... 14
THE RIPARIAN ZONE AND VEGETATION DYNAMICS: A REVIEW...... 14 River Processes ...... 14 Plant Ecology ...... 18 Vegetation dynamics ...... IS Restoration ecology ...... 22 Plant life-history traits ...... 23 Disturbance ...... 25 Seed and Seedling Ecology ...... 27 Vegetative regeneration ...... 31 Regeneration of riparian zones ...... 33 Conclusions ...... 37
CHAPTER 3 ...... 39
STUDY SITE DESCRIPTIONS ...... 3 9 Introduction ...... 39 Location ...... 40 Ord River ...... 40 Blackwood River...... 43 Climate ...... 44 Ord River ...... 44 Blackwood River ...... 45 Comparison of climates ...... 46 Hydrology ...... so Geology ...... , ...... 54 OrdRiver ...... , ... : ...... 54 Blackwood River...... 55 Geomorphology and Soils ...... 56 Ord River ...... 56 Blackwood River...... 57 Vegetation ...... , ...... 59 Ord River ...... 59 Blackwood River, ...... 60 Land Use .: ...... :.: ...... 61 vi
Ord River ...... 61 Blackwood River ...... 63
CHAPTER 4 ...... " ...... 64
DESCRIPTION OF RIPARIAN VEGETATION WITH REGARD TO FLORISTICS AND LIFE FORM ...... 6 4 Introduction ...... 64 Methods ...... 66 Results ...... 69 Floristic structure ...... 69 Floristic patterns ...... 74 Life form groups ...... 76 Discussion ...... 81 Floristic structure ...... 81 Life forms ...... 81 Floristic patterns ...... 83
CHAPTER 5 ...... : ...... 86
IDENTIFYING THE NATURAL FLOW REGIME AND THE RELATIONSHIP WITH RIPARIAN VEGETATION ...... •...... 86 Introduction ...... 86 Methods ...... 89 Natural flow regime ...... 89 Riparian flows ...... 90 Vegetation plots ...... 91 Flow regime and vegetation relationships ...... 92 Results ...... 93 Natural flow regime ...... : ...... 93 Riparian zone flows ...... 97 Vegetation and flooding regime ...... 99 Discussion ...... 103 Natural flow regime ...... l04 Discharge and vegetation relationship ...... 105
CHAPTER 6 ...... 108
EFFECTS OF LIVESTOCK GRAZING AND FIRE DISTURBANCE ON RIPARIAN VEGETATION...... •...... 108 Introduction ...... 108 Methods ...... 111 Results ...... l13 Compruison of grazed and ungrazed riparian areas ...... 113 Grazing exclusion from grazed riparian land ...... 118 Effects of fire on riparian vegetation ...... 124 Discussion ...... 127 Grazing effects on dparian vegetation ...... 127 Effects of relief from grazing ...... 131 Fire ...... 133 vii
CHAPTER 7 ...... 136
SOURCES OF SEED FOR RECRUITMENT OF RIPARIAN COMMUNITIES ...... 136 Introduction ...... : ...... 136 Methods ...... 139 Soil seedbank ...... 139 Propagule dispersal mode ...... 140 Flood debris samples ...... 140 Seed floating ...... 141 Reproductive phenology of target species ...... 141 Seed longevity ...... 142 Seed predation ...... 143 Resuits ...... 143 Soil seedbank ...... 143 Dispersal mechanisms ...... 149 Flood debris samples ...... 151 Seed floating ...... 153 Reproductive phenology ...... 154 Fate of seed ...... : ...... 159 Discussion ...... 159 Riparian soil seedbank ...... 159 Dispersal mechanisms ...... 161 Dispersal by water in debris and seed floating ...... 163 Reproductive phenology ...... 163
8 ...... 166
SEEDLING RECRUITMENT OF SELECTED RIPARIAN SPECIES ...... 166 Introduction ...... ,, ...... 166 Methods ...... 168 Seed germination ...... 168 Seedling establishment...... 170 Effects of flooding ...... 172 Results ...... : ...... 173 Seed germination ...... 173 Seedling establishment...... 178 Effects of flooding ...... 186 Discussion ...... 192 Seed germination ...... 192 Seedling establishment...... 194 Effects offlooding ...... 196
CHAPTER 9 ...... 199
POPULATION STRUCTURE AND SPATIAL DISTRIBUTION OF SELECTED SPECIES IN THE RIPARIAN ZONE ...... 19 9 Introduction ...... 199 Methods ...... 201 Results ...... 204 Tree population structure ...... 204 Size class distribution ...... 209 Long term hydrograph ...... 225 Tree age and hydrology ...... 226 Discussion ...... 226 Population structure ...... 231 viii
Tree age ...... 233 Tree vegetation dynamics ...... 234
CHAPTER 10 ...... 237
SUMMARY AND GENERAL DISCUSSION ...... 237 Riparian vegetation cornrnunities ...... 239 The fluvial regime and the influence on riparian vegetation ...... 241 Livestock grazing and fire ...... 242 Sources of seed for regeneration ...... 244 Mechanisms of seedling recruitment...... 246 Vegetation dynamics and population structure in the riparian zone ...... 248 Management implications ...... 251
REFERENCES ...... 254
APPENDIX 1 ...... 282
List of species found at the study sites on the Blackwood River...... 282
APPENDIX 2 ...... 290
List of species found at the study sites on the Ord River...... 2 9 0
APPENDIX 3 ...... 295
List of submitted publications and conference presentations related to the thesis topic ...... 295
APPENDIX 4 ...... 297
Copy of Manuscript accepted for publication in Wetland Ecology and Management ..... 297 1
CHAPTER 1
INTRODUCTION
Understanding the ecological processes in river systems is acknowledged as fundamental to sound management of river systems in Australia (Lovett & Price 1999; Pen 1999) and elsewhere
(Decamps 1993; Malanson 1993; Naiman & Bilby 1998). This thesis provides baseline information on the ecological processes involved in the recruitment and regeneration of riparian vegetation. As there has been a paucity of basic ecological studies on riparian vegetation in
Australia, the project is broad in scope, giving a general picture of the factors influencing the regeneration of riparian vegetation and providing a starting point for more detailed work. The project focuses on factors determining recruitment events and the life history traits of particular species on a river in the cool temperate zone of south-western Australia (Blackwood River) and on a river in the dry tropics of the Kimberley region of north-western Australia (Ord River). By studying two contrasting river systems in different climatic zones, the influence of the physical environment and biotic factors on vegetation persistence and recruitment in the riparian zone can be distinguished. This will contribute to providing an ecological basis for the rehabilitation and management of riparian vegetation in these types of systems.
Definition
Riparius is the Latin word meaning "of or belonging to the bank of the river" and therefore
"riparian" refers to the biotic communities living along streams (Naiman et al. 1998). The riparian zone can be thought of as the interface between terrestrial and aquatic systems and often contains sharp gradients in environmental and community processes (Naiman et al. 1998).
Riparian environments are recognised as being highly productive and a key component of the landscape (Bunn 1993; Malanson 1993). However, what constitutes the riparian zone is not always clear. There are several definitions that can be applied, the efficacy of each depending on the purposes for which it is used. The legislative definition traditionally used in many areas 2
of Australia usually defines the riparian zone as a strip 20 to 40 metres wide along the banks of a
river or stream. In natural systems, this is not very useful as river ecosystems rarely fit neatly
within these boundaries. A landform definition includes the area between low-flow level of the
water course and the highest point of transition between the channel and the floodplain.
However, in many rivers the floodplain is an integral part of the riparian zone. A definition
based on differences in the vegetation is clearly not adequate because of the many varied plant
communities that can make up the riparian zone. A definition based on riparian functions would
therefore be the most useful and would describe that part of the landscape which exerts a direct
influence on stream channels and on the aquatic ecosystems within them and which in turn is
affected by the geomorphology of the stream channel and flow.
In the field the transition to the riparian zone is often easily recognised by changes in the
vegetation as it becomes more dense and lush and has a different suite of species in both the
overstorey and understorey. The distribution of plant communities is mainly influenced by
tolerance to waterlogging, which is itself determined by micro-topography and soil type,
particularly sediment build-up from flooding (Malanson 1993). In a generalised riparian zone,
along a transect perpendicular to the stream, the vegetation may consist of several distinct zones.
The littoral zone is the area between the summer low and winter high water marks which
consists of aquatic floating, submerged and emergent macrophytes. Immediately above this are
the stream or levee banks which are formed after the deposition of sand during flooding. These
areas are usually well-drained and contain species which are not highly tolerant of
waterlogging. Away from the levee banks is the floodplain, the width of which may vary
greatly from virtually absent in deeply incised stretches of rivers to tens of kilometres wide on
low relief plains. Plant communities on the floodplain vary and usually exhibit some tolerance
to periodic inundation. Within the floodplain old abandoned river channels may periodically
fill with water or contain permanent water. These areas may also have distinct vegetation communities which resemble lentic wetlands. In deeply incised stretches of the river, riparian vegetation can be restricted to the stream bank with dryland vegetation coming right up to the 3
top of the bank. Such areas, in view of the previous functional definition, should be regarded as
part of the riparian zone.
The core of riparian ecosystems is the zone between unflooded (upland) and continuously
submerged (aquatic) areas (Nilsson et al. 1993). The upland boundary often coincides with the
lower limit of particular upland species which cannot tolerate even short periods of flooding.
Hydrologic, geomorphic and ecological processes interact to form the riparian landscape
(Malanson 1993) and the riparian zone can be described as the three dimensional zone of direct
interaction between terrestrial and aquatic ecosystems (Gregory et al. 1991). Dimensions of the
riparian zone of influence for specific ecological processes are determined by unique spatial
-patterns and temporal dynamics. In this study, the area treated as the riparian zone for both
study rivers generally encompassed from the low water mark at summer base flow levels to the
beginning of the floodplain, where there is usually a change of slope and the vegetation is
affected by flooding only very occasionally.
Importance of Riparian Vegetation
The vegetation within the riparian zone performs an important ecological function for in-stream
processes (Bunn 1993). Trees on the river edge provide shade which can reduce primary
productivity through the reduction of light and also reduce water temperature (Bayly &
Williams 1981; Bunn 1993). This effect is implicit in the river continuum concept whereby,
largely because of shading, heterotrophic processes pre-dominate and macroinvertebrate fauna
are dominated by shredder, collector and to a lesser extent predator species (Vannote et al. 1980;
Williams 1983). The process of limiting primary productivity by shading is particularly
important in maintaining the health of rivers which drain predominantly agricultural land,
where there can be high inputs of nutrients (particularly N and P) into the river from surface and
groundwater flows (Peterjohn & Correll1984; Chauvet & Decamps 1989). Streamside vegetation
also contributes carbon to the water through leaf litter and other woody material (Davies 1994).
However, in Australia,. the sclerophyllous leaves of most streamside trees break down very 4
slowly and are shed in small numbers throughout the year1 thus limiting nutrient input by this
means. The leaf litter can also be high in tannins which stain the water1 reducing light penetration and thus limiting photosynthesis (Campbell 1?93).
Riparian vegetation provides a diversity of habitats for in-stream fatma that are important in maintaining biodiversity. This can be in the form of leaf litter and fallen branches as well as whole trees which can create 'riffle' zones/ an important habitat for stream macroinvertebrates
(Malanson 1993). Other riparian vegetation/ such as aquatic submerged/ floating and emergent macrophytes can provide a wide variety of habitats (Bunn 1993; Bunn et al. 1999). Large woody debris plays a significant ecological role in the colonisation and establishment of many plant species on the flood plain and in the river channel in areas of sediment deposition (Naiman et al.
1998; Pen 1999 1 Lovett & Price 1999). It also has an important geomorphic role in the development of channel morphology (Rutherfurd et al. 1999).
Off-stream riparian vegetation can provide many different habitats for terrestrial fauna. It contributes to higher species diversity and abundance of flora in the riparian zone as it forms an ecotone between the different plant communities (Wissmar & Swanson 1990; Nilsson et al. 1993;
Bretschko 1995). The riparian zone may not strictly fit the definition of an ecotone however (see
Risser 1993) and may be better described as a transition zone between aquatic and terrestrial habitats.
Many animals prefer the moist1 productive riparian habitat to adjacent upland areas (Leach &
Edge 1994). The high levels of natural disturbance and the great diversity of habitat formed by these corridors of natural vegetation makes them very important for regional biodiversity
(Naiman et al. 1993). The importance of the riparian wne for wildlife habitat becomes particularly significant in fragmented landscapes in a matrix of cleared agricultural land.
Riparian corridors have been considered useful for connecting isolated remnant patches of native vegetation (Watson 1991) as well as in facilitating the movement of avian fauna/ affecting regional biodiversity (Knopf & Samson 1994). Species diversity and abundance of avian and 5
mammalian fauna has been shown to be greater in riparian corridors than in the surrounding
landscape (Decamps et al. 1987; Crome et al. 1994). Insectivorous birds, sheltering and nesting in
riparian vegetation, are important in controlling populations of herbivorous insects which are
potentially harmful to agriculture and native plants (Loyn 1987). Riparian systems can be key
elements in the landscape as they are usually interspersed throughout the landscape, are high in
species richness and are corridors for the flow of nutrients, energy and species (Nilsson & Jansson
1995).
Riparian vegetation is also important in buffering the flow of nutrients and sediment from the
catchment into the river system. Sediment and nitrate concentrations in a river in the United
States were reduced by up to 90% in water flowing through riparian areas, with less effective
removal of phosphorus (Gilliam 1994). Riparian vegetation is also considered a good short-term
filter for removing N, P, Ca and Mg flowing from upland areas, thus protecting water quality
(Lowrance et al. 1984). The wet soils of the riparian zone, together with the vegetation, can
have an effective nitrate buffering capacity with a 30 metre wide buffer reported as being
sufficient to remove all nitrates seeping through in the groundwater (Pinay & Decamps 1988;
Pinay et al. 1993). Streamside vegetation is also an important buffer, reducing agricultural runoff
of Nand Pinto the river on the south coast of Western Australia (Weaver & Reed 1998).
Rivers are naturally high energy systems, especially in periods of high flow such as during floods, and consequently some erosion is a natural occurrence. However, without riparian vegetation high levels of bank erosion can occur. This can have in-stream and off-stream impacts such as increased turbidity from increased sediment loads and greater flooding of surrounding land due to destruction of stream levee banks which help to contain the water within the stream channel (Thorne 1990). Riparian vegetation can limit channel erosion and down-cutting by reducing the tractive force of flowing water, protecting the bank from direct impacts and by inducing deposition of sediments (Malanson 1993). Accumulation of sediment promotes the establishment of vegetation which in turn tends to reduce near-bank velocities and tractive forces, reducing erosion and increasing deposition of sediment (Thorne 1990). The accumulation of 6
woody debris also contributes to the deposition of sediment. These sites of deposition are also
important for the recruitment of riparian vegetation (Nanson & Beach 1977). Bank stability is
maintained by the roots of perennial vegetation which provide a stable base for binding the soil
and holding the bank in place (Pen 1999). Large roots of trees anchor the river bank and the roots
and rhizomes of perennial understorey vegetation, such as shrubs and sedges, hold the surface
soil in place between these larger roots. Subsidence of the bank can occur if perennial vegetation
is missing, particularly if the soil is wet, and the bank can collapse under its own weight (Thorne
1990). Vegetation is particularly important in reducing damaging effects when the river is in
flood, by reducing the velocity and spread of floodwaters and, through evapotranspiration, reducing bank wettness (Pen 1999).
Many riparian habitats are subject to one or more forms of external disturbance such as land clearing, livestock grazing and trampling and increased fire frequency. In most cases, these disturbances interfere with the natural processes of plant recruitment and regeneration of riparian vegetation, eventually resulting in the disappearance of these communities as mature trees die. This can threaten the long-term sustainability of the whole river and riparian ecosystem.
Impacts on Riparian Vegetation
In most countries, changes to land use in river catchments has had a major detrimental impact m riparian zones. Activities include industrial and urban pollution (Jeffries & Mills 1990) and agriculture and forestry (Abido 1985; Shah & Thames 1985; Fleischner 1994). In Australia, riparian zones are regarded as the most degraded natural resource zones (CEP A 1992).
Clearing of land adjacent to the riparian zone can result in rising watertables causing stress and even death of the vegetation through permanent waterlogging and/ or salinity (Peck 1978; Busch
& Smith 1993). Groundwater under cleared agricultural land can be high in nutrients, particularly nitrates, which can flow into the riparian zone and the river itself, causing 7
eutrophication (Pinay & Decamps 1988). Surface water flows from agricultural land can also cause problems by carrying nutrients into the riparian zone, particularly where this land has been fertilised for crops or carries livestock (Lowrance et al. ·1984; Weaver & Reed 1998). High levels of nitrates in the water may also result in acidification of waterways, impacting on in- stream macroinvertebrates and other fauna (Jeffries & Mills 1990).
High nutrient levels combined with high rates of disturbance and the large edge to area ratio of riparian zones make them particularly vulnerable to invasion by exotic plants. mwoodlands in the wheatbelt of Western Australia these factors, together with livestock grazing, have led to invasion by exotic species and the disappearance of natives (Scougall et al. 1993). m the riparian zone in Australia the replacement of native plants, usually by soft-leaved exotics which break down quickly adding to nutrient inputs, can drastically alter in-stream ecology
(Pidgeon & Cairns 1981; Pen 1999).
Some of the most serious degradation of the riparian zone is caused by livestock. Livestock, usually cattle or sheep, can have direct impacts such as destruction and trampling of vegetation.
They can cause erosion (by trampling streambanks), compaction of the soil, an increase in nutrient levels (both in-stream and in the soil) as well as increased stream turbidity (Askey-Doran & . Pettit 1999). In many rivers in the agricultural and pastoral zones of Australia, livestock have free access to the riparian zone for summer watering and grazing. m the Kimberley region of north-western Australia, cattle overgrazing of the native vegetation has caused major erosion and siltation problems in many rivers (Williams et al. 1996). In addition to the loss of native species and reduction in cover, increases in nutrient levels and disturbance of the soil surface, livestock can also promote invasion by weeds (usually annual ruderal species) and bring about changes in vegetation structure (Fleischner 1994). m many semi-arid areas, livestock tend to congregate in the riparian zone as this area is the most productive and may be the only source of water and shade. Livestock affect stream morphology, stream-side vegetation, shape and quality of the water column and the structure of stream bank soil and have been implicated in the reduction of trout numbers in many U.S. rivers (Fleischner 1994). Grazing by livestock severely 8
restricts recruitment of most riparian plants, particularly that of overstorey species, mainly by
grazing of new seedlings. Livestock can also cause changes to the soil (e.g. compaction and
erosion) and promote the introduction of competing weeds (Kauffman et al. 1983). This leads, in
the long term, to profound changes in the structure of riparian vegetation and to consequent
impacts on stream ecosystems.
Fencing to exclude stock, or to at least manage grazing, is a fundamental step towards achieving
rehabilitation of degraded riparian sites. Whether grazing needs to be excluded completely
will depend on the type of community and the extent of existing degradation. In environments
that have had a long history of grazing, and where the vegetation has adapted to this form of
disturbance, the exclusion of livestock may result in deleterious changes to the vegetation
structure such as invasion of woody plants and reduction in species diversity (Milchunas &
Lauenroth 1993). Experiments on grazing exclusion in riparian vegetation have shown a reduction
in species richness and an increase in plant cover (Kauffman et al. 1983). These studies advocate
management of grazing in riparian zones so that grazing is excluded for some period of the year,
or in particular years, to allow. the vegetation to recover and recruitment to take place. In the
winter rainfall areas of Australia, exclusion of grazing for the summer period only, when most
damage to the vegetation and the stream bank occurs, may be enough to prevent further
degradation (L. Pen, Water and Rivers Commission, pers comm). However, successful recruitment
of many plant species may be episodic and rely on the coincidence of several factors such as
winter flooding, early receding of floodwaters corresponding with seedfall and summer rainfall.
Recruitment opportunities limited by the necessity for the accordance of particular environmental conditions, such as disturbance and climate, have been documented in other plant
communities (Wellington & Noble 1985a; Enright & Lamont 1989) and allowing intermittent grazing may interfere with the 'window of opportunity' for recruitment. Predicting which particular species are most affected by livestock grazing and which species are likely to return after stock exclusion is important in planning for the rehabilitation of degraded riparian areas.
This may depend on particular traits of individual species such as life form, ability to resprout after defoliation, seed production, seed dispersal strategies, seed dormancy and the ability to 9
form a seedbank. Abiotic factors are also important, such as microtopography and 'safe sites',
waterlogging, seasonal drought, soil types and climatic factors. In a study on remnant woodlands,
life form and reproductive strategy were both important in determining species persistence under
livestock grazing disturbance (Pettit et al. 1995).
Altered fire regime is another impact within riparian zones that may have detrimental effects
on regeneration. In areas of native vegetation wildfires would probably have occurred historically from time to time and, as for upland vegetation, many plants in the riparian zone may require a disturbance such as fire to provide gaps for recruitment (Kellman & Tackaberry
1993) and to stimulate seed release (Lamont et al. 1991) and germination (Shea et al. 1979). In managed areas, prescribed fires for fuel reduction or to stimulate new growth of edible species may be too frequent, resulting in the loss of many native species and an increase in weeds (Busch
& Smith 1993), or fire may be too rare to stimulate adequate recruitment of native riparian species.
At least 70% of the worlds freshwater flows are regulated (Petts 1984). In regulated rivers, altered flow regimes and flooding patterns can restrict regeneration opportunities and cause changes to species composition and structure in the riparian zone (Stromberg 1993; Decamps et al.
1995; Walker et al. 1995). On the Murray River in eastern Australia changes in degree, frequency and season of flooding has resulted in the lack of recruitment of Eucalyptus camaldulensis
(Dexter 1978; Bren & Gibbs 1986). Dams can act as dispersal barriers and cause fragmentation of continuous river populations into small isolated areas. Reduced water-level fluctuations in regulated rivers can also reduce the area of the riparian zone (Nilsson et al. 1997).
Significance of Study
To preserve the riparian landscape, consideration should be given to conservation of the whole catchment, including the maintenance of the hydrological regime as well as managing energy, materials and species flows (Malanson 1993). Restoration usually has the objective of restoring a l
10 system to a previous state that is self-sustaining through natural processes (Werner 1990). This goal is difficult to achieve both practically and because the nature of the pre-degraded state may not be known or current conditions may not support a r~tui:n to that state. The maintenance of an altered community that fulfils appropriate ecological functions within the riparian ecosystem, such as acting as a nutrient trap, is more achievable (Hobbs & Norton 1996).
Rehabilitation is thus directed at developing a vegetation community that is reasonably stable and allows natural recruitment, and the cessation of processes that contributed to the initial degradation.
In many rivers in the south-west of Western Australia, such as the Blackwood River, rising salinity can affect re-establishment of plants. Careful selection of salt tolerant species and various site amelioration methods are required to improve establishment (Pettit & Froend 1992).
For rehabilitation of these saline sites attention should also be paid to upland recharge areas which are the source of rising saline groundwater in the valley bottom and the riparian zone
(Peck 1978).
Ecology-based studies on the regeneration of plant communities have been carried out for many types of plant communities including grasslands (e.g. Grime & Hillier 1992; Mcintyre et al. 1995), . heath (e.g. Bell et al. 1982) and forest (e.g. Bell & Koch 1980) as well as lentic wetlands (e.g. van der Valk 1981; Boutin & Keddy 1993; Hills et al. 1994). However, little work has been done specifically on riparian vegetation (e.g. Nilsson et al. 1994) and especially not in Australia (e.g.
Dexter 1978; Bren & Gibbs 1986; Bren 1992; Roberts 1993). In many areas of Australia, such as in the Blackwood River catchment in south-west Western Australia, there is a high level of community interest in restoration of rivers. However, there is a lack of ecologically-based knowledge of the processes of regeneration within the riparian vegetation. In order to manage these systems knowledge of those ecological processes that are important for regeneration of the riparian vegetation is required. This thesis will contribute to this understanding by providing specific knowledge of the riparian vegetation of two rivers in Western Australia and by 11 developing some general principles on riparian vegetation recruitment that can be more generally applied.
An overall aim of this project is to provide baseline information on the processes involved in the recruitment and regeneration of riparian vegetation. Large scale disturbances, such as high water flows and flooding, are likely to be important factors in the dynamics of riparian communities. Therefore a general hypothesis for this thesis is that the fluvial regime has a strong influence on the vegetation dynamics of the riparian vegetation on these rivers. As a means to understanding the extent of this influence, the ecology of the riparian vegetation on two rivers with contrasting fluvial regimes are compared. In addition, congeneriC species occurring m each river, selected for detailed study, are functionally similar and therefore can help clarify the adaptations of each to the particular fluvial regime.
Specific aims are to: e Describe the riparian vegetation community on the two rivers and the environmental factors
that affect their distribution in the riparian zone as well as the relative importance of
different life form groups.
11 Characterise the river fluvial regime and channel morphology of the two rivers and the
relationship of these with the riparian vegetation.
• Describe the effects of two significant disturbances on the riparian vegetation, livestock
grazing and fire, and the response of the vegetation to relief from grazing.
• Investigate the sources of seed available for the regeneration of the riparian vegetation,
including timing of seed production and seed release, dispersal of seed and the existence of a
soil seedbank.
11 Examine the mechanisms of seedling recruitment in the riparian zone, including the
temperature requirements for germination, survival and establishment of seedlings in the
field and the adaptation of seedlings of riparian species to flooding. 12
• Develop a model of vegetation dynamics for each river based on examination of the
population structure and spatial distribution of selected species in the riparian zone and
their relationship to river hydrology.
• Outline management implications based on the results described in this thesis.
The thesis aims to establish the link between the fluvial regime and vegetation dynamics m these rivers. This will contribute to providing an ecological basis for the rehabilitation and management of riparian vegetation in these types of systems.
Thesis Outline
The thesis is structured to consider at community level riparian vegetation on both rivers and from this, particular species (Eucalyptus rudis and Melaleuca rhaphiophylla on the Blackwood
River and E. camaldulensis and M. leucadendra on the Ord River) have been selected for more detailed investigation of population structure and ecology and the influence of environmental factors on recruitment. Chapter 2 provides a background for regeneration processes of riparian vegetation considered in later chapters. It explores the linkages between hydrologic, geomorphic and ecological processes in the riparian zone. This includes a discussion of river processes, particularly in an Australian context; related aspects of vegetation theory; vegetation regeneration, including restoration ecology and plant life history traits; and seed and seedling establishment, as well as vegetative regeneration. Previous studies and current knowledge of riparian vegetation regeneration are also summarised. A detailed description of the two study rivers and their environments is given in Chapter 3, including sections on climate, hydrology, geology, soils, vegetation and land use. The next six chapters are structured to address the six specific aims outlined above. Chapter 4 provides a general description of the vegetation in the riparian zone at the study sites on the Blackwood and Ord Rivers. This includes examination of patterns of species richness and diversity of the plant communities in the riparian zone as well as exploring the key functional groups in terms of life form and species origin. Chapter 5 details the relationship between historical river flows and the riparian vegetation. Historical flow records can be used to develop a picture of the variability of the natural flow regime which can then be 13
related to patterns of vegetation development and plant commtmity zonation in the riparian zone. Chapter 6 is an assessment of the effects of livestock grazing on riparian vegetation and of the initial response of commtmities to the exclusion of live$tock after long term grazing. This includes examining the effects livestock grazing has had on the floristics of sites on the
Blackwood and Ord rivers, particularly the effect on species richness and diversity, structural changes in terms of life form and seedling establishment and continued recruitment of particular species. The effects of fire on riparian vegetation will also be examined in this chapter.
Chapter 7 investigates the potential sources of seed that are available for recruitment in riparian communities, through considering the diversity and abundance of the soil seedbank and whether this changes with topographic position and distance from the river, the fate and longevity of seed of specific species (and therefore their likelihood of entering the soil seedbank). The spectrum of propagule dispersal types and the role of flood debris as a potential source and means of dispersal of plant propagules in the riparian zone will also be investigated.
In addition, this chapter documents the reproductive phenology of the two target species on each river. Chapter 8 examines the influence of temperature on seed germination and the survival and growth of seedlings of the four target species in the particular riparian environments of the Ord and Blackwood Rivers. This includes determining the optimum temperature requirements for germination of seed of these species and the conditions likely to allow successful early . establishment of seedlings. The adaptation of the target species to various degrees of flooding is also determined. In Chapter 9 the population structure ofthe target species is examined. This is done by recording the size class structure of these species on different river reaches and their spatial variation within the riparian zone. This can be used to assess the population stability and successional trends or vegetation dynamics of these species (Veblen 1986). The final chapter
(Chapter 10) brings the various elements of this study to a synthesis and discusses various aspects of recruitment patterns and vegetation dynamics in riparian vegetation on the Blackwood and
Ord rivers. Management implications of this work will also be discussed in this chapter. 14
CHAPTER2
THE RIPARIAN ZONE AND VEGETATION DYNAMICS: A REVIEW
This chapter provides a background to the regeneration processes of riparian vegetation
considered in later chapters. It explores the linkages between hydrologic, geomorphic and
ecological processes in the riparian zone. This includes a discussion of river processes,
particularly in an Australian context; related aspects of vegetation dynamics; vegetation
regeneration (including restoration ecology and plant life history traits); seed and seedling
establishment; and vegetative regeneration. Previous studies and current knowledge of riparian
vegetation regeneration are also summarised. The concluding section discusses how these relate
to the specific questions addressed in this thesis.
River Processes
Riparian zones are shaped by spatial and temporal patterns of hydrology and geomorphology as well as by terrestrial .and aquatic ecosystems (Gregory et al. 1991). The geomorphic structure of the riparian zone results from the interaction of landform geology, climate and hydrology as well as from inputs of materials from adjacent hill slopes and vegetation. Geomorphic effects therefore create a mosaic of stream channels and flood plains within the valley floor (Gregory et al. 1991), the functions and relationships of which are summarised in Figure 2.1. These processes modify the riparian zone and exert their influence over various tim.e scales, ranging from regular events to much less frequent episodic events. Different types and scale of physical disturbance, along with landforms and hydrology, can determine the structure and dynamics within the riparian zone (Wissmar & Swanson 1990). Spatial scales can also vary widely from local deposition or erosion to local channel shifts or whole catchment flooding. Hierarchical classification of spatial scales for river channels would include geomorphic province, watershed and valley segment at higher scales, down to channel reach and pool and riffle zones at smaller 15
scales (Montgomery & Buffington 1998). Within the river zone, scale-dependent relationships
between biological, geomorphic and hydrologic features and ecological processes also exist
(Walker et al. 1995). These include macro-habitat, which .reiates to the whole river ecosystem
and flow regime; mesa-habitat (at the community or population level) which is related to flow
history; and micro-habitat at the organism level.
Valley floor landforms
Plant successional
processes Geomorphic and hydrological processes I Riparian vegetation
Channel structure
Figure 2.1: Summary of river processes influencing riparian vegetation (adapted from Gregory et al. 1991).
The structure of riparian environments can also be considered from various aspects within the
landscape including longitudinal (e.g. from headwaters to river mouth), transverse (river
profile) and internal (including reaches, pools, riffles, meanders and backwaters) (Malanson
1993). In a characteristic river system, the longitudinal pattern usually begins with small (first
order) streams at high elevations in the landscape at the headwaters of the catchment. These
streams tend to be fast flowing and erosive. At lower elevations the river slows down and meanders across a plain before entering the sea. This lower section can be subject to flooding and is usually an area of deposition where more material is deposited than eroded (Bayly & 16
Williams 1981). As the gradient of a river decreases the bed materials become finer, so that while source streams tend to have gravel beds, mobile beds of silt and sand are found at the river mouth. River morphology comprises many different features and at different scales which provide a variety of habitats. These features include channel form (comprising width, depth, slope, roughness, velocity and discharge) and channel pattern. The formation of these features is influenced by regional geology, climate and sediment transport in the form of erosion and deposition. Within the river channel are pool, run and riffle zones which are influenced by, and which in turn influence, water flow. These pool and riffle sequences tend to alternate along the river (Jeffries & Mills 1990). River bends are major sites of erosion and deposition as material is eroded from the outside of the bend and deposited on the inside. In many rivers these depositional areas (point bars) provide a major substrate on which primary succession of riparian vegetation communities can take place (Malanson 1993).
In floodplain rivers, along a cross-section perpendicular to the channel, several zones can be discerned, influenced mainly by microtopography and groundwater, that provide different habitats for various biotic communities. The littoral zone, that area between the seasonal normal high and low water levels, provides a habitat that is subject to long periods of inundation and high levels of disturbance. The levee bank is immediately adjacent to the river channel.
Here rising floodwaters lose velocity and deposit sediment, creating an area of slightly higher elevation and higher nutrient levels. The floodplain, which is subject to intermittent flooding, is developed by sediment deposition and can be of variable width and made up of several terraces of different elevation. Terraces are generally considered to be old floodplain levels which have become redundant as the river has incised into the landscape and, through meandering, created new floodplains at a lower level. Finally, 'abandoned' channels occur on the floodplain, usually parallel to the existing channel, which may have standing water for long periods and can behave much like lentic wetlands.
The river processes considered important for structuring riparian vegetation include water flow, water level fluctuation, sediment transport (deposition and erosion), litter transport and seed 17
dispersal (Nilsson et al. 1993). Distribution and composition of riparian plant communities
reflect mainly fluvial disturbances but also some non-fluvial disturbances such as fire or wind.
Geomorphic surfaces provide a physical template for the development of riparian communities
(Gregory et al. 1991). The flood pulse concept Gtmk et al. 1989) acknowledges the connection of
the river to its floodplain which may result in several successional stages in vegetation
development (Ward & Stanford 1995). The flood pulse provides a recharge of nutrients to the
floodplain via deposition which can result in productivity benefits. It can also lead to a
recharge of soil water which can be important for the survival of many riparian plant species m
semi-arid and arid rivers (Roberts 1993).
The existing models of rivers have been predominantly derived from studies of northern
hemisphere rivers and, although they can be applied to Australian lotic systems, there are
features of many Australian rivers that are unique (Williams 1983). Other than for the eastern
seaboard, the two major characteristics of the Australian environment that influence lotic
systems, are aridity and the flatness of the terrain. The arid climate not only results in very low
average total annual runoff compared with other countries but also in a low percentage of annual
precipitation as runoff (apart from rare large events). This has led to large river basins and
relatively long rivers with low annual flows, such as the Murray-Darling River in eastern
Australia (Bayly & Williams 1981). There is also wide variation in seasonal and annual river
flows. This is evident in most rivers in southern Australia but is most pronounced in the northern
rivers (Walker et al. 1995). In these areas, rivers carry very large volumes of water in the wet
season after the monsoon rains but can be reduced to a series of pools in the dry season, with a
great deal of annual variability. In Australia, large floods tend to be of greater magnitude than
those of other continents, with 100 year flood levels ten to twenty times the magnitude of mean
annual flood levels, compared with two to four times for other parts of the world (Lovett & Price
1999). As erosive power of flow increases disproportionally with increasing discharge, 100 year floods are extremely powerful agents for landscape change (Lovett & Price 1999). Where river basins are draining arid and semi-arid areas of the continent, rivers may be dry most of the time and only flow after a large episodic rainfall event (Walker et al. 1995). The lack of relief over 18
most of Australia results in many rivers, such as the Murray-Darling, having very low gradients.
As a consequence, short areas of erosion occur in the upper reaches and extensive areas of
deposition occur in the lower reaches, which can lead to high turbidity (Williams 1983). Many
Australian rivers, especially those draining low rainfall areas, have high salinity levels by world standards (Williams 1983). The fringing vegetation m Australian rivers also contributes to an ecology which is different from most northern hemisphere rivers. In Australia, most riparian vegetation is evergreen. Leaves are shed throughout the year, usually with a pulse in summer, instead of all at once at a particular time of year, as do deciduous trees (Pen 1999; Bunn et al. 1999). The leaves that are shed are harder, more sclerophyllous and take longer to break down and many contain chemicals that stain the water, thus affecting in-stream ecology and productivity (Pidgeon & Cairns 1981; Bunn 1988).
Plant Ecology
Vegetation dynamics
The riparian zone experiences a naturally high level of disturbance, both in frequency and intensity, due mainly to the flow of water and flooding (Malanson 1993). The high level of disturbance can result in a mosaic of successional communities (Kallida & Puhakka 1988).
Therefore, a study of the natural recruitment of riparian vegetation needs to consider the processes of vegetation dynamics under natural disturbance regimes. Furthermore, the rehabilitation of degraded areas usually follows successional pathways and sound knowledge of these processes is necessary for successful planning and implementation of the re-establishment of vegetation (Werner 1990).
Vegetation dynamics usually refers to the ways in which vegetation may change over time and incorporates the concepts of succession and plant community. It involves the successive nature of the development of plant communities over time, including the interaction of populations in uniform environments (van der Maarel 1988). Vegetation dynamics can be viewed in terms of 19
temporal scales, ranging from short-term through to long-term changes, as well as on several spatial scales including individual patches, communities, landscapes and bioregions. The plant community has three major components relating to species response to habitat; interaction between species; and chance and random processes both in the environment and in populations
(Noy-Meir & van der Maarel1988).
Earliest theories of succession developed by Clements (1916) see (Glenn-Lewin & van der Maarel
1992) were based on the idea that plant communities develop in sequence to reach a climax stable community after some perturbation. Others stressed the importance of individual responses of species to the environment and to chance events following disturbance (Gleason 1927; Egler 1954).
More recent discussion on ecological succession has considered that physical stresses on plants and competition for resources between plants are the main mechanisms determining the course of succession (Drury & Nisbet 1973). A model proposed by Connell and Slatyer (1977) suggested three mechanisms for succession which included elements of these earlier theories. These were facilitation or 'relay' floristics, where earlier species modify the environment so that later successional species can establish; tolerance, where different species evolve different strategies or life history traits for exploiting available resources; and inhibition, where species resist the invasion of competitors and the first occupants prevent the establishment of later colonisers until they are damaged or die and thus release resources. These mechanisms differ mainly in the way later species establish. The facilitation model is most likely to apply in primary succession where the substrate has not been influenced by species beforehand, whereas tolerance and inhibition models will apply to secondary succession (Connell & Slatyer 1977). Connell and
Slatyer (1977) emphasised the importance of competition in the succession process and in the determination of species composition. Competition between species for limited resources was also considered critical for species composition and succession in the resource ratio hypothesis, where the ratio of limiting resources is important rather than a single resource (Tilman 1986).
This hypothesis suggested that community composition is largely influenced by limiting resources and that patterns of succession will change with relative availability of these resources. 20
The initial floristic composition model proposed that many species which initially establish may not necessarily give way to other species later in the -successional sequence but may become dominant themselves (Egler 1954). Evidence of this has been reported after fire in forests in eastern Australia (Purdie & Slatyer 1976). This model may also apply in the jarrah forest of south-western Australia due to the predominance of resprouter species in the vegetation (Bell &
Heddle 1989). In the high disturbance riparian zone, where vegetative resprouting is probably very significant, this initial floristic composition model may also apply.
Models of succession are difficult to apply to spatially heterogeneous ecosystems such as the riparian zone and the concept of ecosystem stability has been suggested (Fischer 1990). This concept allows the ideas of disturbance, stability and resilience to be directly incorporated into the description of vegetation dynamics. Rather than unidirectional succession, dynamics within vegetation communities may result in a series of patches made up of transitions between more or less stable states mediated by disturbance, climate or management (Hobbs 1994). Different vegetation types within an area may thus not be predictable by current environmental factors and species interactions only, rather the present state is the result of a particular history of disturbance, climatic factors and/or management. The regeneration phase and longevity of the vegetation can be critical in determining the stage and rate of transition states. For example, a tree-dominated community, where there is little or no recruitment, would take a long time to change because of the long-lived nature of the trees, unless they are killed by some disturbance.
Hobbs (1994) suggested that a synthesis of qualitative transition state models (ie. Westoby et a I.
1989) with the quantitative, statistical Markov models (Hobbs 1983) could be used in a geographical information system (GIS) to predict landscape scale changes to vegetation due to disturbance. State and transition models have been suggested as a tool for rangeland management
(Westoby et al. 1989; George et al. 1992), and as an alternative to linear succession, to allow for different vegetation community composition (states) depending on the transitions. Management can be used in conjunction with natural disturbances to direct vegetation composition to a desired 21
state (Hobbs & Norton 1996). This model has been applied to riparian vegetation dynamics on a
montane river in the U.S.A. (Baker & Walford 1995).
Large scale disturbances, such as high water flows and flooding, are likely to be important
factors in the dynamics of the riparian community. Flood induced disturbance may instigate the
transition from stable states of plant composition and population structure. Recruitment,
establishment and survival of riparian trees is thought to be closely tied to the hydrological
regime (Barnes 1985; Bradley & Smith 1986; Wissmar & Swanson 1990; Johnson 1994; Naiman et
al. 1998) and there is obviously a strong link between river processes and the dynamics of
riparian vegetation. The dynamics can vary longitudinally along the river as well as laterally
and are a function of geomorphology, vegetation history and vegetation life history strategies
(Naiman et al. 1998). The interaction of plant life history processes and the physical processes
of rivers can determine the typical patterns of vegetation colonisation, establishment and
succession. For example, for many riparian species seed dispersal coincides with seasonal retreat
of floodwaters, when moist sediments are exposed which enhance germination and colonisation
Gohnson 1994; Naiman et al. 1998). Floods can affect vegetation patterns by destroying and
subsequently excluding plants, by creating new areas for colonisation and by forming elevational
gradients where plants show varying tolerance to flows and sediment movement (Wissmar &
Swanson 1990). High frequency flooding can keep vegetation in an early stage of succession
(Naiman et al. 1998). How fluvial processes can affect vegetation life history patterns can be assessed by gathering information on magnitude, frequency and duration of flows. Patterns of these flows will influence successional stages of the vegetation. For example, where flooding disturbances are chronic to frequent, riparian forests are young, as vegetation patch development is continually reset and remains in stand initiation and exclusion successional stages (Wissmar &
Swanson 1990). In environments with low magnitude, seasonally predictable floods the vegetation is likely to be more stable and hence more developed. Floodplain succession is therefore an interaction between stochastic flood events and life history traits as well as the extent, intensity and timing of flooding, which determines patterns of vegetation colonisation.
Lateral patterns of the riparian vegetation can also occur, with vegetation of low terraces and 22
floodplains mostly influenced by factors such as moisture gradients and flooding tolerance. On higher terraces and slopes the vegetation is more influenced by biotic factors such as competition and herbivory. Heterogeneity of soils and habitat and .microtopography are also important influences on riparian vegetation dynamics.
Restoration ecology
As discussed earlier, the majority of riparian zones in Australia have suffered some form of degradation (CEP A 1992). Discussion of restoration ecology is therefore appropriate when considering the recruitment of native species in the riparian zone. Characteristics of the natural ecosystem that need to be considered in restoration ecology include composition, structure, pattern, heterogeneity, function and resilience (Hobbs & Norton 1996). In restoring a degraded ecosystem, removal of the factors causing the degradation is the essential first step. The likely success of natural regeneration once these factors have been removed will depend on how they have affected ecosystem function (Brown & Lugo 1994). If the degradation stress has affected only biotic components of the ecosystem, such as through the removal of the vegetation, then removal of this stress may result in fairly rapid recovery without further management intervention (Allen et al. 1994). However, if the degradation stress has affected the resource base or ecosystem processes, such as the soil or hydrology, restoration may be more difficult
(Milchunas & Laurenroth 1995). Hobbs and Norton (1996) identified a number of key processes that need to be considered in ecosystem restoration. These included identifying degrading processes, developing methods to reverse the degradation, determining realistic goals, developing observable measures of success, developing practical techniques and communicating results. The restoration of degraded systems to some desired state (whatever this may be) in ecological terms also requires the re-initiation of successional processes by accelerating biotic change (Luken 1990). The identification of factors important in recruitment of riparian species will provide important information in developing a restoration framework for degraded riparian systems. 23
Plant life-history traits
Modelling succession and species composition using life form, life history traits, functional groups or vital attributes has been widely used. These models ·are usually developed for particular habitats or types of disturbance and, as generalisations to describe succession, they may be oversimplified (Brown 1992). Noble and Slatyer (1980) devised a qualitative model of vegetation dynamics in habitats disturbed by fire, characterising the plants into three life history groups with certain attributes important in succession. Using the interaction between categories (groups) they could predict changes in composition and dominance of species following disturbance. The three categories were method of persistence and arrival after disturbance; ability to establish and grow in a developing community; and time taken for species to reach critical life stages. This model has obvious application in a study of grazing disturbance and provides a basis for discussing the role of disturbance frequency in community types such as the riparian zone. In communities which are subject to more than one recurring disturbance, such as grazing and flooding, then more than one vital attribute may be appropriate for each species.
Where a community is subject to a continuous disturbance, such as grazing (which prevents reproduction) Noble and Slatyer (1980) predicted that plants in all but two categories of attributes would become locally extinct. That is, only those species with widely dispersed seeds and those species with the ability to withstand disturbance with little harm, as well as having widely dispersed seed, would be able to survive in the long term.
Plant strategies in response to disturbance have been described to explain species co-existence for plant species of the British Isles (Grime 1979). This scheme suggests that vegetation develops as a result of equilibrium between competition, stress and disturbance where disturbance is the destruction of extant growth by outside factors (e.g. fire), stress is the reduction in rate of growth by outside forces causing a shortage of resources (e.g. aridity) and competition is limitation to growth through shortage of resources due to capture by other plants (Grime et al. 1986). This system distinguishes plants into three main groups (Grime 1980):
1) Competitors: capable of high productivity, vegetative expansion and usually flowering
each year after maximum vegetative production. 24
2) Stress tolerators: slow growing, with long life histories and for which onset of flowering is
usually delayed and intermittent.
3) Ruderals (disturbance tolerators): usually small annual herbs with dormant seeds, short
life histories and early flowering.
Identification of the different strategies among plants within a community can be useful in providing a means of predicting the community response to continued disturbances such as grazing.
For example, this scheme predicts that under a regime of high stress and high disturbance, ro strategy exists to allow plants to survive and therefore no vegetation is possible. This situation may occur in riparian areas that are continuously grazed. These plant functional strategies were applied to regularly disturbed riverine wetland ecosystems in which competition and stress tolerance were found to be the most important traits (Hills et al. 1994). Grime's plant strategies have been criticised as being too simplistic and too specific to British grasslands (Glenn-Lewin & van der Maarel1992).
Life history traits have been used to predict species response to fire in Australia (Gill 1981) and the effect of fire on sandplain species in Western Australia (Bell et al. 1982). Reproductive strategies of plants that occur in the riparian zone are not well understood (Malanson 1993) but may be important for predicting the outcome of succession in this highly disturbed zone.
Classification of plants by life history strategies can be useful in predicting riparian vegetation succession (Naiman et al. 1998). Naiman et al. (1998) categorised plant species according to their strategy in coping with flooding disturbance as invaders, endurers or resisters. A model for predicting the outcomes of succession in wetlands has also been developed using three key life history traits (van der Valk 1981). These traits include life span (annual, perennial and vegetative perennial), propagule longevity in the seedbank and propagule establishment requirements. This model relies on using the composition of the seedbank to predict the vegetation composition after a disturbance. Many other attempts to classify vegetation communities into functional groups in order to predict the outcome of changes have been made.
For example, herbaceous wetland plant traits were selected that reflected function and included ruderals, matrix specie~ and interstitial species (Boutin & Keddy 1993). Grouping plants into life 25
and growth form groups has also been used (Rodolfo & Sala 1993; Mclntyre et al. 1995), as have reproductive and regeneration characteristics (Cowling et al. 1994), to help describe successional processes and species co-existence in plant communities. Grime and Hillier (1992) have suggested that accurate predictions of plant communities can be made with a good theoretical framework of regenerative functional types for a flora.
Disturbance
Natural disturbance and patch dynamics are important concepts in plant ecology when determining the structure and function of communities. Disturbance is considered to be important in maintaining species diversity by creating gaps in vegetation that has exploited most of the resources and, through competition, excluded other individuals (Abugov 1982). The creation of patches through disturbance allows germination and establishment of new individuals which may be of different species. In the riparian zone, patch dynamics may allow the continuous recruitment of individuals between infrequent large scale disturbances such as flooding.
Disturbance reduces the dominance of a site by established plants and, by altering the availability of light, water and soil nutrients, creates openings for colonisation and growth of new individuals not necessarily of the same species (Canham & Marks 1985). Patch dynamics may be defined as the internal dynamics generated by patterns of natural disturbance and subsequent patterns of succession (Pickett & Thompson 1978). The first three forms of vegetation dynamics detailed by van der Maarel (1988), that is, fluctuation, gap dynamics and patch dynamics, describe the types of processes that make up this definition of patch dynamics. These different terms relate mainly to the size of disturbance and the subsequent size of patches created. For example, fluctuations are considered on the level of the individual plant and gap dynamics may result from the death of an individual plant, resulting in the opening up of the area to light, thus affecting plants in the immediate vicinity.
A plant's ability to exploit patches will depend on reproductive traits, such as the ability to expand vegetatively, seed disperal, the ability to form a seedbank and also to exploit increased resources quickly (Sousa 1984). The natural disturbance regime will determine the size, density 26
and temporal frequency of patches and the range in temporal and spatial distribution of patches may allow the calculation of the natural disturbance regime within a community (White &
Pickett 1985). How disturbance shapes community structure will depend on the frequency, magnitude, predictability and type of disturbance (Sousa 1984). In the riparian zone, naturally high levels of disturbance can cause stream migration, and erosion and deposition are continually
destroying and creating new sediment substrate which allows colonisation by plants (White
1979). Disturbances such as fire, flooding and drought are very important for the recruitment of salmon gum (Eucalyptus salmonophloia) in woodlands in the wheatbelt of Western Australia
(Yates et al. 1994) and flooding is essential for river red gum (Eucalyptus camaldulensis) recruitment on the Murray River in eastern Australia (Dexter 1978; Bren & Gibbs 1986). Small scale disturbance has been shown to be influential in maintaining species diversity in mediterranean old fields (Lavorel et al. 1994), in wetland communities (Mclntyre et al. 1988) and in temperate grasslands and herbaceous vegetation in eastern Australia (Mclntyre & Lavorel
1994; Mclntyre et al. 1995).
Many river systems (including the ones studied here) are subjected, at least for some part of their length, to livestock grazing. This has serious consequences for riparian systems in terms of loss of native species through grazing pressure, prevention of seedling recruitment and the replacement by better adapted exotic weeds and pasture species, resulting in a change of structure and species composition. Livestock, through trampling, can cause compaction and soil erosion, thus changing the physical properties of the soil. With the exclusion of livestock grazing, recovery of the native vegetation can be prevented by competition from exotic species (such as annual grasses) and changed soil properties, even where propagules of native species are available. Effects of livestock on the riparian zone have been reported in North America (Kauffman et al. 1983;
Armour et al. 1994), documenting changes to the structure and composition of the riparian vegetation, and for other landscapes in Australia (Wilson 1990) but there has been little work specifically on the riparian zone in Australia. 27
Seed and Seedling Ecology
Once disturbance has created gaps in the vegetation for recruitment of new individuals, successful
recruitment will depend on particular reproductive traits and the seed ecology of particular
species (Canham & Marks 1985). Factors that are important in successful regeneration include
abiotic factors such as the creation of 'safe' sites for seeds and favourable climatic conditions.
Important biotic factors include ability to reproduce vegetatively, production of seed, timing of
seed release, dispersal mechanisms of seed, seed dormancy and the ability to create a soil
seedbank as well as germination requirements of seed and requirements for seedling
establishment.
The regeneration niche theory (Grubb 1977) proposes that replacement of species is primarily
driven by the biotic and physiochemical heterogeneity of the environment. Therefore the
variable environment may provide new niches and determines whether there is a place for the
co-existence of species. This regeneration involves vegetative and reproductive parts of plant
life histories, the relative importance of each depending on the type of corrmmnily. Grubb (1977)
proposed five processes important for the regeneration of plants, including production of seed,
dispersal, germination, establishment of seedlings and the development of the mature plant.
These elements incorporate the factors which are considered important in the differentiation of
the regeneration niche. These can be usefully applied to regeneration traits of species within the riparian zone.
The production of seed is obviously critical for successful recruitment of species which have ro ability to resprout vegetatively (obligate reseeders). Timing of seed production and seed fall is also very important to ensure that seed is released at times of most favourable conditions, such as after receding floodwaters when fresh soil is exposed (White 1979), soil moisture levels are high
(Roberts 1993) and there is peak abundance of seed vectors (Willson 1992). Production of large numbers of seeds may reduce the impact of seed predators (Ashton 1979; Wellington & Noble
1985a) and can also compensate for low seed germination rates (De Steven 1991). 28
One aspect that can influence the coexistence of species in plant communities is propagule
dispersal traits (Levin 1976). The development of seed dispersal syndromes has been influenced by the probability of finding a 'safe' site, avoidance of sibling interaction and the avoidance of seed predators (Willson 1992). Dispersal to a safe site is particularly important for species that have no capacity for seed dormancy. The determination of dispersal spectra can be useful in identifying the relative advantage of different dispersal types in particular habitats (Willson
1992). In riparian landscapes, Malanson (1993) described three main vectors of seed dispersal including air, water and animals, with seven sub-groups. The relative importance of different seed dispersal types may change in various parts of the river with greater occurrence of wind dispersal at the headwaters and animal dispersal being more important in the lower reaches
(Malanson 1993). In fragmented landscapes with narrow riparian corridors, wind dipersal of fugitive species may become more prevalent. In rivers in Sweden, floodwaters carrying seed greatly extended the dispersal range of wind-dispersed species, dependent on the ability of the seed to float and the length of time they could float (Skoglund 1990). In fact, the long hairs on a seed which appear to aid in dispersal by wind may also increase ability to float which may be the prime means of dispersal (Herrera 1991). Floating time can also be important for dispersal distance with larger, heavier seeds establishing closer to the point of release (Nilsson et al.
1991b).
To understand the role of the seedbank in vegetation dynamics Thompson (1992) suggested that it is necessary to estimate the buried seed density, measure seed production and seed 'rain' and also determine the source of colonists. The seedbank should also be examined at different times of the year in order to assess seasonal differences and to separate the long term seedbank from the transient (Grime 1979). Soil seed storage is an important aspect of natural regeneration (Onans &
Parsons 1980). Species with seeds which remain dormant in the soil have the advantage over resprouting species of being able to remain viable for a very long time. Seeds of herbaceous species such as Digitalis purpurea and Hypericum paletrium have been found to remain viable in the soil for 100-200 years (Thompson 1987) while seed of some Acacia spp. have remained viable for at least 50 years (Mott & Groves 1981). Species with a viable long-term seed bank are 29
' very important in restoration of grasslands (Thompson 1992), while in temperate woodlands few
perennials occur in the seedbank and those that do are usually short-lived early colonisers. In
the jarrah forest of Western Australia, where there is a high. proportion of resprouting species in
the flora (Bell et al. 1990), only a limited number of species are represented in the seedbank
(Vlahos & Bell 1986). Overstorey species are usually not present in any number in the seedbank
as there is usually ample opportunity over time for regeneration of long-lived perennials. In
contrast, the seed of annual herbs and grasses provide the only opportunity for regeneration, and
dormancy ensures germination only under favourable conditions (Schneider & Sharitz 1986;
Thompson 1987). In lentic wetlands the seedbank is very important in determining the
vegetation community (Thompson 1992), and determination of the composition of the soil
seedbank is used to predict the outcome of succession after a disturbance (van der Valk 1981). The
role of the seedbank in riparian vegetation is unclear and little work has been done on this,
however Malanson (1993) suggests that the highly disturbed nature of the riparian zone and the
frequent movement of the soil sediments would indicate that the role would be minimal.
Generally, most of the seed of eucalypts that is released is collected by seed-harvesting ants.
Heavy predation of eucalypt seed has been observed in eucalypt woodland in Queensland (Drake
1981) and in mallee scrub in northern Victoria (Wellington & Noble 1985a), where up to 100% of
eucalypt seed was removed in a short period of time when seed was placed in feeding stations. It
has also been observed in salmon gum woodlands in the wheatbelt of Western Australia (Yates et
al. 1995). In these circumstances a build-up of eucalypt seed in the soil occurs only after a
disturbance (such as fire) causes massive seed release from the canopy and results in predator
satiation (Ashton 1979; O'Dowd & Gill1984; Wellington & Noble 1985a). However, even where
seed predation levels are very high, the limiting factor in recruitment of seedlings may in fact be
the number of safe sites for germination (Andersen 1989; Crawley 1992; Stoneman & Dell 1994).
Many plants maintain a seedbank for several years on the plant itself, usually in woody fruits. A
large proportion of species in the Myrtaceae family which have woody fruits retain large stores
of seed in the fruit for several seasons, releasing only after desiccation, through fire or death of a
limb (Beardsell et al. 1993). For many temperate eucalypts, seed maturation can take several 30 months and seeds are retained in the capsules with a continuous low level of seed release occuring, which is accelerated by dry conditions (Cunningham 1957; Wellington & Noble 1985a;
Davies & Myerscough 1991). Eucalypt seed is rarely held- in the canopy for more than three or four years (Ashton 1975). A canopy-borne seed store is relatively common in families such as
Myrtaceae and Proteaceae in Australia (Bellairs & Bell 1990).
Seeds germinate in response to several different stimuli including light, temperature and moisture, and the germination of seed of a particular species can be related to life history traits and response to fire or seed storage (Bell et al. 1993). In many species, seeds germinate soon after they are shed, regardless of conditions. In other species, seeds germinate only under favourable conditions (enforced dormancy) and can remain for long periods in the soil seedbank with intermittent germination from the population (Murdoch & Ellis 1992). Seed dormancy may be innate where germination does not take place even though suitable conditions prevail. This may be due to under-development of the embryo and innate dormancy occurs in many annual species which form long-lived seedbanks (Murdoch & Ellis 1992). Forced dormancy occurs where prevailing conditions are unsuitable for germination and, for example, in many Acacia spp. which have a hard, impermeable seedcoat. Germination is stimulated when the seedcoat is cracked or inhibitors are denatured by heat, usually from a fire (Bell et al. 1993). In experiments on jarrah forest species, greatest germination was recorded at temperatures typical of winter when moisture conditions would be suitable for survival of germinants. Most species also germinated best in continuous dark which simulates seed burial (Bell 1994). Light requirements of germinating seed are also important in terms of the level of shade tolerance of particular species and whether they will successfully establish under shade or within canopy gaps (van der
Valk 1992; Bell 1994). In lentic wetlands, because of the unpredictable nature of water levels, temperature provided the best cue for germination of seed from the soil seedbank (Britton &
Brock 1994).
Once seed has germinated, it is critical that adequate moisture is available, at least until the seedlings have developed a substantial root system (Wellington & Noble 1985b; Stoneman et al. 31
1994), Therefore, the coincidence of seedling establishment and adequate rainfall is critical,
particularly in Mediterranean-type climates during summer drought. In riparian areas, flooding
provides a means of restoring soil water which is critical ·for seedling establishment (Roberts
1993). For seedlings to establish in the riparian zone, fast growth rates may be necessary (White
1979) and/or some sort of protective barrier (Hancock et al. 1996) so that they can withstand
floodwaters the following season. For seedlings establishing on riparian floodplains suitable
microsites may also be critical, with some elevation to reduce the effects of waterlogging, but
also within a slight depression so that moisture can gather during the dry summer months (Titus
1990; Bren 1992). Degree of shading can be very important for the establishment of many shade
intolerant species, with seedlings only establishing in tree-fall gaps where there are high light
levels (Ellison et al. 1993). In a riparian habitat, seedlings growing in heavy shade were two
and a half times smaller than seedlings growing in the open (Jones & Sharitz 1989). High levels
of leaf litter can also have a large impact on seedling establishment, principally through
forming a physical barrier to root penetration in the mineral soil. Flooding can exert a
considerable influence on riverbank vegetation as a result of litter displacement, allowing seeds
and seedlings access to the mineral soil (Nilsson & Grelsson 1990). Study of the seedling flora of
a vegetation community can give important insights into the relative importance of the different
seedling strategies within that community and therefore be useful in predicting community
dynamics after disturbance (Grime & Hillier 1992).
Vegetative regeneration
In many types of vegetation, the majority of perennial species possess the ability to reproduce vegetatively by a variety of mechanisms such as rhizome, stolon or lignotuber. Vegetative resprouting is particularly common in communities which receive frequent periodic disturbance
(James 1984) such as in many wetlands and rivers (van der Valk 1992). For many vegetatively propagating species, such as rushes and sedges, the ability to colonise a new area depends on how far new ramets can establish from the parent plant (van der Valk 1992). For example, many species have a clumped growth form with new ramets produced very close to the parent plant.
These species will colonise an area very slowly, if at all, whereas species which can send out 32
runners and produce independent ramets at some distance from the parent plant can colonise an area more rapidly (van der Valk 1992). The common bracken fern (Pteridium esculentum) is an example of a plant that can rapidly colonise areas by producing new ramets from long rhizomes.
However, for the majority of species that are able to propagate vegetatively, rapid colonisation of a new area would require establishment from seed. The major role of vegetative propagation for plants is therefore, in most cases, to allow them to persist at a site, particularly following a disturbance.
Many woody plants have the ability to resprout from underground storage organs after defoliation. This has been extensively studied in relation to recovery of vegetation after fire
(Gill 1981; Bell et al. 1982; Keeley 1986). Resprouting shrubs and trees usually possess one of a number of structures such as a lignotuber or root crown in which carbohydrate reserves and dormant buds are located. The dormant buds are released from inhibition and begin to sprout after the defoliation of the plant. Resprouting species have a greater dependence on the survival of the individual after defoliation whereas non-resprouting species (obligate seeders) rely on the availability of seed to continue the local existence of a species after a disturbance such as fire (Carpenter & Recher 1979). Consequently, obligate seeders tend to put more resources into reproductive effort. Resprouters usually take longer to reach reproductive maturity, produce less flowers less often and have a lower rate of seed production per plant (Hansen et al. 1991; Bell et al. 1993). Seedlings of resprouter species tend to be slower growing as energy is directed towards building up reserves of carbohydrates for resprouting. As a result they have a much smaller shoot to root ratio and a greater proportion of starch reserves in the roots than obligate seeder species (Pate et al. 1990). However, the rapid recovery of established resprouters gives them a competitive advantage over seedlings so that resprouting species may dominate the post disturbance habitat (Malanson & O'Leary 1982). Replenishment of starch reserves is an important factor in the ability of reprouters to continually resprout after disturbance and very short disturbance intervals increase the chance of mortality as starch reserves may not fully recover until shoot biomass has reached pre-disturbance levels, possibly taking several seasons
(Bowen & Pate 1993). . 33
In highly disturbed riparian habitats, the ability to resprout as well as to produce large quantities of seed would be highly advantageous. Many species that occur in riparian habitats in North America, such as Salix spp. and Populus spp., possess the ability to resprout and produce large quantities of seed Q"ohnson 1994). This is also the case in some of the dominant riparian trees in Australia, such as Eucalyptus camaldulensis and E. rudis. Seedlings that establish m new substrates must be able to withstand high water flows in the following or subsequent seasons which could damage and defoliate plants. The time it takes for a species to build up sufficient carbohydrate stores, and therefore the ability to resprout, as well as the interval between floods, are likely to be critical in determining survival during subsequent flood events. An ability to resprout would also enable mature plants to survive physical damage caused by flooding.
Regeneration of riparian zones
The processes that affect regeneration of vegetation in riparian zones, such as dispersal, establishment and growth, are ultimately dependent on the landscape structure and function as well as on landscape scale disturbance regimes. Spatial changes among species are affected by temporal changes in the distribution of physical structures and functions and also by the interaction between species (Malanson 1993). In riparian zones, species reproductive strategies may be adapted to the particular high frequency disturbance regime that occurs. These traits may include regular annual seed crops, light seeds which are dispersed by wind and/or water, fast growth rates, ability to resprout and high flood tolerance (White 1979). Floodplain ecosystems tend to be disturbance dependent with a high sub-system instability but a broader system stability (Ward & Stanford 1995).
Recruitment, establishment and survival of riparian trees has been thought to be closely tied to the hydrological regime. There has been evidence of this in North America where cottonwood recruitment on point bars correlated with years of maximum daily flow greater than second year 34
flood return and these coincided with seed release approximately once in every five years
(Bradley & Smith 1986). For Populus spp. and Salix spp. on the Platte River, sununer flooding
reduced seedling recruitment and initial survival as it coincided with the normal period for seed
germination, whereas low spring flows exposed more river bed and therefore more substrate for
seed germination and survival Gohnson 1994). Late sununer survival of seedlings depended m
river stage, elevation of seedlings and rainfall. On a river island in Wisconsin, successful
establishment of vegetation occurred on bare mineral soil when weather conditions were
favourable and when there was no prolonged or extensive flooding until the seedlings had become
established (Barnes 1985). Regulation of a central U.S. river brought about the loss of significant
water level fluctuations resulting in an absence of areas of bare moist substrate and lead to a lack
of regeneration of riparian tree species (Friedman et al. 1995). In south-west Colorado, stands of
Populus angustifolia originated in discrete cohorts in years with both high spring and autumn
peak dicharges (Baker 1990). On this river, key environmental determinants of riparian vegetation development were cohort age, surface sediment type and soil development, with
cohort age correlated with past flooding. Therefore the vegetation mosaic is determined by the periodic disturbance of floods (Baker & Walford 1995).
The effects of fluvial disturbance can be widespread in the river valley, with numerous patches of early successional vegetation close to the river channel (Gregory et al. 1991), while on the floodplain, further from the influence of the active channel and low level flooding, older plant communities are more common (Hawke & Zobel 1974). The vegetation structure can therefore be strongly influenced by the timing and character of recent large scale disturbances (i.e. floods) and the age of stands is related to height above the channel (Baker 1988). Failure of the vegetation to reach later successional stages due to high frequency flooding disturbance has been documented for New Zealand podocarp forest (Duncan 1993). This importance of hydrology was also observed for rivers in southern France, with flood frequency and duration being important in determining species composition. For the rivers of this region, a strong correlation existed between species composition and hydrology, while other significant factors included soil type and geomorphology (Pautou & Decamps 1985; Tabacchi 1995). For a mediterranean riparian 35 shrub species, seed release of wind- and water-dispersed seeds coincided with peak river flows.
Once seeds had dispersed and germinated the principal cause of seedling mortality was somewhat evenly divided between drought (46%) and flooding (41%) (Herrera 1991).
Other factors have been reported as being important in determining establishment success for riparian species. For example, the critical importance of microsites for establishment of various species has been demonstrated in a flood plain wetland (Titus 1990) and for river red gum on the
Murray River floodplain in eastern Australia (Bren 1992). In these examples, microsites are important to elevate seedlings from excessive waterlogging but also to provide a site to trap moisture, thus reducing desiccation in the dry period of the year. In a floodplain forest in Texas,
U.S.A., effects on seedling mortality included flooding, drought, light and herbivory, with some factors being more important in different years (Streng et al. 1989). For a river in Sweden, the effects of shade and leaf litter were found to be critical in successful establishment (Skoglund &
Verwijst 1989). A strong effect of shading on seedling growth has been shown for floodplain forest species in eastern North America Gones & Sharitz 1989). Litter can have either positive or negative effects on the germination and establishment of riparian vegetation (Xiong & Nilsson
1997). Litter can improve water availability under dry conditions, protect seed from predation or reduce competition but can have negative effects through shading, obstructing emerging seedlings or reducing temperature. Floating litter debris can enhance dispersal of plant diaspores (Xiong &
Nilsson 1998), however seeds may not germinate in deposited litter as it may be too loose to retain moisture for germination (Nilsson et al. 1993).
Species diversity of vegetation in the riparian zone can be influenced by many factors including level of disturbance, tolerance to waterlogging (which is influenced by topography), soils and geomorphology, life history traits and climate. On a floodplain, adaptation of species to flooding can result in the development of different communities, with those in flood-prone areas consisting mainly of flood-resistant stress-tolerators while at higher elevations species with competitive traits tend to dominate (Blom et al. 1990). Ability to establish in the riparian zone can be related to some extent to physiological tolerance to flooding, in both adults and seedlings 36
(Pezeshki & Anderson 1997). This can be through etiolation of the stem above water levels or through the development of adventitious roots and arenchyma tissue in the roots (Kozlowski
1984). Riparian vegetation communities can be described ·in relation to topographic position, from the river flat to various levels of river terrace, with the youngest communities on the river flat (Fonda 1974). In a gradient analysis of floodplain forest in Illinois, species diversity and richness increased with decreasing flooding stress (Bell & del Moral 1977). Seasonal variation in climate, particularly rainfall, can also be important in determining species composition in riparian habitats. In dry years disturbed sites are most likely to be invaded by dryland species whereas in wet years invasion tends to be by species dispersed by water (Decamps et al. 1995).
Life history traits are also very important and riverine species of Salix and Populus are ideally suited to this habitat, producing large seasonal seed crops which are effectively dispersed by wind and water, having rapid germination and growth as well as an ability to reproduce vegetatively Gohnson 1994). In a riverine wetland, peat thickness and water level were the abiotic variables most strongly correlated with plant community composition. However, only half the species were accounted for in multivariate analysis of the environment variables, with historical factors such as change in water regime and fire frequency also explaining a large proportion of the variation (Jean & Bouchard 1993).
In Australia, recruitment of river red gum (Eucalyptus camaldulensis) seedlings on the floodplain of the Murray River is limited by poor establishment of seedlings rather than lack of seed germination. As well as seeds requiring safe sites, flooding of the surrounding forest is necessary every two years to replenish soil moisture (Dexter 1978). Flooding should occur in late winter and recede by late spring to coincide with seed fall in areas of receding floodwaters. Lack of recruitment in these forests has been attributed to increased flooding in summer due to river regulation (Bren & Gibbs 1986). On the Murray River, establishment of the plant community in the littoral zone of the river bank is dependent on falling water levels exposing new substrate for seed germination. Establishment is also affected by water currents, wave action and herbivore grazing so that in many. cases regeneration can only take place through vegetative reproduction 37
(Roberts & Ludwig 1991). On an ephemeral creek in the semi-arid region of eastern Australia,
analysis of size-class structure of Eucalyptus coolabah trees indicated that recruitment was
episodic and related to past large rainfall events where flooding was considered important to
replenish soil moisture, allowing seedlings to survive the dry season (Roberts 1993). Flooding
tolerance of establishing seedlings of Tristaniopsis laurina and Acema smithii could also
influence the distribution of these two species in a riparian rainforest in Victoria. Seedlings of T.
laurina survived waterlogging for a much longer period and therefore were found on the wetter
sites along the river (Melick 1990).
Conclusions
From this review there are clearly many attributes of riparian vegetation that are important in
terms of possible regeneration and rehabilitation. These include concepts of vegetation dynamics
and how these relate to the unique physical environment of the riparian zone. Added to this are
the effects of disturbance, particularly flooding and livestock grazing, on the processes of
regeneration. By describing species composition and life history attributes and how these vary
across the elevational gradient of the river, mechanisms that are significant for regeneration of
riparian vegetation may be shown. The review has shown that particular life history traits can
be important in different plant communities depending on the disturbance regime, management
inputs and the physical environment. Although this has been documented for many different
types of plant communities, little work has been done specifically on riparian vegetation and
particularly not in Australia. The life history traits of both native and exotic plants in the
riparian zone will be examined here and can be used to predict the outcome of disturbance.
The literature emphasises the importance of disturbance in the dynamics of vegetation
communities, particularly in the riparian zone. In many riparian communities, populations are established by mass recruitment events after some large disturbance, principally flooding, and
the population may be maintained between large disturbance events by occasional recruitment 38 after some small disturbance (patch dynamics). The role of disturbance is certainly crucial and methods used in the current study focus on determining the relative importance of large scale disturbance (i.e. episodic events) and patch dynamics. (i.e. opportunistic recruitment) in recruitment of riparian species. Regeneration in the riparian zone is also clearly tied to the hydrological regime so that an understanding of the regime of the particular river system studied is very important. The regeneration strategies that will predominate in riparian communities will be adapted to regular high frequency disturbance by flooding. These strategies include high annual seed production, wind- or water-dispersed seeds, fast growth rates and the ability to resprout vegetatively. The combinations of these strategies which occur in the riparian zone are not well documented and this will be undertaken as part of this study.
Overseas studies, particularly many in North America, have shown the importance of hydrological events on various aspects of seed ecology and phenology, including timing of seed release, dispersal of seed, exposure of moist sediment, the germination and survival of seedlings and the population dynamics of particular species (Bradley & Smith 1986; Johnson 1994; Naiman et al. 1998). These aspects will be explored, in an Australian context, for the two rivers described in this study. 39
CHAPTER3
STUDY SITE DESCRIPTIONS
Introduction
The studysites for this project were located on two rivers, the Blackwood River in the south- west and the Ord River in the Kimberley region of north-west Western Australia (Figure 3.1 &
3.2). The Blackwood River was chosen as being representative of river systems in the south-west
of Western Australia, with a Mediterranean-type climate and predictable seasonal patterns of rainfall and river flow. In contrast the Ord River is a sub-tropical northern Australian river with river flows that are greatly affected by monsoon rains, resulting in large wet season flows and virtually no flow in the dry season {Table 3.1). The riparian zone on each of these rivers has, at least in some sections, become degraded by livestock grazing and as a consequence of agricultural clearing.
Table 3.1: Catchment statistics for the Ord and Blackwood Rivers.
Catchment Rainfall Mean Max. Water Disturbance Area Range Annual Recorded Quality 2 1 (km ) (mm yr ) Flow Discharge 6 3 1 (x 10 m3) (m sec- ) Ord River 46,100 400- 700 4,320 31,000 Fresh Cattle grazing, river regulated in lower reaches Blackwood 28,000 350-1000 659 1,190 Brackish 85% cleared River farmland, sheep & cattle razin 40
Location
Ord River
The Ord River stretches from its source, in the Durack Ranges, north-west of Halls Creek, to empty into Cambridge Gulf north-west of Kununurra (Figure 3.1). Vegetation in the catchment is predominantly open grassy woodland (Beard 1979) which is grazed by cattle. On the Ord River plains in the upper catchment, degradation of the vegetation through a long history of overgrazing has resulted in severe erosion and extremely high sediment loads in the river. Over large areas, such as at Ord River Station, rehabilitation through fencing and de-stocking has taken place (Fitzgerald 1967). The river has also been dammed in two places, resulting in three discreet sections. Upstream from the Ord River Dam and Lake Argyle, the river flow is essentially natural, with flows influenced by a climate of wet season monsoon rainfall and a long dry season when the river dries up to a series of pools. The middle section of the river, between the Ord River Dam and the diversion dam at Kununurra, has been flooded and water is permanent with levels fluctuating little throughout the year. Downstream of the diversion dam flows are artificially regulated to provide water for irrigation. Water levels are therefore higher in the dry season than under natural conditions and lower levels are experienced in the wet season. Study sites were located in a longitudinal sequence from downstream to upstream, with Site 01 being located downstream of the diversion dam near Carlton Hill Station. Site 02 was located in the middle section of the river, between the Ord River Dam and the diversion dam on Lake Kununurra. Monitoring sites upstream of Lake Argyle were located on the Behn
River (Site 03) near the Duncan Road crossing, on the Negri River (Site 04) near the Duncan
Road crossing and at the Old Ord River Station (Site 05) adjacent to a Water and Rivers
Commission gauging site (Figure 3.1). 41
Ord River Catchment
. I I D I I I I I
I· 'I I I I "I• I I nurturraI
I SitG02
t N O:::km~,.,.....;;;;50~k~m~,_,.:1.;;:,;;00km
Figure 3.1: Location of the study sites within the Ord River catchment. Blackwood River Catchment
Western Australia
200 km 1-----l
Okm 50 km 100km =
Figure 3.2: Location of the study sites within the Bla>Ckwood River catchment >!::> N 43
Blackwood River
The Blackwood River rises in the low rainfall area of inland south-western Western Australia and flows westward to enter the sea at the south-west tip of the continent (Figure 3.2). The river is around 300km long and carries the largest flow of any river in the south-west. Rainfall in the
-1 -1 catchment ranges f rom around 400 mm yr in the east to 1200 mm yr near the coast. The upper and middle reaches are influenced by the large predominance of cleared agricultural land, with many areas having been cleared for 100 years and livestock grazing taking place right up to the river's edge (Olsen & Skitmore 1991). Clearing in areas of naturally highly saline soils of the upper catchment has resulted in rising salinity levels of the river and in the waters of the upper
Blackwood tributaries (Arthur & Beaufort Rivers) becoming saline to brackish (Schofield et al.
1988). Water in the middle to lower sections, which are drained by mainly forested areas, is marginal to fresh before becoming influenced by the salt wedge as the river nears the coast. As the riparian zone on the Blackwood River is heavily influenced by intensive livestock grazing, study sites were set up in pairs at each location, with one site that was heavily grazed and, in close proximity, a site that has had little impact of stock grazing. The main study sites on the
Blackwood River included two sites near Nannup (downstream section), one of which had been heavily grazed by livestock so that the composition of the understorey was highly modified
(Site Blb), and the other in a forested area which has had no disturbance from livestock (Site
B1a). A grazed site and an ungrazed site were also located near the Condinup Bridge near Boyup
Brook (middle section) (Site B2a &b). There were another two sites on the Beaufort River
(upstream section), an ungrazed site (B3a) and a grazed site (B3b) (Figure 3.2). Two additional sites were also surveyed on the Blackwood River for the effects of relief from grazing and fire m riparian vegetation (Chapter 6). These sites were located in the middle catchment on a tributary of the Blackwood River (Site B4) and on the Blackwood River 5km west of the town of
Boyup Brook (Site B5) (Figure 3.2). 44
Climate
Ord River
The climatic region of the Ord River catchment (between latitudes 14° and 19° S) is described as tropical with a short rainy season (Gentilli 1972). Climatic data for this East Kimberley region was collected from Kununurra Research Station in the north (records from 1944 - 1996) and from
Halls Creek Post Office in the south (records from 1890 - 1996)(Bureau of Meteorology). This region has a warm dry monsoonal climate characterised by a rainy season from November to
April and a period of virtual drought for the remainder of the year. Gentilli (1972) identified five periods in the seasonal cycle: the cool dry season (May to mid-August); the warm dusty season (mid-August and September); the hot dry-wet transition (October to November); the wet season (December to March); and the hot wet-dry transition (April). There is a marked trend in climate from the coastal region in the north to the inland area in the south. In the northern portion there is generally higher total rainfall, less extremes of temperature and greater humidity than in the south. Mean mmual rainfall ranges from 784 mm at Kununurra to 491 mm at
Halls Creek, with a rapid onset of rain in the wet season from November and a rapid decline in rainfall after March (Figure 3.3a&b). Rainfall intensity and reliability also decreases to the south. The period of effective rainfall or 'growing season' (i.e. when rainfall exceeds evaporation rates) is around eight months for Kununurra and substantially less for Halls Creek.
Highest monthly median rainfall is for January and February (170 mm for both months in
Kununurra and 113 mm and 88 mm respectively in Halls Creek) with 0 mm median rainfall for
June to September for both sites. Wet season rainfall forms by far the greatest proportion of the annual total with 80-82% of rainfall occurring in the months of December through to March.
There are an average of 65 and 51 rain days per year for Kununurra and Halls Creek respectively, with the months of January and February having the highest number of days of rain (14 for both months in Kununurra and 11 and 10 in Halls Creek). For the months of May to September, there is, on average, less than one rain day per month for both Kununurra and Halls Creek. Average rainfall per rain day for the period December to March is 17.5 mm at Kununurra and 9.6 mm at
Halls Creek. Annuall,"ainfall variability is moderate but increases as rainfall decreases, with 45 deviation from the mean being 21% in Kununurra and 31% in Halls Creek (Slatyer 1970). Due to the proximity to the coast, temperature extremes and diurnal variations in temperature are ameliorated in Kummurra compared with Halls Creek. . The hottest month for both sites is
November, with an average temperature of 38°C and with, on average, 10 days above 40° C. The onset of the wet season has the effect of slightly reducing maximum temperatures (Figure
3.3a&b). July is the coolest month, with average maximum temperatures of 30.3°C and 26.8°C for
Kununurra and Halls Creek respectively. Average minimum temperatures range from 24.7°C in
November to 14.4°C in July in Kununurra and 24.2°C to 8.9°C in Halls Creek. Humidity is lower in
Halls Creek (53% in February to 21% in July, annual mean 40%) than in Kununurra (70% in
February to 32% in August, annual mean 47%), due mainly to geographic location and lower rainfall. Humidity levels are in general lower than would be expected for the humid tropics
(Slatyer 1970). Annual evaporation rates are similar at both sites at around 2600 mm, with highest evaporation rates occurring in the pre-wet season period between September and
November.
Blackwood Rivet·
The regime of prevailing winds exerts a strong influence on the climate of the Blackwood River region, with the sub-tropical belt of high pressure (as anticyclones) extending across the region during summer bringing easterly prevailing winds. In autumn, these gradually move northwards and incursions of cyclonic low pressure zones bring most of the winter rains (Gentilli 1972).
Frequency of moisture laden unstable squalls, which characterise winter storms, reduces with distance in an easterly direction from the coast. The strongest climatic features of the south west region are abundant winter rains and intense summer droughts, which are characteristic of the 'Mediterranean' climate (Gentilli 1972). Climatic data for this region was collected from
Bridgetown Post Office (records from 1887- 1996) in the central region and Wagin Post Office
(records from 1891 - 1996) at the eastern headwaters of the Blackwood River (Bureau of
Meteorology) (see Figure 3.2). There is a marked trend in climate from the coastal region in the west to the inland area in the east. In the west there is higher annual rainfall, less extremes of temperature and greater humidity than in the east. The period of effective rainfall or 'growing 46
season' (i.e. when rainfall exceeds evaporation rates) is around six months in the east and eight
months closer to the coast (Gentilli 1972). Mean annual rainfall ranges from 836 mm in
Bridgetown to 440 mm in Wagin. Mean monthly rainfall t_ogether with maximum and minimum
temperature data are summarised for these sites in Figure 3.3c&d. Highest monthly median
rainfall is for July (144 mm in Bridgetown and 66 mm in Wagin), with lowest values of 6.4 mm and
2.8 mm for January at Bridgetown and Wagin respectively. Winter is the period of greatest
rainfall, with 45% of the annual total falling from May to August at Bridgetown and 59% at
Wagin. There are an average of 140 rain days per year in Bridgetown and 93 per year in Wagin,
with July having the most number of days of rain (22 in Bridgetown and 14 in Wagin). For
January there are on average 3.3 rain days in Bridgetown and 2.3 rain days in Wagin. Average
rainfall per rain day for the period May to August is 7 mm in Bridgetown and 5 mm in Wagin and
for December to March is 4 mm and 4.5 mm respectively. Annual rainfall variability is
moderate, with deviation from the mean of 14% in Bridgetown and 15% in Wagin. The hottest
month for both sites is January, with an average temperature of 29.7°C in Bridgetown and 31 oc in
Wagin. July is the coolest month, with average maximum temperatures of 15.7°C and 15.2°C for
Bridgetown and Wagin respectively. Average minimum temperatures range from 12°C in January
to 4.4°C in July for Bridgetown and 14.6°C in February to 5.SOC in August for Wagin. Humidity is
lower in Wagin (80% in June and July to 54% in December and January, annual mean 70%) than in
Bridgetown (91% in June and July to 60% in December and January, annual mean 75%), due mainly
to geographic location and lower rainfall. Annual evaporation rates at Wagin and Bridgetown
are between 1600 and 1800 mm yr-1 with highest evaporation rates occurring in summer between
December and March.
Comparison of climates
The difference in the climate of the two regions that most affects fluvial processes is rainfall
intensity. The principal causes of rain in the Ord River region are instability thunderstorms,
monsoonal incursions and tropical thunderstorms (Gentilli 1972). These give rise to a large
proportion of high intensity rainfall events in the wet season. For the Blackwood River region, winter rains arise from the advection of moist air in the westerly air streams in the form of rain- 47
bearing depressions in cyclonic low pressure systems. Rainfall throughout the year tends to be less intense and to occur over a longer period. For example, at Kununurra there is an average of 64 raindays per year (i.e. 11.9 mm rain per rain day), with 14 .rain days in the wettest month of
January (i.e. 13.6 mm rain per rain day). This compares with an average of 140 rain days for
Bridgetown (i.e. 5.9 mm rain per rain day), with 22 rain days in July (i.e. 6.9 mm rain per rain day). Kununurra receives, on average, 80% of its rainfall in the four wettest months, from
December to March, whereas Bridgetown receives only 45% of its rainfall from May to August.
Similar trends are seen for the drier areas of each region. When mean monthly rainfall is regressed against mean number of rain days per month, the pattern of greater intensity of rainfall events on the Ord River is clearly evident (Figure 3.4). The slope of the regression lines indicate that rainfall intensity is less in the high rainfall months in the Blackwood region, whereas intensity is fairly even throughout the year, if not slightly more in the high rainfall months, for the Ord. This large difference in rainfall intensity of the two regions has obvious implications for river flows and fluvial processes with, for example, high intensity rainfall resulting in greater runoff into drainage channels. Other aspects of climate that are contrasting in the two regions, such as temperature, humidity and evaporation rates would also contribute to shaping the nature of the riparian vegetation on these two rivers. 48
200 40 0 -E 35 - E 150 30 -f 25 .....::I - (U (U 100 20 a.. Q) -·a;s::: 15 c.. E a:: 50 to Q) 5 1- 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
200 40 - ...... -(> ...... A b) Halls Creek .... <>·······<>·······<> 0 v········<> ()···· 35 ~ -E ········o...... o········· E 150 .... -o-...... o···· 30 f 25 ::I - ~ (U 100 20 a.. -·a;s::: 15 ~ a:: 50 10 Q)E 5 1- 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
200 c) Bridgetown 40- 35 ~ 150 30 ~ 25 ::s ~ -m 1oo 20 a.. s::: 15 ~ -·a; 10 a:: 50 E 5 ~ 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
200 d) Wagin 40- -E 35 ~ E 150 30 ~ - 25 .a .! 100 20 ~ s::: «< 15 ~ 10 a:: 50 EQ) 5 1- 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
r-:::1 1::.::.1 Rainfall ...... <>...... Maximum ----0---- Minimum Temperature Temperature
Figure 3.3: Mean monthly rainfall and maximum and minimum daily temperatures recorded at weather stations on the Ord River (a) & (b) and the Blackwood River (c) & (d). 49 a)
25 y= 0.126x + 2.970
20
<> y =0.074x + 0.600 r 2 =0.976 [?"' 15 '"0 ·ea~ <> "' ""'0 0 10 z D Kununurra
<> Bridgetown 5
0 0 ...... 0
Rainfall (mm)
b)
20 y = 0.081x + 0.935 r 2 = 0.932
rJl 15 [? '"0 D Halls Creek '@~ ""'0 10 <> Wagin z0
5
Rainfall (mm)
Figure 3.4: Relationship between rainfall and number of raindays for (a) higher rainfall sites on the Ord River (Kununurra, 784 mm/yr) and the Blackwood River (Bridgetown, 835 mm/yr) and (b) low rainfall sites on the Ord River (Halls Creek, 490 mm/ yr) and the Blackwood River (Wagin, 440 mm/yr). 50
Hydrology
The Ord River rises in the sandstone country of the Durack Range at an altitude of around 400
metres above sea level, gradually sloping down to around 150 metres in the middle catchment at
Spring Creek and to about 50 metres at Carlton Gorge and onto the floodplain downstream (see
Figure 3.1). The Ord is considered a degraded river (Williams et al. 1996), with cattle grazing
in the upper reaches causing extensive erosion and two dams on the lower section resulting in an
altered flow regime downstream. Hydrological statistics quoted below are for the unregulated
section upstream of Lake Argyle.
Many westward draining rivers in south western Australia, such as the Blackwood River, have
an unusual drainage pattern. The Blackwood River rises in terrain of little relief (~250 m asl at
1 Wagin) and with low annual rainfall (<500 mm yr- ), is 'rejuvenated' in the middle reaches as it
flows through an area of greater relief down an escarpment and slows again as it meanders across
the coastal plain to the sea. The upper catchment is affected by an area of north-south oriented
paleo-drainage which is marked by a system of saline lakes. Extensive agricultural clearing in
this area has resulted in rising saline water tables and therefore secondary salinisation
(Malcolm 1983). These two factors contribute to high salt loads, particularly in the upper reaches of the river. The Blackwood River appears to have a reverse salinity profile, with high salinity in the headwaters, especially in early winter, and fresher middle reaches which drain higher rainfall forested areas (Morrissy 1974).
Monthly river discharge records over an 18 year period from 1975-93 for the Ord and Blackwood
Rivers show the much greater discharge for the Ord River (Figure 3.5). The variable nature of flows in terms of magnitude are clearly seen over this period, particularly for the Ord River. On the Ord River, flows are generally restricted to the wet season (December - March), with significant flows being recorded outside this period only once. During the period of records there have been three flows. of greater than 10,000 m3 sec·1 and two years with very little flow 51
recorded. Blackwood River discharge is much lower, with the highest levels in winter (May-
August) and with some flow continuing for most of the year. Greatest discharge on the
3 1 Blackwood River for the period of records (1187 m sec- ) wa~ in summer 1982 when a cyclonic low pressure system drifted down from the north dropping a large amount of rain and resulting in flash flooding.
The degree of variability of flows of the Ord and Blackwood Rivers is given in Table 3.2 along with a comparison of flows with other rivers throughout the world in different climatic zones.
Compared with these other rivers, the Ord River is highly variable in terms of monthly discharge, with only the arid zone rivers of Australia having a higher coefficient of variation
(CV), skew and spread. The Blackwood :River is also quite variable, with values similar to the
Murray and Darling Rivers in eastern Australia.
Table 3.2: Summary statistics of Australian and world rivers from different climate zones 3 1 showing mean monthly discharges (m s- ) over periods of 15- 25 years (modified from (Walker et al. 1995).
River Cotmtry Climate Min. Max. Mean CV Skew Median Spread Ord Australia Trop,D 0 1285 53 2.86 5.16 1.42 8.87 Blackwood Australia W/Temp,SA 0.1 286 18 1.47 3.15 2.73 3.24 CooperCk. Australia W/Temp,A 0 7593 142 4.54 8.51 6 4.3 Diamantina Australia W/Temp,A 0 5955 101 4.66 9.04 1 40.66 Darling Australia W/Temp,SA 0 764 97 1.27 2.13 47 2.7 Murray Australia W/Temp,SA 6 5710 975 1.06 2.06 646 1.48 Mekong Thailand Trop, D 1006 30582 8394 0.94 0.99 4390 2.87 Niger Mali Trol}' D 30 2658 1161 0.71 0.06 1158 1.4 Danube Romania W/ ,C 1370 13300 5121 0.4 0.74 4920 0.6 1 Trop=tropical; D= dry; W= warm; Temp= temperate; SA=semi-arid; A=arid 2 CV = coefficient of variation. 3 Spread = ratio of interquartile range to the median.
Comparison of the hydrographs for 1993/94 for the Ord and Blackwood Rivers highlights the differences in volume and shape (Figure 3.6). The much greater discharge for the Ord River is related to a combination of factors including rainfall intensity, topography I slope and geology and soils, with rainfall intensity being the greatest influence as discussed above. Figure 3.6 also 15000 (a) Ord River 24455 a_Cl) :t..,...m • 10000 .c UJ OM .~ E C- 5000
1 0 1"I' 0' oA • •I • • I• • • • • '6{ .. 1• • • • • {,i 1l>l 1 ~" • fi1 ,', h'a 11 • h 1l>ln oo • o • • o 1"" o• o o {, '' •.ln"" 6i •oh a• • a(, • '' ol •• • ... ~," •o o • 1 J, 1,} .. o' 11 o/,1 oI o 11 o1 a{o ,I 1 oaa nJ.l9,{,1, o11 {,i r)=i;," oo, 1,1 olo moo 1 o•'il• 11", nli i, I CD r-.. 0'1 0 ..- N (") ' 500 (b) Blackwood River \\11 87 400 I Cl) tn ll-et! 'I"""- 300 .c'0 U) UJM ·-c_E 200 100 0 !.0 CD r-.. CO 0'1 0 ..... C'\1 (") ' Figure 3.5: Monthly river discharge data over a 18 year period for (a) the Ord River and (b) the Blackwood River. CJl N Ord River daily discharge 130000 Blackwood River daily discharge 8000 .....- •Cl) 120000 • M ...... • 7000 E -I .."' Cl) 110000 .. Cl) M ·: -0) . :I.. E 100000 ,.i . ea ,... 6000 .c Cl) .. - •' . () 0) •' Cl) :1.. 90000 •' .. ea ..•' '•. "'0 .c ' .' . () 80000 .'. 5000 (5 Cl) '. .' ''' .' 0 .' ' . '"0 70000 .' .... ' . - ... ~- ea !!~ .' 4000 0 60000 . ·ea ' ·' . '"0 - .' '' . ~ : ..,p 3 . :I.. - 50000 ' . . Cl) '(ij ' . 3000 .' ' . .::: "C . . . :I.. 40000 ' a: ' ' . .~ Cl) ' . ' ' ." . ... '"0 > •,' ..... ·.,: ·•. 2000 0 a: 30000 . ... 0 "C (1\).: .. :I.. ' . . ~ 20000 () ' ' .._ ... '> 0 ... '. 1000 ea . <), q . q"'~ . ..s . 10000 ...... aJ 0~- '~•····· 0 M M M .::r .::r .::r .::r .::r .::r .::r .::r <:t .::r <:t <:t <:t <:t '<:!' <:t <:t .::r <:t <:t <:t <:t -:!' '<:!' en en en en en en m m m en m m en en m en en en m m m m m m m m m I I I ,_ I I I I I u• ' • ' ... .!. lj') lj')' • • u u' c c .c' .c ... ' >- >-' c' c' c' :I' ::11 ' c.' c. u u' > > > Cl) Cl) Q) m m Cl) Cl) m m c. c. m m ::11 ::11 :I ..., :I ::11 Cl) Q) 0 0 0 c c c -::11 -::11 1.1.. 1.1.. :E ::E <( <( ., -::11 -:I -:I <( <( 00 00 0 -0 -I z z I I :E :::: (") • z I .,.. l.t)• C\1• CD' 0• • • l.t) ' !'-- .,.. l.t) m' • en .,.. m' M' en ' M' CD' <:t cc .....' ' m' .,.. C\1 0' .::r' !'--' ' C\1' CD 0 .,.. C\1 C\1 C\1 C\1 C\1 .,... .,... C\1 .... C\1 C\1 ...... M Figure 3.6: Daily total river discharge for the Ord River (left hand axis) and the Blackwood River (right hand axis) for 1993/94 highlighting the differences between volume and the shape of the hydrographs. Ul (JJ 54 highlights the influence of groundwater flow on river base flows. There is negligible groundwater influence and therefore base flow on the Ord River, with large steep sloped peaks in flow. In contrast, there is a large base flow on the Blackwood River, with a much gentler slope m both the rising and falling limb of the hydrograph, so that groundwater inflows become the greatest influence on discharge. Discharge figures for the Ord River were taken from above Lake Argyle at Old Ord River Station and cover a catchment area similar to the area of the Blackwood catchment from just below the town of Nannup (Darradup ). These results highlight the large 'flashy' nature of flows on the Ord River, with large flows occurring in quick response to rainfall events but quickly returning to very low flows. Flows on the Blackwood River are usually maintained throughout the rainy season and into the drier late spring/ early summer period. These features of the hydrology have implications for the adaptation of the riparian vegetation. Geology Ord River The geology of the East Kimberley region is represented by rocks from almost all geologic periods, the oldest including the granite country of the Bow River Hills that were formed in the pre-Cambrian period, some 1900 million years ago (Department of Mines 1972). The geologic formations that are associated with the Ord River basin include the area of low hills of metamorphic rocks to the west of the Ord River from which the Bow River, a major tributary of the Ord River, rises. To the west of this area are the Durack Ranges of mainly sandstone sedimentary rugged ridges and escarpment which formed in the Tertiary Period where the Ord River and another major tributary, the Dunham River, have their headwaters. This area also forms the divide between the Ord and Fitzroy Rivers. To the east and south of Lake Argyle, in the area around Ord River Station, is the Ord River plain. This extensive dissected plateau consists of basalt lava flows 550 million years old which were subsequently covered by shallow seas resulting in sedimentary deposits of limestone, sandstone and mudstone. This plateau 55 country was uplifted some 20 million years ago and has been extensively incised by creeks and rivers. The Halls Creek fault is a major fracture which extends from the south of the Ord basin to the north into the Northern Territory and has caused horizontal and vertical movements of the landscape. The escarpment along the fault line is characterised by cliffs or higher hills of more resistant rock on one side of the fault or by resistant quartz veins (Stewart et al. 1970). The flat country of the lower Ord River as it flows into Cambridge Gulf is made up of Quaternary alluvial sediment plains. Blackwood River Of the south-west province of Western Australia, around 95% is composed of granite-migmatite terrain of Archean age known as the Yilgarn Block (Department of Mines 1972). Small, irregular, patchy areas of metamorphic gneiss rocks, with indistinct boundaries, are scattered throughout the province. The regional tectonic trend is mainly north-westerly and an active seismic zone with a north-west trend across the province is the Meckering line (Mulcahy & Bettenay 1972). This fault line marks the boundary between the westerly limit of wide alluviated valleys of ancient drainage containing salt lake systems and rejuvenated incised river valleys characteristic of the western areas. Archean rocks are poorly exposed east of the Meckering line and are commonly deeply weathered and covered by thick sandy superficial deposits. To the west, fresh rock has been exposed along the sides and floors of the incised drainage systems. The Archean rocks are intruded by mafic dykes of altered quartz dolerite which also trend north westerly and increase in frequency to the west near the Darling Scarp (Department of Mines 1972). There are small areas of shallow, unconsolidated sediments throughout the area, usually in the upper part of the landscape. The lower catchment on the coastal plain, below the Darling escarpment, is formed of depositional materials of either fluvial or aeolian activity. This includes the Blackwood Plateau of flat-lying Mesozoic sandstones overlain with lateritic remnants. To the west is the Leeuwin Block of Pre-Cambrian gneiss and granulites with overlying sedimentary deposits of limestone and Pleistocene and Holocene aeolian sands (Department of Mines 1972). 56 Geomorphology and Soils Ord River The Ord River area lies in a relatively small range of altitude, with most of the area less than 300 metres above sea level and with only small ridges in the western part more than 600 metres above sea level. A significant proportion of the area consists of rugged rock outcrops with shallow stony soils. The remaining area is undulating to flat with gently undulating plateau surfaces, erosional plains with moderately deep soils formed from weathering of the underlying rock and alluvial plains along the major stream valleys (Stewart et al. 1970). The gently undulating erosional plains in the upper catchment are formed on Cambrian shale and limestone. The area has a clearly defined dendritic drainage pattern and is interspersed with occasional volcanic mesas and buttes of eroded basalt material. The soils consist of brown powdery calcareous loams over deep clay loams on the slopes, with brown cracking clays in depressions and along drainage lines (de Salis 1993). There are also smaller areas of gently undulating country with deep yellow sands over sandstone. This sandstone parent material is particularly resistant to erosion and is strongly jointed which has a significant effect on drainage. To the north, towards Lake Argyle, is more hilly country of basalt rock outcrops derived from Cambrian volcanics and with pockets of red clayey soils. There are also areas of volcanic structural benches and erosional plains on moderate slopes, with brown loams merging to red clays and brown skeletal soils around the basalt hills and structural benches. Large areas consist of gently undulating erosional plains of calcareous, dolomitic and shale sediments, with soils of brown and grey cracking clays (black soil plains) (Stewart et al. 1970). The Carr-Boyd Ranges between Lake Argyle and the Dunham River form rugged hilly country up to 600 m high of Proterozoic sandstone, siltstone and shales. The soils are shallow and stony with sands and loams on the ridges and duplex soils on the slopes. Most of the drainage is superimposed but is modified to an irregular drainage pattern by the underlying bedrock. To the west of the Ord River plain and forming part of the Halls Creek fault system is the area of the Bow River Hills, consisting of pre-Cambrian granite boulder-strewn hills and granite ridges with open textured 57 dendritic drainage patterns. Soils are mainly shallow stony sand with much rock outcrop m ridges and upper slopes and with duplex soils on lower slopes and in the valleys. The lower Ord River, below Lake Argyle, is predominantly gently sloping· fine textured (black soil) fluvial plains on Quaternary alluvia. This is the flood plain of the Ord River and it is interspersed with broadly meandering channels with broad levees associated with brown sand or sandy loam soils over permeable reddish brown subsoil. Minor areas of sandy country consisting of deep red or yellow sand over Permian sandstone are scattered throughout this area, together with some areas of rock outcrop and skeletal soils of sandstone and shales. Blackwood River By the end of the Tertiary Period the Archean shield of the great plateau in the south-west drainage division of Western Australia was mantled by a lateritic soil which developed when climatic conditions were wetter and warmer than the present day climate (McArthur 1991). The laterite consists of a sandy surface and gravelly clay sub-soil underlain by weathered mottled and pallid zone materials and saprolite above basement rock. The original lateritic profile has had a strong influence on the make-up of the present day soils. Today the lateritic profile is rarely found intact (Bettenay et al. 1980) and only remains as remnants, with stripping of the lateritic mantle being more extensive with distance to the east. The soils reflect varying degrees of erosion, colluviation, sorting and leaching, depending on factors such as slope, aspect and landscape position. Erosion has exposed the mottled and pallid zones and in some cases has led to the deposition of erosional material in the valley bottoms. The original sandy surface soils can also be replaced by gravelly sand or sandy loam so that the resultant soil can have a colluvial surface layer and underlying soils of a quite different origin. The Blackwood River basin can be divided into the upper, middle and lower catchments on the basis of landform, topography, vegetation and land use (Grein 1995). The upper catchment is a zone of ancient drainage of gently undulating plateaux, between 280 m and 400 m above sea level. It is marked by a series of saline lakes and only flows (and therefore only contributes to the Blackwood River discharge) in years of high rainfall. The broad valleys have very low grades, 58 with local relief of less than 50 m, and contain deep grey to yellow sandplain soils with some laterite remnants on the divides and with loose grey gravelly soils. Shallow duplex soils of the stripped laterite profile and grey to red-brown loams formed on fresh rock can also occur. The zone of rejuvenated drainage west of the Meckering fault line contains clearly defined drainage lines that carry water each year. Most of the landscape is formed of the mottled and pallid zones of remnants of the laterite profile and contains undulating hills, with sandy duplex soils en the slopes and in the broad valley floors, where the duplex soils are underlain with sodic, alkaline clay sub-soils (Northcote et al. 1967). There are also numerous areas of exposed basement rock and laterite breakaways on the ridges with sandy gravel soils. The middle catchment consists of undulating dissected terrain between 260 m and 360 m above sea level, with large remnants of the lateritic plateau consisting of yellow-brown and grey sandy gravel soils. There are also broad flat upland areas of Eocene sedimentary deposits separated by valleys dissected into the underlying bedrock. Valley soils include yellow-brown loamy and sandy gravels and yellow-red loamy duplex soils grading into clays. Further west from Bridgetown, where the Darling Scarp rises from the coastal plain, is a broad undulating laterite plateau where sandy and loamy gravel soils are dominant. The Blackwood River is deeply incised into the plateau, forming a deep-sided valley and exposing fresh rock. Red-brown and yellow-brown loams grading into clays have formed on the fresh rock. Along valley floors river flats contain soils of deep loose brown loams and deep red loams. West of the Darling Scarp is level, slightly undulating terrain formed on Cretaceous and Jurassic Perth Basin sedimentary rocks with elevations of 20 m to 180 m above sea level. This is dissected by the river to form shallow valleys with grey-brown sands and loams. These merge to the west with the low-lying coastal plain of deep loose white sands on Quaternary sediments (Tille & Lantzke 1990). 59 Vegetation Ord River The vegetation of most of the Ord River basin is open woodland or savanna with a grass understorey. In much of the area the composition, density and structure of the understorey has been greatly altered by over 100 years of cattle grazing, the introduction of exotic species and the effects of soil erosion. In severely degraded areas, such as the calcareous soils of the Ord River plains, a rehabilitation project has taken place to control soil erosion and to reduce sediment loads into the river (Fitzgerald 1967). This has seen the de-stocking of large areas and re vegetation with exotic species such as Cenchrus ciliaris (buffel grass), C. setiger (birdwood grass) and Aerva javanica (kapok bush). The vegetation described here is in the undegraded condition (Beard 1979) and species names have been updated where appropriate from Wheeler et al. (1992) and the State Herbarium records. In the upper catchment on the Ord River plain, just south of Lake Argyle, the vegetation is a low tree savanna of Eucalyptus opaca and E. argillacea with an understorey of 'Tippera' tall-grass association and dominant grasses including Themeda triandra, Sehima nervosum, Sorghum plumosum, Chrysopogon fallax, Aristida pruinosa and Heteropogon contortus. In the southern portion trees become more sparse and restricted to lower slopes and flats and include Lysiphyllum cunninghamii, Terminalia oblongata and T. arostrata. T. platyphylla and other Terminalia spp form a fringing woodland along drainage lines. The ground layer changes to a short-grass (<30 an tall) association of Enneapogon spp. and Aristida contorta. On the rocky hills Eucalyptus brevifolia and some E. opaca and Terminalia spp. with a ground layer .of Triodia spp. (spinifex) can be found. Large areas of the plain are treeless, with Astrebla spp. (Mitchell grasses) dominant on the areas of black soil and short grasses such as Enneapogon spp. most abundant m the calcareous soils. On the levee frontages of the major streams E. gymnoteles and Lysiphyllum cunninghamii form sparse woodlands, while on the river banks themselves fringing forests of Eucalyptus camaldulensis, E. gymnoteles, Melaleuca leucadendra and Ficus sp. can be found (Perry 1970). 60 On the lower Ord River where sandy soils overlay sandstone, the vegetation comprises a high grass savanna woodland of annual sorghum (Sorghum spp.) and Plectrachne pungens with an overstorey of Eucalyptus tetrodonta and E. miniata. On the rocky limestone areas the overstorey changes to open woodland of Terminalia spp., Adansonia gregorii, Lysiphyllum cunninghamii and Ficus spp. with a similar understorey. The alluvial plains, other than black soil areas, support a savanna of Eucalyptus tectijica and E. byrnesii with some areas of Melaleuca viridiflora and M. nervosa and an understorey of tall grasses such as Sorghum spp.. The black soil plains are sparsely wooded with Eucalyptus microtheca, Terminalia spp. and Lysiphyllum cunninghamii, common shrubs Acacia farnesiana and Carissa lanceolata and a ground layer of blue-grass and tall-grass community including Dichanthium spp. and Sorghum plumosum. Frontage woodlands on the levees of the major stream channels consist of Eucalyptus houseana, E. polycarpa, E. patellaris, and Adansonia gregorii with a tall-grass understorey of Sorghum stipoideum, Chrysopogon latifolius and Panicum spp.. Fringing forests on the river banks consist of mixed forest with dominants including Eucalyptus camaldulensis, Melaleuca leucadendra, M. argentea, Pandanus spiralis Terminalia platyphylla, Lophostemon grandiflorus, Nauclea orientalis, Barringtonia acutangula and Ficus spp .. Blackwood River Most of the south-west of Western Australia, including the Blackwood catchment, is within the South-West Botanical Province which contains the Roe, Avon (upper catchment) and Darling (middle and lower catchment) Botanical Districts (Beard 1981). Species names have been updated from State Herbarium records where appropriate. The majority of the catchment is composed of eucalypt woodlands or forest with a mixed shrub understorey and with the most common families being Myrtaceae, Proteaceae and Papilionaceae. The vegetation composition and density is strongly influenced by landforms, soils and climate. The eastern extent of the catchment is a mosaic of shrub-heath on sandplains, mallee on the slopes and woodland in the valley floors. The dominant mallee species are Eucalyptus eremophila, E. angustissima and E. transcontinentalis with. a shrub understorey dominated by Melaleuca spp.. Low woodlands are 61 chiefly composed of E. gardneri and E. falcata with E. celastroides on laterite ridges and a sparse mixed shrub understorey including Melaleuca spp. and Acacia spp.. There can also be a mid storey of Acacia acuminata, A. microbotyra, Allocasuarina· heugeliana and Hakea preissii. On plateau areas with heavy soils which may be subject to waterlogging there is an association of E. wandoo and E. gardneri. Further west of this area there is heath vegetation (dominated by Dryandra spp) on laterite residuals, while in undulating country there are woodlands of E. wandoo on the upper slopes, E. loxophleba on the mid-slopes and E. salmonophloia and E. longicornis on the heavier soil flats. Along the major creeks E. rudis is the dominant tree with Callistemon phoeniceus forming a major part of the understorey. Where creeks are salty Casuarina obesa and Melaleuca hamulosa are common with a shrubby understorey of salt tolerant species including Halosarcia lepidosperma and Sarcocornia blackiana. In the middle catchment the principal elements include woodlands of Eucalyptus wandoo on laterite residuals, with E. wandoo and E. loxophleba on undulating country and E. loxophleba and E. occidentalis m sandy patches. E. astringens and E. marginata are found on laterite breakaways, with E. rudis and Melaleuca rhaphiophylla occuring along creek lines. Further west and south, open forest of E. marginata occurs on the laterite hill tops, with E. wandoo woodlands on the lower slopes and scattered Corymbia calophylla in both areas. On the poorer drained soils E. occidentalis occurs with E. rudis on the drainage lines. In the lower catchment, particularly along the Darling scarp, are large areas of E. marginata and C. calophylla forest where laterite or ironstone gravels are present. Along riverbanks E. rudis, with M. rhaphiophylla as a mid-storey, forms a fringing forest. On sandier soils closer to the coast are areas of E. marginata and Banksia spp. woodlands, with low woodlands of Banksia spp. and Melaleuca preissiana where drainage is poor. Land Use Ord River The East Kimberley region and the area around the Ord River were settled in the 1880's when cattlemen brought stock from Queensland, with some of the earliest leases taken up around the 62 Ord River plains at Argyle Downs and Ord River Station (Stewart 1970). Cattle numbers expanded rapidly after introduction because of the abundant water and productive pasture, with cattle becoming concentrated on the river frontage where water was readily available. The grasslands, which had evolved through selective grazing by marsupials, could not withstand the greatly increased grazing pressure associated with open range production of hard-hoofed cattle (Fitzgerald 1967). By the 1930's large areas of land were severely degraded and eroded by overgrazing and this was exacerbated by the effects of drought, fire and feral animals (Fitzgerald 1967). Bare soil surfaces were exposed to the effects of wind and water erosion, particularly the friable calcareous soils, contributing enormous sediment loads to the Ord River. With the contemplation of the building of a dam on the river for irrigation, concerns were raised about the effect of high sediment loads on the storage capacity of the reservoir and a rehabilitation programme on the Ord River plains was instigated in the early 1960's (Fitzgerald 1967). Land was resumed by the government, stock removed and a revegetation plan implemented. This regeneration programme has seen the establishment in many areas of the introduced grasses Cenchrus ciliaris and C. setiger and the shrub Aerva javanica, but many areas are still without vegetation cover (de Salis 1993). At present cattle grazing for beef production continues throughout the Ord River catchment, except in the area of Purnululu National Park in the south-west. The Ord river is also affected by the construction of two dams along its length. The diversion dam was constructed in the early 1960s near the town of Kununurra as a pilot dam to test the viability of an irrigated agricultural industry on the lower Ord floodplain. This flooded the section of river above the dam to Carlton Gorge some 45 km upstream, turning it into a reservoir and drowning large areas of riparian vegetation. The river below the dam now runs year round at a fairly constant level, with some input of floodwaters from the Dunham River in the wet season. This area contains the Ord River irrigation scheme, where 15,000 hectares of land on the Ivanhoe plains are irrigated for agriculture including sugar cane, fruit and vegetable production. In 1972, a large dam was constructed at Carlton Gorge to allow the envisaged expansion of the 63 irrigation scheme. This resulted in the flooding of the lower, northern section of the Ord plains to form Lake Argyle (Water & Rivers Commission 1997). Blackwood River The eastern area of the Blackwood catchment in the Narrogin and Wagin districts was sparsely settled in the late 1890's with the main activity being livestock grazing on the native vegetation. Greater settlement and land clearing for agriculture occurred after 1890 with the building of the railway (McArthur 1991). With further settlement, the expansion of sheep production and the use of fertilisers to improve nutrient deficient soils to allow cereal cropping, the area was gradually extensively cleared of native vegetation so that in many areas <10% of the native vegetation remains. The area has extensive, highly saline soils so that secondary salinity, particularly in the lower part of the landscape, is a major problem. Many parts of the _ middle catchment were settled in the 1880's where the first commercial enterprise was native sandalwood (Santalum spicatum) harvesting. The eastern part of the middle catchment around Boyup Brook has been extensively cleared for improved pasture, mainly for sheep production and with some cropping for fodder. ·Recently, there has been an increase in plantation forestry m cleared farmland, with Eucalyptus globulus for paper pulp and Pinus pinaster for timber. Much of the land consists of nutrient poor soils which require fertilising to maintain productivity. This region is also susceptible to waterlogging and secondary salinisation, through rising groundwater, in the lower sections of the landscape. To the west, the main activity is timber harvesting in the native E. marginata (Jarrah) forest. Native vegetation on the fertile soils of the Blackwood valley has been extensively cleared for plantations of Pinus radiata (for timber production) as well as for livestock grazing and small areas of orchards. The coastal plain of the lower catchment was settled in the 1830's around Augusta at the mouth of the Blackwood River. Large scale clearing for agriculture took place in the 1920's and was stepped up in the 1950's. This was mainly for improved pasture for dairy and beef cattle with soils regularly treated with fertilisers to correct nutrient deficiencies. There was some timber harvesting in earlier years of karri (E. diversicolor) around Karridale and some planting of Pinus spp. plantations. 64 CHAPTER4 DESCRIPTION OF RIPARIAN VEGETATION WITH REGARD TO FLORISTICS AND LIFE FORM Introduction Plant communities in the riparian zone are highly productive and can be more species rich than adjacent upland sites (Gregory et al. 1991). Indeed, riparian vegetation is rarely homogenous and usually reflects the effects of disturbance that has created longitudinal and transverse environmental gradients which have led to the formation of complex vegetation communities and high species diversity (Gregory et al. 1991). The riparian zone has also been considered an ecotone where different plant communities overlap, combining elements of both communities to produce an area of higher species richness (Bretschko 1995). In the riparian zone edaphic and lotic processes can both influence community structure and composition. Vegetation patterns, and therefore species richness and diversity, in the high disturbance riparian zone are strongly influenced by species colonising ability (White 1979). In addition, plant colonisation results from the interactions between existing vegetation, spatial patterns in the environment and morphology and physiology of propagules (Glenn-Lewin & van der Maarel 1992). Therefore, to understand recruitment patterns on the rivers assessed in this study, knowledge of the floristics and structure of the specific vegetation communities that occur is required. There have been m specific studies of the vegetation communities on the Blackwood or Ord rivers, only broad scale assessments which describe the vegetation in a very general sense (Beard 1979; Beard 1981). This Chapter provides a description of the vegetation in the riparian zone at the study sites on the Blackwood and Ord rivers in terms of floristics and the major life forms. 65 Describing the range of different plant life forms that occur is a convenient way of identifying the important functional groups in the riparian zone of ·these rivers. These groups may be important for the ecological functioning of the system (Hobbs 1992). For example, overstorey trees and emergent macrophytes may have a critical role as a functional group through soil stabilisation and nutrient cycling. Functional groups that have been suggested for terrestrial plants have been based on phenology, life form or life history (Armstrong 1993). In Australian grasslands, life form was the most useful characteristic for gauging community response to disturbance (Mclntyre et al. 1995). In the specific environments of the riparian wne particular life forms may be more successful. For example, in highly disturbed environments annual species are favoured due to their fast growth rates and early and prolific seed set (Grime 1979). In contrast, high stress environments (as a result of drought or low nutrients for example) tend to have more perennial species which are generally slow growing, take a long time to reach reproductive maturity and are frugal with resources. Flooding impacts on plants can be described by both disturbance and/ or stress depending on intensity of flow or duration of inundation. Therefore the abundance of the· different life forms on each river can provide an insight into whether disturbance or stress is the most important 'driver' of vegetation community dynamics. Species life history traits, functional groups or vital attributes have been widely used for modelling succession and species composition after disturbance (Brown 1992). Functional grouping of species can also be useful in identifying species tolerance to one or more forms of disturbance, such as flooding and grazing as occurs on some of the Blackwood River sites and all of the Ord River sites. The three main categories of vital attributes used by Noble and Slatyer (1980) included; method of persistence and arrival after disturbance, ability to establish and grow in a developing community and time taken for species to reach critical life stages. This chapter provides an introduction to the floristics of the vegetation found in the riparian zone on the Blackwood and Ord rivers and investigates the environmental factors that affect the floristics between and within sites. It places species in functional groups based on life form in order to develop an insight into which aspects of the vegetation community are important in the 66 establishment of riparian vegetation on these rivers. This enables identification of particular features of the vegetation and the riparian environment that influence regeneration. These features include the effects of disturbance and stress, the fate of propagules and the population dynamics of important species, all of which will be explored in greater detail in the following chapters. Methods Vegetation surveys were completed in July and August 1997 on the Ord River and in October and November 1997 on the Blackwood River. Descriptions of the surveyed sites on both rivers is given in Chapter 3 and location of each site is shown in Figures 3.1 & 3.2. At each site two transects perpendicular to the river were established, transect length varying for each site depending on the width of the riparian zone. Along each transect, five 10 metre x 10 metre plots were established at variable distances located to coincide with the bands of vegetation, giving a total of 10 plots at each site. All plants rooted within each plot were identified and an estimate of percentage cover was made, where cover is the vertical crown or shoot projection per species. Species richness and diversity were measured for each plot. Diversity is derived from species richness (number of species within a plot) and evenness (relative abundance). The index of diversity used here was the Shannon- Weiner Index (H'): H' = Lin Pi X log Pi where p = the proportion of cover or abundance of a species and log is the natural logarithm (Kent & Coker 1992). The Shannon-Weiner index makes the assumption that individuals are sampled from an infinite population. It is based on predicting the likelihood of a particular species being sampled, the higher the value indicating the less certainty of predicting a randomly selected sample and therefore higher diversity. Species evenness (J) can also be calculated from diversity and is an expression of the observed diversity as a proportion of the 67 maximum possible diversity. An evenness of 1 equates with maximum evenness, with all species having the same abundance or cover. A description of the soil profile to a depth of 0.5 m was made for each quadrat and soil samples were taken over a depth of 0 - 10 an for analysis of particle size using the hydrometer method (Page et al. 1986). As there is a gradient of increasing salinity with distance upstream on the Blackwood River, soil electrical conductivity was also measured in each quadrat using a 1:5 soil to de-ionised water solution (Page et al. 1986). Salinity levels on the Ord River are low, and soil salinity was not measured as it was not considered an important environmental attribute that may affect vegetation community composition. A species importance value (IV) was allocated for all species occurring in plots along the transects. This was calculated for each site as IV = relative frequency +relative occurrence+ relative cover. Nomenclature used for plant species names followed Green (1985) and Marchant et al. (1987) for the Blackwood River species and Wheeler et al. (1992) for the Ord River species. Species identification and nomenclature were verified with Western Australian Herbarium records. To identify the functional groups that are important in the riparian zone of these rivers, a 11 plant species were categorised according to life form (i.e. trees, shrubs, perennial herbs, perennial grasses, annual grasses, annual herbs or lianas). Life form groups can, in turn, be categorised by the system proposed by Raunkaier (1934) (see Mueller-Dombois & Ellenberg 1974), based on the height above ground of the perennating organ and on the assumption that species morphology is related to climate (Kent & Coker 1992). Life form groups used here relate to Raunkaier categories thus: trees and shrubs = phanerophytes; perennial herbs = chamaephytes, hemicryptophytes and geophytes; perennial grasses = hemicryptophytes; annual herbs and grasses = therophytes; and lianas = lianas. Multivariate analysis of the floristic and environmental data from the plots using Detrended Correspondence Analysis (DCA), detrending by segments (ter Braak 1987), was performed to provide a summary of ,.plant species composition and abundance. This enabled a quantitative 68 comparison of plots so that floristic differences could be identified. DCA axes are scaled in units of average standard deviation (SD) of species turnover where a 50% change in quadrat composition occurs at around 1 SD (Kent & Coker 1992). DCA is an indirect ordination technique designed to overcome the problems of other types of ordination, namely the arch effect and compression of points at the ends of the first axis (Hill & Gauch 1980). DCA has been criticised as being inelegant and arbitrary and performing poorly with skewed species distributions (van Groenewoud 1992; Palmer 1993). Despite these criticisms, it is considered an effective, powerful and robust indirect ordination technique suitable for ecological data (Peet et al. 1988; Kent & Coker 1992). The ordination resulting from DCA is a product of the variability in species data between plots. The floristic differences between sites can be represented by the relative position of plots on the four floristic axes examined. Sites with similar floristics will be positioned close together. The eigenvalues given for each axis in the DCA analysis provide an estimate of the proportion of the variation in the data explained by each axis so that an axis with a low eigen value is unlikely to be readily interpretable or ecologically meaningful. In this Chapter, only those axes that explain the greatest proportion of the variation in the data are presented in the results and related to the floristics as well as to environmental and life form variables. Environmental variables, floristics and lifeform variables were correlated with the values for the important DCA axes after ordination analysis. This gives an indication which of these attributes are influencing the position of each plot along that particular DCA axis. Vegetation plot variables that were correlated with the DCA axes values included species richness, percentage of exotic species and percentage of each of the different life form groups. Environmental variables correlated with DCA axes values included mean annual rainfall, soil type, height (elevation in metres) above base flow water level and distance (in metres) from low water level. 69 Results Floristic structure Total species richness for the six Blackwood River sites was 156 species, with a mean of 41.2±8.7(SE) per site and ranging from 65 species at Site B2a to 15 species at Site B2b (Table 4.1a). There was a total of 41 families represented across the sites, the most abundant family being Poaceae with 27 species from 21 genera (19 exotic species). Other well represented families included Asteraceae with 12 species in 12 genera (eight exotic species), Cyperaceae with 12 species in seven genera and Myrtaceae with 13 species in six genera. Species richness, diversity and evenness was greatest at the sites undisturbed by grazing. The mean species diversity (H) overall for the Blackwood River sites was 2.2±0.5 and species richness was 40.7±8 but for the ungrazed sites alone species diversity and richness was 2.9 and 56 respectively. Species richness increased with increasing distance from the river for both grazed and ungrazed sites, with a sharp decline in species richness in plots closest to the water's edge (Figure 4.1a). Exotic species were common, particularly at the sites subject to sheep grazing, making up 67% of species and 82% of cover, while at the sites notaffected by grazing exotic species conh·ibuted 33% of species and 36% of cover. The majority of the exotic species were annual herbs and grasses from the families Poaceae and Asteraceae. A full list of species found on both rivers is given in Appendix I. The general structure of the riparian vegetation on the Blackwood River is of an overstorey of almost exclusively Eucalyptus rudis throughout the profile and at all sites. Other eucalypts such as E. wandoo, E. marginata and Corymbia calophylla occurred in areas which are rarely flooded, usually some distance from the river. Casuarina obesa tended to replace E. rudis in the upper reaches, in areas affected by secondary salinisation such as at Sites B3a&b. There is a mid storey of Melaleuca rhaphiophylla fringing the area nearest the water's edge, with other mid storey species occurring further along the profile at higher elevations including Banksia seminuda, Kunzea ericifolia, Jacksonia sternbergiana and Dodonea viscosa. The understorey consists of various Table 4.1a: Summary of floristic data and species origin for the Blackwood River sites. Data are from ten 10m x 10m quadrats on two transects at each site. Site Bla Site Blb Site B2a Site 2b Site B3a Site B3b (grazed) (grazed) (grazed) Species Richness 65 38 68 15 35 26 Species Diversity (H) 3.1 2.57 3.13 1.16 2.38 0.83 Species Eveness (J) 0.74 0.71 0.74 0.43 0.67 0.26 No. spp Cover No. spp Cover No. spp Cover No.spp Cover No.spp Cover No.spp Cover (%) (%) (%) (%) (%) (%) Origin: Native 53 82.3 8 5.0 41 46.0 4 0.7 20 63.0 13 37.1 Exotic 12 21.3 30 68.0 27 43.0 11 43.3 15 44.0 13 47.5 Table 4.1b: Summary of floristic data and species origin for the Ord River sites. Data are from ten lOm x lOm quadrats on two transects at each site. Site 01 Site 02 Site 03 Site 04 Site 05 I Species Richness . 40 22 25 25 30 Species Diversity (H) 1.3 0.63 1.9 1.96 2.01 Species Eveness (J) 0.57 0.37 0.59 0.61 0.59 No. spp Cover No. spp Cover No.spp Cover No. spp Cover No. spp Cover (%) (%) (%) (%) (%) Origin: Native 31 52.0 19 61.6 20 34.16 17 16.9 16 29.0 Exotic 9 46.0 3 0.4 5 0.9 8 18.9 14 52.1 ~ -- '-~- ~· ------ ~ 71 a) Blackwood River Ill Ungrazed sites 35 ...... 0·····•·· Grazed sites -0 30 c. C\1 25 E 0 Lt)... 20 Cl) a. 15 a. c. 2 ...... ~ (f) 10 ...... ···~=:...... n=4 z0 5 2 n=6 0 0-10m 10-30m 30-50m 50-70m Distance, from river channel (m) b) Ord River C\1 E 10 0 Lt) ,.. ~ ...... 2 ~ ...... ~ ?l. 'n=9 ~:,; ...... ~ n=B 5 6 n=17 0-10m 10-30m 30-50m 50-70m >70m Distance from river channel (m) Figure 4.1: Change in mean (±SEM) species richness in 10m x 10m plots with distance from the river for combined sites on a) Blackwood River and b) Ord River. Note change of vertical scale for a & b. n = the number of plots at each distance. 72 a) Blackwood River - Site B2a 8m 6 Kunzea ericifolia ~ Hypocalyma angustifolium . Acacia sa/igna Acacia saligna M. rhaphtophyll~ Lo/ium ridigum 4 . Ehrhata longtfolta Jacksonia furcel/ata ,.-_.. __.,...... Chorizandra enodis Conostylis aculeata \M. rhaphiophylla Patersonia occidentalis 2 Astartea fasicularis Atriplex prostrata 0 20 40 60 80 100 120m b) Ord River- Site 05 E.gymnota/es Cenchrus spp. I E. camaldulensis ______,_,... /M./eucadendra ~ Ficusspp. ~ 1 Cynodon dactylon Cenchrus spp. 1 Crotalaria spp. Cenchrus spp. Aerva javanica Achyranthes aspera M. /eucadendra ~-- Alternanthera nodiflora Cyperus sp. 0~·.------r------~~--~------.------.------.------.- 40 20 0 20 40 60 80m River Channel width -120m Figure 4.2: Schematic diagrams showing the distribution of species at a) Site 2a on the Blackwood River and b) Site 05 on the Ord River. 73 shrub species in less frequently flooded areas such as Melaleuca incana, Hypocalymma angustifolium and Acacia saligna, as well as perennial herbs such as Chorizandra enodis, Carex appressa, Lepidosperma drummondii, Conostylis aculeata and Patersonia occidentalis. Near the water's edge sedges such as Juncus spp. and Lepidosperma spp. were common as well as exotic species such as the annual herb A trip lex prostrata (in the more saline reaches at Sites B2 and B3) and the perennial grass Cynodon dactylon. In areas subject to livestock grazing, the understorey tends to become dominated by exotic grasses such as Ehrharta spp. and Lolium rigidum. A schematic diagram showing the distribution of species along a profile at Site B2a is shown in Figure 4.2a. At the five Ord River sites a total of 98 species were recorded. There were a total of 36 families represented across the sites. The most common families were Poaceae, with 16 species in 12 genera, Cyperaceae with nine species in six genera and Papilionaceae with seven species in 4 genera. A summa1y of floristic data and species origin for the Ord River sites is given in Table 4.1b. Across all sites, mean species richness was 27.8±3.4 and species diversity was 1.9±0.05. Compared with the Blackwood sites these values are relatively low and indicate the tendency of particular species to dominate. As occurred at the Blackwood River sites, species richness increased with increased distance from the river, with a large decline in species in plots closest to the water (Figure 4.1b). Exotic species comprised a mean of 28% of species and 34% of cover with Site 02 (Lake Kununurra) and Site 03 (Behm River) having the lowest proportion of exotic species. Structure of the riparian vegetation at the Ord River sites consisted of a much greater diversity in the overstorey, with common species including Melaleuca leucadendra, Cathormion umbellatum, Barringtonia acutangula and Nauclea orientalis occurring nearest the water's edge and Eucalyptus camaldulensis, Lophostemon grandiflorus and Terminalia platyphylla usually occurring further from the low water mark. Higher up the bank, in areas which are infrequently flooded, Eucalyptus gymnoteles was found and at the fringes of the riparian zone Eucalyptus houseana occurred with Lysiphyllum cunninghamii. Common mid-storey species included Ficus coronulata, Flueggea virosa and the exotic Parkinsonia aculeata in areas close to the water and Acacia holosericea higher up the bank. The understorey tended to be dominated by perennial 74 grasses at all sites. In particular, the exotic Cenchrus spp. were dominant at the upstream sites (Sites 03, 04 & 05) mainly higher up the bank, and the perennial grass Cynodon dactylon was dominant in the more frequently flooded areas close to the water. The perennial herb Achyranthes aspera was a rommon understorey species at all sites, with the shrub Crotalaria spp. found at higher elevations. A schematic diagram showing the distribution of species along a profile at Site 05 is presented in Figure 4.2b. Floristic patterns Site ordination (DCA) of floristic data for the vegetation quadrats on the Blackwood River sites showed a separation of quadrats along the first floristic axis and, to a lesser extent, the second floristic axis with 88% of the variance in these first two axes (Figure 4.3a). The length of the axes are around seven and six standard deviations, indicating large floristic differences between quadrats at the extremes of the axes. Along the first axis a geographic gradient from downstream to upstream sites (Site B1 to B3) can be seen. This can be represented by rainfall changes across the sites with a strong negative correlation between mean annual rainfall and DCA axis 1 (r = 0.72, p<0.001) and with mean annual rainfall decreasing along this axis. Soil salinity was also strongly correlated with DCA axis 1 (r = 0.792, p<0.001), indicating the change to more salt tolerant species, especially in quadrats at Site B3. This also reflects the trend of increasing salinity from downstream to upstream which is documented for the Blackwood River (Morrissy 1974). The separation of quadrats along the second DCA axis appears to be between grazed and ungrazed sites. There was a positive correlation between species richness of quadrats and DCA axis 2 (r = 0.623, p<0.001) and the proportion of native perennial herbs and DCA axis 2 (r = 0.524, p<0.001), with species richness and proportion of native perennial herbs increasing along this axis. There was a negative correlation (r = -0.542, p<0.001) with proportion of exotic species decreasing along DCA axis 2. For ungrazed sites there was a positive correlation between DCA axis 2 and distance from the water's edge (r = 0.439, p = 0.01), showing a change in floristics with distance from the river. It indicates the zonation of the 75 a) Blackwood Floristic plots 6 DCA 3a 5 2a ~ (\I 2a£~ (/) '>( 4 ~al «< ~ 1~a 3a .~ •• • 3a (/) 3 3a ·;:::- • 3i 3b3a 0 Si>Jt) • u:: 2 1b 1b 1~b •• • 1a1b •• 1b~ • 1 • • • 1b it~· 'I 0 • • 0 1 2 3 4 5 6 7 8 Floristic .axis1 Eigen value = 0.8855 b) Ord River Floristic Plots 6 DCA 4 • 5 N (/) 4 4 '>( 4 4 • • 5 «< • • (.) 2 4 5 :;:: 3 (/) • ·;::: .3 • 0 u:: 2 3~ • • .. i~ 1 5 5 • ·; 5 0 • 0 2 4 6 Floristic axis 1 Eigen value = 0.8195 Figure 4.3: Detrended Correspondence Analysis (DCA) showing scatter of plots along the first two floristic axes for a) Blackwood River with numbers corresponding to the various sites along the river with Site 1 the most downstream site and Site 3 the most upstream site with a = ungrazed sites and b = grazed sites and b) Ord River with Site 1 the most downstream site and Site 5 the most upstream site. 76 vegetation, with fewer species and increased proportion of annuals and exotic species closer to the water's edge and the occurrence of flood tolerant species such as sedges. Ordination by DCA for the Ord River sites also showed a separation of sites along the first DCA axis according to the geographical distribution of sites from downstream (Site 01) to upstream (Site 05) (Figure 4.3b). This can be represented by decreasing rainfall across the sites, with a negative correlation between mean annual rainfall and DCA axis 1 (r= -0.47, p = 0.051). For the Ord River quadrats 78% of the variance could be attributed to the first two axes and the length of the axes are around 6 and 5 standard deviations indicating large floristic differences between quadrats at the extremes of these axes. There was also a positive correlation between DCA axis 1 and distance from the water's edge (r = 0.437, p = 0.05), with distance increasing along DCA axis 1. In addition, distance was positively correlated with elevation (r = 0.609, p<0.001), species richness (r = 0.332, p<0.001) and percentage of perennial herbs (r = 0.766, p<0.001) and negatively correlated with soil clay content (r = -0.565, p Life form groups Life form groups for the Blackwood River sites are averaged for number of species and cover in Figures 4.4a&b. Annual and perennial herbaceous species were most abundant in terms of number of species, whereas annual grasses had the greatest cover, with cover of shrubs and annual herbs also being prominent. For individual sites, shrubs gave the highest values of both number of species and cover and tree species were highest for cover at the ungrazed sites. Annual grasses were dominant in terms of cover at grazed sites. This included pasture weeds such as Lolium rigidum at all sites and Hordeum leporinum in the upstream (more saline) sites (Sites B2 & B3) (Table 4.2a). The overstorey was dominated by only a few species on the Blackwood River, with cover of these generally high across all sites. The dominant tree species were Eucalyptus rudis and Melaleuca rhaphiophylla with Casuarina obesa dominant at the saline upstream site (Site B3). 77 Ord River sites showed a spread of life form types in terms of numbers of species, with trees and shrubs being the most common, but in terms of cover sites were dominated by trees and perennial grasses (Figure 4.4c&d). This trend was common across all sites except for Site 02 (Lake Kununurra) where perennial herb species (in particular the emergent macrophyte Typha domingensis) replaced perennial grasses as the second most prominent life form (Table 4.2b). There was a greater richness of overstorey species at the Ord River sites, with the families Myrtaceae (Eucalyptus spp., Melaleuca spp. and Lophostemon grandiflorus), Moraceae (Ficus spp.) and Combretaceae (Terminalia spp.) being represented by more than one species. The understorey was much less diverse in comparison with the understorey on the Blackwood River sites. The five top scoring species on the Blackwood River in terms of importance value were all exotic species (Table 4.3a). These were common pasture weeds and included species which are widespread in south-western Australia such as the annual grass Lolium rigidum (rye grass) and the annual herb Romulea rosea (guildford grass). The exotic perennial grass Cynodon dactylon (common couch) is a cosmopolitan weed that occurs throughout Australia. The perenniating organ is a rhizome and couch is a common invader of bare, disturbed sites and is particularly common in the riparian zone. The annual grass Hordeum leporinum (barley grass) is somewhat salt tolerant and becomes more dominant in the middle to upper sections of the river (Sites B2 & B3). As for the native species, Casuarina obesa dominates in the saline conditions of Site B3 and only occurs at this site. The tree species Eucalyptus rudis occurs at all sites and Melaleuca rhaphiophylla occurs at all except the saline site B3. These two species are the most common overstorey species in the riparian zone on the Blackwood River. The other two most important native species are sedges in the Cyperaceae family and many of these are common in damp to waterlogged conditions. 78 a) Blackwood R. m Life Forms b) Blackwood R. - Life Forms Mean No. Species Mean % Cover 20 35 !I) 30 !I) Q) c 15 25 .c ... (,) ·;: ~ 20 10 0 .. .. !I) 0 .. .. Q) 15 .. .. 'if?.. .. (,) .. Q) 10 c. 5 (f) 5 0 0 Cl) .c .c .c Ill Ill Ill Cl) .c .c .c Ill Ill Ill ...... Ill Ill IJ) Ill :I Cl) Cl) ea :I ... .. ea ~ ... ea c ~ ... Cl) Cl) ea ea c 1- .c ::r:: ::r:: ... !!! ea 1- .c ::r:: ::r:: ...... ea Cl) m m Cl (!J :J Cl) m m (!J (!J :J :I '2 m m :I '2 m m c c :I '2 c c :I '2 c c c c Cl) c c ~ c Cl) .... c Cl) < Cl) ... < Cl) ... !l. < Cl) !l. < Cl) !l. !l. C) Ord R. m Life Forms d) Ord R. m Life Forms Mean No. Species Mean % Cover 20 35 Ill !I) 30 Q) c 15 ... .c Q) 25 (,) > ... 0 20 0 !I) 10 Q) 'if?.. 15 (,) Q) c. 10 (f) 5 5 0 0 Cl) .c .c .c 1/1 Ill 1/1 Cl) .0 .c .c Ill 1/1 IJ) :I ...... IJ) IJ) ea Cl) :I ...... 1/1 1/1 ea ~ .... Cl) Cl) ea c ...... Cl) Cl) ea ea c 1- .c ::r:: ::r: !!! ... ea 1- .c ::r:: ::r: ...... ea Cl) m m 0'1 Cl :J Cl) m m 0'1 (!J :J :I m :I m '2 :I m "2 :I m c c c "2 c c c '2 c Cl) c c c c c ... Cl) ~ Cl) < Cl) < .... < Cl) < ... !l. Cl) !l. Cl) !l. !l. Figure 4.4: Number of species and% cover of the major life form types on the Blackwood River (a & b) and Ord River (c & d). Values are means (±SEM) for six sites on the Blackwood River and five sites on the Ord River. Table 4.2a: Summary of life form groups at the Blackwood River sites. Life Form Site Bla Site Blb Site B2a Site 2b Site B3a Site B3b (grazed) (grazed) (grazed) No.spp Cover No.spp Cover No. spp Cover No. spp Cover No.spp Cover No. spp Cover (%) (%) (%) (%) (%) (%) Tree 2 20.9 1 10.1 3 3.0 2 2.6 1 26.3 1 5 Shrub 20 37.7 2 4.4 13 30.9 0 - 6 28.1 0 - Annual herb 20 11.9 16 25.4 22 20.4 7 1.5 11 13.5 6 30 Perennial herb 16 14.5 7 7.2 15 12.5 1 0.2 4 6.3 8 24.9 Annual grass 4 16 8 12.3 9 18.4 5 41.9 6 30.9 7 22.7 Perennial grass 3 2.6 3 22.1 6 6.7 0 - 5 1.9 4 2.1 Lianas 0 - 0 - 0 - 0 - 0 - - - Table 4.2b: Summary of life form groups at the Ord River sites. Life Form Site 01 Site 02 Site 03 Site 04 Site 05 No. spp Cover No. spp Cover No. spp Cover No. spp Cover No. spp Cover (%) (%) (%) (%) (%) Tree 8 55 4 48 9 35 7 43 5 45 Shrub 5 4 4 13 7 3 6 2 6 2 Annual herb 7 0.7 3 9 1 11 1 0.4 5 0.4 Perennial herb 11 9 8 30.1 3 27.9 3 0.6 4 0.4 Annual grass 0 - 0 - 0 - 1 5 1 0.8 Perennial grass 5 29 0 - 4 1 5 43 5 52 Lianas 4 3 1 2 1 0.5 2 0.3 2 0.4 ~ 80 Couch (C. dactylon) was also the second highest species in terms of importance value for the Ord River sites (Table 4.3b). The two most common overstorey species in the riparian zone of the Ord River, Eucalyptus camaldulensis and Melaleuca leucaden~ra, were the highest scoring native species and were found at all sites. The exotic perennial buffel grasses Cenchrus ciliaris and C. setiger were introduced as part of the rehabilitation programme on the upper Ord at Old Ord River Station. These species tended to occupy the outermost section of the riparian zone and the frontage lands beyond. The exotic tree legume Parkinsonia aculeata is a common woody weed throughout warmer parts of Australia and is particularly common along watercourses (Wheeler et al. 1992). Table 4.3a: Highest scoring native and exotic species according to their importance value for the Blackwood River sites. Importance value == relative frequency + relative cover + relative occurrence. Species Frequency Cover Importance Life Dispersal (%) (%) value fonn1 mode2 Lolium rigidum * 67 9.2 18.42 ag w Cynodon dactylon * 29 6.1 11.42 pg w Romulea rosea * 46 3.4 9.76 ah m Briza maxima* 32 3.7 9.21 ag w Hordeum leporinum* 21 4.8 8.44 ag ve Casuarina obesa 24 4.7 8.42 t w Melaleuca rhaphiophylla 21 2.5 6.97 t w Eucalyptus rudis 51 1.1 6.71 t w ]uncus pallidus 32 1.6 5.14 ph m Isolepsis cernua 32 0.3 4.0 ah w * Exotic species 1 Life forms are ag =annual grass: pg =perennial grass; ah =annual herb; ph = perennial herb; sh = shrub; t = tree. 2 Dispersal modes are w = wind; un = unknown; wa = water; ve = vertebrate external; vi = vertebrate internal. Table 4.3b: Highest scoring native and exotic species according to their importance value for the Ord River sites. Importance value = relative frequency + relative cover + relative occurrence. Species Frequency Cover Importance Life Dispersal (%) (%) value fonn1 mode2 Melaleuca leucadendra 65 15.2 42.2 t w Cynodon dactylon * 42 11.2 30.4 pg w Eucalyptus camaldulensis 67 7.2 28.6 t w Cenchrus ciliaris* 21 4.5 14.1 pg w Cenchrus setiger* 17 5.5 13.4 pg w Flueggea virosa 13 2.7 8.8 sh vi Lophostemon grandiflorus 23 1.7 8.7 t w Parkinsonia aculeata* 21 0.5 8.3 t m Achyranthes aspera 15 1.5 8.1 ph ve 81 Discussion Floristic structure For the rmgrazed sites on the Blackwood River and all sites on the Ord River, species richness and species diversity were fairly low, indicating the dominance of species which are well adapted to the particular conditions that prevail in the riparian zone. These adaptations include tolerance to high disturbance by flooding, scouring and high stress through waterlogging and periodic dry conditions (Szaro 1990). Variations in species diversity in riparian communities have been considered to be the result of the rate of disturbance (Ward & Stanford 1983) and along disturbance gradients in general, species richness has been reported to be highest at intermediate levels of disturbance (Grime 1979; Huston 1979). However, species diversity may also be influenced by other factors such as productivity (Tilman 1986). In a riparian wetland, species richness has been best described in terms of gradients of disturbance, productivity and spatial heterogeneity (Pollock et al. 1998). For both rivers, species richness tended to increase with distance from the river and therefore probably with reduced frequency of flooding. In particular, a sharp decline in species richness was seen in plots near the water's edge in the littoral zone, which was subject to the greatest flooding disturbance and greatest stress from inrmdation and waterlogging. The results indicate that for these Western Australian rivers, the riparian zone is not an area of high species richness or diversity and does not follow the generally accepted idea of the riparian zone being an ecotone and of higher species richness as the river collects propagules of species throughout the watershed (Nilsson et al. 1994; Bretschko 1995). Greater species richness in upland sites adjacent to the riparian zone has been found for other rivers in south-western Australia (Hancock et al. 1996). Life forms Structure of the vegetation was contrasting on the two rivers with the greatest diversity on the Ord River occurring in the number of overstorey species and the rmderstorey tending to be dominated by only a few species. Species diversity on the Blackwood River was most influenced by the understorey which, particularly on the ungrazed sites, included a high diversity of shrub 82 species. The overstorey was composed almost exclusively of Eucalyptus rudis which is taxonomically similar to E. camaldulensis and occupies a similar niche in wetlands and rivers in south-western Australia (Boland et al. 1989). The prevalence of the different life forms also differed sharply between the two rivers, with a more diverse structure on the Blackwood River and a number of shrubs, perennial and annual herbs and annual grasses. This differed between grazed and ungrazed sites, with shrubs more dominant at the ungrazed sites and annuals more common in the grazed sites. This is related to the increased disturbance on the grazed sites favouring ruderal species, whereas at the ungrazed sites the prevalence of stress tolerators such as shrubs indicate that stresses brought about by periods of inundation or drought are likely to be an important influence on vegetation structure. On the Ord River, vegetation structure was dominated by trees and perennial grasses,· while annual herbs and grasses were only a very small component of the riparian vegetation. The perennial grasses are usually hemicryptophytes which behave like annual species, with the above ground parts dying back to a rhizome or stolon in the dry season and growing rapidly and producing large quantities of seed when conditions are favourable. They are therefore disturbance adapted species. On the Ord River shrub species are more common further from the river at higher elevations where disturbance is less but stress from drought may be greater. Lianas were also common on the Ord River sites but absent from the Blackwood River sites. Soil stability will also affect vegetation physiognomy (Friedel et al. 1993) and favour ruderal species and will affect the ability of riparian species to form a soil seedbank. This will be explored further in Chapter 7. At a landscape scale, differences between the rivers are primarily driven by differences in climatic history, present climate, geology and soils (Barlow 1994). The natural disturbance regime and the level of stress caused by flooding m the two rivers are also likely to influence the factors affecting recruitment of the vegetation. Both rivers had a high level of exotic species, both in terms of number of species and cover, particularly at sites grazed by livestock. The addition of disturbance from livestock grazing in this naturally high disturbance environment makes the riparian zone highly vulnerable to invasion by exotic species which tend to be ruderal species (Grime 1979). Further, the river is an area of high productivity and a natural conduit for dispersal of propagules of exotic species 83 which can be rapidly transported over long distances. Examples of this include Atriplex prostrata and Rumex spp. on the Blackwood River and the exotic lianas Passiflora foetida and Cardiospermum halicacabum on the Ord River. The riparian zone may also provide a reservoir of exotic species within the landscape. On the Blackwood River there was a clear separation of sites where livestock grazing has resulted in the invasion of exotic species and a decrease in native species richness and diversity. This is typical of other environments that have not had a long history of intensive livestock grazing and where the vegetation is therefore not adapted (Mclntyre & Lavorel 1994; Pettit et al. 1995). On the Ord River no sites that have not been subject to cattle grazing could be found. All sites have been grazed by cattle for at least the past 50 years (Perry 1970). Therefore, m effects of grazing could be discerned for these sites. The dominance of grasses and lack of slu·ubs in the understorey is most likely related to climate and natural disturbance regimes and this vegetation, described as woodland with a grassy understorey, is typical of the low latitude summer rainfall areas across northern Australia (Perry 1970). The exotic perennial buffel grasses Cenchrus ciliaris and C. setiger dominant at Site 05 were introduced as part of the rehabilitation programme on the upper Ord at Old Ord River Station (de Salis 1993). For environments with a long evolutionary history of grazing, species richness tends to be higher with some grazing pressure (McNaughton 1985; Milchunas & Lauenroth 1993). The effects of livestock grazing will be more fully explored in Chapter 6. Floristic patterns Ordination of the sites showed the longitudinal and transverse trends in the floristics of the plots, with the longitudinal gradient of climate (such as decreasing rainfall) influencing a change in floristics. Other influences along this gradient may be changes in geology and soils. For example, on the Blackwood River there is increasing laterisation of the landscape westward (downstream), and on the Ord River there is a change from a basalt plateau overlain by sedimentary deposits of limestone, sandstone and mudstone on the Ord River plains (at the upstream sites) to the more recent alluvial sediment of the Carlton Plain downstream. On the 84 Blackwood River floristics of the riparian vegetation is also influenced by increasing river salinity upstream, and is reflected in a change to species which have greater salt tolerance, such as species in the Chenopodaceae family. On the Blackwood River sites the transverse gradient for plots (identified by the second axis of the DCA ordination) indicated a greater proportiort of annual species and reduced species richness closer to the river with increasing proportion of woody perennial species further from the river. Exotic species, which were nearly all annuals, were also more prevalent near the water. This may reflect the increased frequency of disturbance near the water, favouring ruderal species (Abugov 1982). Transportation of seed by the river and the continual exposure of fresh sediments and high soil moisture levels, as well as the reduced competition (which disadvantages competitive dominants (Pickett et al. 1989)), favour these ruderal species which in the south-west are predominantly exotic species. The only perennial species that usually occurred near the river are hydrophytes, such as from the Cyperaceae family, and the trees Eucalyptus rudis and Melaleuca rhaphiophylla which have some waterlogging or inundation tolerance. These generally establish above the summer low water mark with M. rhaphiophylla restricted to the area close to the river channel. The establishment of these species will be investigated further in subsequent chapters. A transverse gradient was also observed on the Ord River sites, with lower species richness and a greater proportion of annual species nearer the water's edge. This is possibly due to the increased frequency of disturbance and stress. Floristics of the plots are affected by the occurrence at the water's edge of semi-aquatic, emergent macrophyte species such as Typha domingensis. The greater clay content closer to the river, from alluvial deposits, will also affect floristics along the river profile. On many sites, the understorey close to the river was dominated by the perennial grass Cynodon dactylon occurring on the freshly deposited sediments. Away from the river, other perennial grasses became dominant, such as the buffel grasses Cenchrus spp. at the upstream sites. The overstorey species on the Ord River are more diverse than on the Blackwood River but the most common species which occurred at all sites were Eucalyptus camaldulensis and 85 Melaleuca leucadendra with M. leucadendra tending to be restricted to areas close to the water channel. The gradient of change in floristics along the transverse profile away from the river is correlated with elevation and probably reflects a moisture gradient and/ or a gradient of flooding disturbance. These factors will be explored further in later chapters. Disturbance and stress have important roles in shaping the vegetation communities in the riparian zone. Stress, which can include extended periods of inundation and waterlogging or drought, affects the ability of plants to grow and persist in the riparian zone. Disturbance from strong river flows can create gaps for recruitment of plants and the transport of propagules, removal and renewal of soil and replenishment of soil water. The degree to which disturbance and stress affect riparian vegetation will depend on river flow regime, distance from the river, elevation and topographical position within the riparian zone. This Chapter has provided an overview of the riparian vegetation on the study rivers and provides a basis for further examination of factors that influence establishment. Aspects raised in this Chapter such as the influence of fluvial processes, effects of livestock grazing, the soil seedbank and the fate of seed and seedlings and will be explored in greater detail in the following chapters. 86 CHAPTER 5 IDENTIFYING THE NATURAL FLOW REGIME AND THE RELATIONSHIP WITH RIPARIAN VEGETATION. Introduction The previous Chapter and evidence from the literature indicated the link between riparian vegetation and hydrology, including water flow and fluctuating water levels (Malanson 1993; Naiman & Bilby 1998). This Chapter describes the natural flow regime and the variability of flows into the riparian zone and how this can be related to the vegetation at sites on the Blackwood and Ord rivers. This chapter provides a detailed investigation of the variability in the flow regime of the study rivers over time, of the influence on the floristics of the riparian community identified in Chapter 4 and the establishment of seedlings and size class structure of common overstorey species. This knowledge of historical flows can be used to understand the distribution of riparian species and life forms in the riparian zone on these rivers. An understanding of the variability of historical flows can also be used to predict changes in the vegetation with changes to river flows in regulated systems or with climatic changes. The fluvial regime and river geomorphology are major influences on the spatial and temporal structure of riparian vegetation (Johnson 1994; Barnes 1997; Johnson 1997). In particular, water flow and fluctuating water levels are important determinants of the structure of riparian vegetation (Gregory et al. 1991; Malanson 1993; Nilsson et al. 1993). The importance of the flow regime on the ecology of aquatic fauna has been discussed for an Australian arid zone river (Puckridge 1999). Puckridge (1999) identified six measures of variability that are important to fish and macro-invertebrate biology including: duration of flow; water depth; shape of the flood 87 pulse; frequency of the flood pulse; connectivity of water bodies; and permanence. Assessment of the range of flows needed to maintain riparian species should consider more than just the aquatic species (Richter 1997) and should therefore include species of the riparian zone. Measures of temporal variability of flow are useful in describing the range of conditions that the riparian vegetation (as well as other biota) experience. The flow regime of a river describes the distribution of water through time, and can be used to determine periods of inundation of the riparian environment. Flow essentially "drives" sediment transport and, in fluvially dominated systems, shapes the river channel and consequently the structure of riparian landscapes (Young 1999). This can result in multiple successional stages (e.g. Malanson 1993). Timing, frequency and level of flooding disturbance are also important determinants of vegetation dynamics in the riparian zone. How natural flow disturbance shapes commtmity structure can. depend on the frequency, magnitude and predictability of the disturbance (Sousa 1984). Several critical aspects of river flow regimes are recognised as being more ecologically "significant" (Richter et al. 1996). These key aspects include flow magnitude; flow variability; rates of flow change; magnitude and frequency of extreme flow condition; and predictability of flows (Young 1999). The timing of river flows is important for the dispersal of vegetation propagules and in triggering germination and subsequent establishment of seedlings. This has been extensively reported in the literature (e.g. Barnes 1985; Bradley & Smith 1986; Bren & Gibbs 1986; Johnson 1994). For example, seasonal timing of flows in the Murray River is an important trigger for recruitment of Eucalyptus camaldulensis (Bren & Gibbs 1986). Flows are also indirectly important for the maintenance of riparian vegetation in recharging soil water (to reduce water stress), increasing vegetative growth rates (Bacon et al. 1993) and enhancing survival of seedlings (Roberts 1993). On the Murray River in south-eastern Australia, the abundance and composition of plant species was strongly correlated with the degree of water level fluctuation (Walker et al. 1994). In wetlands, changes in frequency an.d duration of flooding can alter plant species richness 88 (Denton & Ganf 1994) and prolonged flooding can affect growth and reproduction of emergent macrophytes (Froend & McComb 1994). To provide a framework for assessing the relationship between the natural flow regime and the structure of riparian vegetation (including life form groups discussed in the previous chapter), fundamental procedures outlined in the "holistic method" of in-stream flow management (Arthington 1998) have been adopted. This approach emphasises ecosystem level and flow dependent processes and aims to provide an assessment of environmental water requirements by using information on the relationship between flow and the various ecologically-important attributes of rivers. It assumes that the natural flow regime, or in regulated systems, the historic flow regime, is of fundamental ecological importance in maintaining channel morphology, in stream biota and riparian vegetation as well as associated floodplain and wetland environments. In regulated rivers, the holistic methodology leads to the development of a modified flow regime based on critical flows to achieve particular ecological, geomorphological and/or water quality objectives. It uses a range of flow-related parameters to determine significant environmental flows and then to develop a suitable (artificial) environmental flow strategy. This Chapter adapts this method for assessing seasonality and variability of natural flows in the Ord and Blackwood rivers and describes the frequency and variability of flows that would influence riparian vegetation. It provides examples of the influences of riparian flows on various aspects of the riparian zone vegetation including floristics, life form structure and population structure. These can be used to compare similarities and differences in the response of the different vegetation components to the different fluvial regimes. This Chapter provides an example of how this technique can be applied in the management of riparian zones and in planning rehabilitation programs of riparian vegetation. The technique is also useful in improving the ecological knowledge base for assessing environmental flows in regulated systems. Clearly, this is useful in the planning and management of riparian zone vegetation restoration. 89 Methods Using the historical daily stream flow record, average monthly flows were assessed to determine the variability of natural flows on the Ord and Blackwood rivers. Stream channel morphology and estimates of Manning's n (Newbury & Glaboury 1993) was measured at a site on each river to calculate flows that would influence riparian vegetation and flows that would inundate the riparian zone. The historic seasonality and variability of these flows was then assessed. Vegetation plots set up along transects perpendicular to the river (i.e. along an elevational gradient), and described in Chapter 4, were used to assess the effects of particular riparian flows on riparian vegetation floristics, the presence of seedlings and the population structure of overstorey species. Natural flow regime 3 1 Historical flow records (as daily discharge figures in m s- ) were obtained from Water and Rivers Commission (Western Australia) gauging sites. On the Ord River the chosen site was upstream of the Ord River Dam and Lake Argyle in the unregulated section of the river at Old Ord River Station (Station No. 809316) (Site 05, see Figure 3.1). On the Blackwood River the site was located at Darradup (Station No. 609025) (Site Bla, see Figure 3.2). Data were analysed from records covering the period October 1970 to November 1995 (25 years). Although a longer period of continuous record was available for the Blackwood River (from 1958}, the 25 year period was used to enable a valid comparison between the two rivers. To determine yearly and seasonal flow patterns and flow variability for the two rivers, mean flows for each month were calculated from daily records along with percentile flows. As a measure of the distributional spread of the daily flows over the period of record, the coefficient of variation (CV) was calculated from daily flows for each month (Zar 1984). However, for skewed data, the mean, and therefore CV, may not accurately represent the central tendency of the data (Zar 1984} so an additional measure of variability (S80 or spread) was also calculated. This measure is derived from the difference between the 90th percentile and the lOth percentile divided by the median (50th percentile) 90 (Walker et al. 1995; Young 1999). The greater the flow variability, the higher the SSO value. Overall seasonal patterns were described by indices of time-series analysis (Colwell 1974) of predictability (P), constancy (C) and contingency (M). Values of P range from 0 (unpredictable) to 1 (totally predictable). A value of 1 would occur only if flows in every month were the same for every year of record. Constancy (C) would be 1 if flows were totally predictable and identical across all months while contingency would be 1 if flows were totally predictable but different every month. The degree of seasonality can also be indicated by MIP. These Colwell indices together with monthly distribution of flows are a useful first step in defining patterns in the natural flow regime (Davies et al. 1998). Colwell's Index was calculated using log scale 2 intervals of the median (Q) monthly flows (i.e. 0.5Q, Q, 2Q, 4Q, SQ, 16Q, 32Q, 64Q) (Gordon et al. 1992) for 25 years of record for each river system. Return period of floods was also calculated for each river based on recorded annual flow peaks plotted against cumulative frequency (p) where p = (rankln+1) x 100%. This provides an estimate of annual return period of particular flows or the probability of particular flows being equalled or exceeded (Petts 1983). However, because of the short period of record, extreme How return periods must be regarded with caution. Riparian flows Water discharge rates to achieve different stage heights at a particular reach of the river can be calculated using "Manning's Equation" (Newbury & Gaboury 1993): 213 112 v = (R x s ) I n and R =Alp 1 where v =mean velocity (m s- ); R =hydraulic radius of How (m); s = average reach slope; n = Manning's roughness factor; 2 A= cross-sectional area of flow (m ); p =wetted perimeter of flow (m). 3 1 Discharge (Q) (m s- ) is then calculated as Q = v x A. Channel resistance to flow, or Manning's roughness factor (n), was estimated for different transects from examples of calculated n values for different channel types (i.e. Newbury & Gaboury 1993). Resistance to flow varies greatly depending on the size of the bed material (including rocks and boulders), the vegetation and depth of flow. Calculation of Manning's Equation allowed the estimation of river discharges, at different stage levels within the river 91 channel, that would be important for riparian vegetation (i.e. inundation). Calculation of discharge values were averaged from three transects. Analysing the historical flow record allowed the timing, frequency and duration of these discharges to be quantified and consequently correlated with the associated vegetation data. To adequately determine river channel morphology at the study sites on the Blackwood and Ord rivers, three cross-sectional transects (including vegetation transects) were measured for distance (m), elevation and slope with a surveyor's automatic level and staff at 10 metre intervals. Channel slope was measured by taking levels 100 metres up and downstream of the transect. Bed materials along the transect (including the presence of gravel, rocks and boulders, large woody debris and vegetation) were recorded to assess channel roughness and hence the Manning's n value (Newbury & Gaboury 1993). An average channel morphology was then determined for a particular reach of the river and river discharge was estimated using Manning's Equation. Critical flows for riparian vegetation for each river were calculated as: 1. Low flows - below active channel levels, which would expose wet sediments and therefore provide sites for the germination of seed and the establishment of seedlings. 2. Riparian zone flows - where water flows into the riparian vegetation zone, facilitating dispersal of seed, providing soil water recharge and redistribution of sediments. 3. Floodplain flows- where water flows out of the river channel and onto the floodplain where slope levels out and the vegetation consists of mainly terrestrial species. The occurrence, frequency and duration of these flows were then calculated for each month from historical flow records. Vegetation plots At each site on each river, two vegetation transects perpendicular to the river were established, the length of each varying for each site depending on the width of the riparian zone. Along each transect, five 10 metre x 10 metre quadrats were established (see Chapter 4). For each plot, all 92 plants rooted within the plot were identified and an estimate of percentage cover was made, where cover was determined as the vertical crown or shoot projection per species. In each plot, the stem diameter at breast height (1.5 m) (DBH) was measlired for all individuals of the two most common tree species, Eucalyptus camaldulensis and Melaleuca leucadendra on the Ord River and E. rudis and M. rhaphiophylla on the Blackwood River. For these species, the number of tree seedlings in each plot was also recorded. For greater precision of determinations of tree DBH and number of seedlings, an additional two transects were established at each site. Multivariate analysis of the floristic and environmental data from the plots was conducted using Detrended Correspondence Analysis (DCA), detrending by segments (ter Braak 1987), which provided a summary of plant species composition and abundance (see Chapter 4). Species were categorised into different life form groups to compare the dominance of the various life forms on each river and the effect of the flow regime. Flow regime and vegetation relationships The relationship between flow regime and floristics, plant functional types and population structure of the selected species at these sites on the two rivers was considered important for understanding the dynamics of regeneration of plants in the riparian zone. For the period of record, river discharge that would inundate each of the vegetation plots along the vegetation transects at these sites was calculated along with mean duration of inundation and number of events per year (Table 5.1). These were correlated with floristic data for each plot, including the DCA scores calculated in Chapter 4. Data from tree seedling plots and tree diameter plots were correlated with mean period of inundation and number of events per year. The relationship between floristic data from the vegetation plots and mean duration of flooding and mean number of flooding events per year into these plots was assessed using stepwise multiple regression (Zar 1984). In this analysis, vegetation plot data were the dependent variables and mean duration of flooding and mean number of flooding events per year were independent variables. 93 Table 5.1: River discharges that would inundate vegetation plots (calculated from Manning's equation) and the mean duration and mean number of these events per year. Blackwood River Ord River Plot 1Depth 2Discharge 3Duration 4Events Depth Discharge Duration Events No. (m)@ (m3s-1) (days) per year (m)@ (m3s-1) (days) per year 275m3s-1 2424m3s-1 1 7.0 14 59.6 1.5 7.5 35 10.2 4.6 2 6.9 14 59.6 1.5 7.3 35 10.2 4.6 3 3.4 66 14.9 1.3 6.3 160 5.4 4.1 4 3.2 66 14.9 1.3 6.2 160 5.4 4.1 5 2.9 70 13.4 1.2 5 304 4.3 2.7 6 2.5 100 10.4 1 4 500 3 2.4 7 0.1 172 9.6 0.375 3.9 588 2.7 2.1 8 0.9 260 6.8 0.25 3.2 790 2.5 1.7 9 1.1 265 6.5 0.22 0.5 2000 1.5 0.7 10 0.5 275 5.7 0.2 0 2424 1.6 0.5 1 Mean depth of water (m) in the vegetatiorl. plot at a discharge of 275 m3s-1 or 2424 m3s-1 2 Minimum discharge that would inundate each vegetation plot. 3 Mean duration of inundation at each plot over the 25 years of record. 4 Mean number of flow events per year into the plot over the 25 years of record. Results Natural flow regime Historic mean monthly river discharge for the Blackwood River showed a strong seasonality, with 80% of discharge occurring in late winter I early spring of July, August and September (Table 5.2). In contrast, summer and some autumn flows (January to May) made up only 3% of the annual flow. The coefficient of variation (CV) of flows was very high during summer, however values for January and February were distorted by an extreme flow event in late January /early February 1982 caused by a cyclonic depression. If this flow event is excluded, the CV for January and February reduces to 48% and 129% respectively. 94 Table 5.2: Historic mean monthly flows in gigalitres (GL) and measures of variability for each month from daily flow records for the Blackwood River near the. Darradup gauging site for records covering the period 1970 to 1995. Mean Median Std. Dev. Std. Err. CV% sso n January 14.75 1.16 67.26 13.45 455 1.25 24 February 2.86 0.7 8.44 1.69 294 1.52 24 March 2.35 0.77 6.25 1.25 265.6 2.3 25 April 2.70 1.61 3.43 0.69 127 2.9 25 May 5.78 4.68 3.30 0.66 57.2 2.9 25 June 32.18 19.13 36.04 7.21 112.2 3.8 25 July 119.88 93.16 90.35 18.07 75.4 3.4 25 August 163.20 117.18 128.18 25.64 78.5 2.3 25 September 95.57 73.27 89.08 17.82 93.2 2.3 25 October 41.02 32.65 31.66 6.33 77.2 2.7 25 November 12.02 11.33 6.76 1.38 56.2 2.2 24 December 3.99 3.15 3.89 0.79 97.4 1.8 24 Table 5.3: Historic mean monthly flows in gigalitres (GL) and measures of variability for each month from daily flow records for the Ord River at the Old Ord River gauging site (Site 01) for records covering the period 1970 to 1995. Mean Median Std. Dev. Std. Err. CV% sso n January 401.35 203.62 491.31 104.75 122 28.3 21 Februaxy 703.34 196.20 943.19 205.82 134 12.9 20 March 276.88 166.81 326.28 69.56 118 13.2 22 April 71.77 10.63 162.39 34.62 226 18.5 22 May 12.47 1.82 30.80 6.72 249 13.4 21 June 2.30 0.31 3.45 0.77 151 23.1 20 July 2.70 0 8.91 1.86 330 23 August 0.89 0 2.02 0.42 227 23 September 2.20 0 7.74 1.61 352 23 October 1.83 0 5.25 1.07 287 24 November 34.18 1.61 101.13 21.08 296 236.3 23 December 110.66 36.54 239.74 51.11 217 29.2 23 Comparing mean and median values indicates the skewed nature of the data, particularly in the low-flow months, suggesting that the median may be a "better" indicator of general flow conditions as the mean is unduly influenced by occasional, but extreme, flow events. Using the median to measure variability (580) gives quite different results than the CV, with the highest values measured during winter. Calculation of the Colwell Index showed that the Blackwood River had a highly predictable flow regime (P= 0.71). Constancy of 0.32 with low contingency (M) of 0.34 indicates a strong seasonality of the flows, with this seasonality highly predictable between years (i.e. M/P = 48%). Using the peak discharge levels for each year, a flood return 95 frequency curve was developed from the available years of data (Figure 5.1a). This indicates that a return frequency flood of 100 years would have a discharge of 1160 m3s·1 and that a discharge of 554 m3s·1 would occur for a 10 year return flood. However, because of the small period of record these estimates are only tentative, particularly for the higher values. Monthly discharge figures for the Ord River showed a very strong seasonality of flows with 92% of the mean yearly flow occurring from December through to March (Table 5.3). Periods of no flow were recorded for July to October, with low median flows (i.e. <2 GL) from May to November. Flow variability was very high (CV of >100% for all months), with highest figures measured during the dry season months. The large difference in mean and median values again indicates skewness. Flow variability, determined using 580, was also very high but greatest in the wet season. However, values for July to October could not be calculated because of no-flow in the 50th percentile range for these months. November showed an extremely high 580 value, reflecting the variability of climatic patterns in the transition of seasons from the dry to the wet. Colwell's Index showed a high predictability of flows (P = 0.67), with this mostly attributed to low, constant dry season flows with constancy (C) of 0.44 and 0.20 for contingency (M). The contribution of seasonality to predictability was low (M/P = 29%). Using the peak discharge levels for each year, a flood return frequency curve was developed from the available years of data (Figure 5.1b). This indicates, at this site on the Ord River, that a return flood frequency of 100 years would have a discharge of 19994 m3s·1 and that a discharge of 9458 ffils·1 would occur for a 10 year return flood. 96 a) Blackwood River 1200 ...... r...... ,.... -I ...... 1' ...... 1 tl) 1000 M E 800 ...... 1' ...... 1 -Cl,) C') !... ea 600 J: 0 tl) 400 ···.··.:.-::.·.:.- ..-.···:::.·· .. ··:.~ .. .1.: ... :.·········::·:.:.-:·::: .. ::.] ·-"C ~ ~ x ...... ea 200 ...... ,...... ~.... , :: 0 1 10 100 log Frequency o/o 2oooo b}.QrQ f!.i.Y~!:····· mm. r·m ·······················m························~ .,.... -I tJ) M -E 15000 ...... ·····r··················· ...... 10000 ...... ·······································································i1 5000 ······································] 0 1 10 100 log Frequency 0/o Figure 5.1: Flood frequency curves for a) Blackwood River and b) the Ord River based on the annual maximum discharge recorded for the period 1970 to 1995. 97 Riparian zone flows Critical low flows for riparian vegetation were calculated. as <12 m3s-1 for the Blackwood River and <30 m 3s-1 for the Ord River. Riparian zone flows were 66 - 275 rrfs-1 for the Blackwood River and 580-2424 m 3s-1 for the Ord River. Floodplain flows were >1050 m 3s-1 for the Blackwood River and >5200 m3s-1 for the Ord River. On the Blackwood River, flows that would inundate the riparian zone occurred on average 2 to 3 times per year and lasted for approximately seven days. These events occurred most frequently during July and August (Table 5.4). Flooding of riparian zones has occurred only twice between November and May over the flow period. This was after the cyclonic rainfall event in January and February 1982 (Cyclone Alby). Most consecutive days of flows >66 m3s-1 was 60 (26/7 /83 to 23/9/83) and there were 10 periods of 20 consecutive days of flows of this volume, between June 3 1 and September (winter). There were 24 days of flows >275 m s- , with 20 recorded from July to September and the other four recorded in January 1982. There was only one day of recorded flow of >1050 m3s-1 (over-bank, floodplain flows), recorded on 31/1/1982. Low channel flow events 3 1 (<12 m s- ) occurred on average 11 times per year and lasted around 23 days, so for the months November through to May these events are likely on most days. During August, such flows are unlikely and only occurred in one in three years, while July and September had flows <12 m3s-1 less than once a year. The greatest number of consecutive days of <12 m3s-1 flows was 317 days from 29/8/79 to 10/7/80. There were 21 periods of greater than 200 consecutive days of flows <12 3 1 m s- , with three periods >300 days. For the period of analysis, there were 70 days of flows >12 m3s-1 in the dry season (November to April), with the longest consecutive period being 18 days (25/1/82 to 11/2/82). Criteria for beginning of winter flows was the first 10 consecutive flows of 3 1 1 >12 m s- but not followed by >14 consecutive days of <12 rrf s- • End of winter flows was 3 1 considered to be 14 consecutive days <12m s· • Winter flows began in June 13 times, in July 11 times and in May once. The earliest start to the winter flows was 30/5/74 and the latest 27/7/76. The end of winter flows with consequent receding water levels occurred in October 16 times, 98 September 7 times and twice in November. The earliest cessation of winter flows was 1/9/79 and the latest was 9/11/73. Table 5.4: Summary of flows into the riparian zone at Darradup on the Blackwood River over a 25 yr period from 1970 to 1995. Flows are between 66 m3s·1 (active channel flow) and 275 m3s·1 (upper limit of riparian vegetation) calculated from profile transects at the site. Month n Total No. Days/Month Events/Month Duration (days) (~rs) Da~s Events Mean± SE Mean± SE Mean±SE January 24 3 1 0.12 0.04 3 February 25 2 1 0.08 0.04 2 March 25 0 April 25 0 May 25 0 June 25 27 3 1.08 ± 0.2 0.12 ± 0.05 9±2 July 25 151 18 6.04 ± 1.6 0.72 ± 0.16 8.4 ± 1.3 August 25 247 21 9.9 ±2 0.84 ± 0.12 11.8 ± 2 September 25 88 11 3.5 ± 1.2 0.44 ± 0.1 8 ± 1.6 October 25 21 4 0.84 ± 0.13 0.16 ± 0.07 5.25 ± 1.1 November 24 0 December 24 0 On the Ord River, discharges that would inundate the riparian zone occurred on average about three times per year and lasted approximately two days. Eighty-five percent of these riparian flows occurred from January to March and none occurred between June and October (Table 5.5). The short duration of riparian zone flows is reflected in the steepness of the flood hydrograph 3 1 for the Ord River (see Figure 3.6). For low channel flows (of <29m s· ) an average of 15 events occurred per year, each lasting for about 21 days. From May to November, low flows occurred virtually all month, while in the wet season (December to March) there was an average of six low flow events per year. There were thirteen periods of at least 200 consecutive days of flows 3 1 1 <29m s· • The greatest number of consecutive days of flows <29 nf s· was 331 days from 19/2/92. For the period of analysis (11/70 to 10/95) there were 25 days in the dry season (May to October) 3 1 of flows >29m s· , with the longest consecutive period being 11 days (1-11/5/74). There were four 3 1 1 days of recorded flows of >5200 m s· (over bank, floodplain flows). The highest, 16456 nf s· , was recorded on 14/2/80. Criteria for the commencement of wet season flows was the first 3 consecutive flows of >29m3 s·1 but not followed by>14 consecutive days of <29 m3 s·1 after October. Endofwetseason flows was considered to be 14 consecutive days <29m3 s·1 after February. Wet season flows began in.December 8 times, in January 8 times, in November 4 times and in February 99 once. The earliest start to the wet season was 10/11/71 and the latest 3/2/92. The end of wet season flows and the beginning of dry season flows occurred in March 12 times, April 4 times and once during February and May. Table 5.5: Summary of flows into the riparian zone at Old Ord River Station over a 25 year period from 1970 to 1995. Flows are between 580 m3s"1 (active channel flow) and 2424 m3s"1 (upper limit of riparian vegetation) calculated from profile transects at the site. Month n Total No. Days/Month Events/Month Duration (days) (yrs) Da;rs Events Mean±SE Mean±SE Mean±SE January 22 39 10 1.9 ± 0.6 0.90 ±0.25 3.9 ± 0.4 February 21 64 25 2.8 ± 0.8 1.04 ± 0.3 2.6 ± 0.4 March 22 29 12 1.3 ± 0.5 0.52 ± 0.2 2.4 ± 0.6 April 22 7 3 0.3 ± 0.2 0.14 ± 0.07 2.3 ± 0.9 May 21 1 1 0.05 0.05 1 June 21 0 July 21 0 August 22 0 September 22 0 October 21 0 November 22 3 2 0.15 ± 0.05 0.10 1.5 December 22 4 3 0.21 ± 0.1 0.16 1.3 Vegetation and flooding regime Vegetation plots showed the large difference in the floristics between the two rivers. The use of historical flows and discharge levels for each vegetation plot showed the different effects of flooding regime on floristics, life form structure and population dynamics of the riparian vegetation. On the Blackwood River, the dominant life forms in terms of species and cover were sluubs (30% of species and 36% of cover) and perennial herbs (30% of species and 20% of cover), but this varied along the gradient away from the river. Zonation of the vegetation along a flooding gradient is reflected in the relationship to duration and frequency of flooding in the vegetation plots (Table 5.6). Floristics, as indicated by the DCA, showed a strong correlation of frequency of flooding with the first floristic axis. Species richness declined with increased duration of flooding as did the cover of sluubs (Figure 5.2). Cover of exotic species, which were mainly annual herbs and grasses (with the exception of blackberry (Rubus sp.)) increased with both increasing duration and 'frequency of flooding (Table 5.6). The cover of both perennial grasses and 100 herbs showed no significant relationship with either flooding variable (Table 5.6). On the Blackwood River, the population structure of the common tree species was not significantly correlated with the flooding variables for either the nlll11ber of tree seedlings per plot or tree diameter (Table 5.6). On the Ord River, the dominant life forms were perennial grasses (22% of species and 52% of cover) and trees (16% of species and 40% of cover). Zonation of the vegetation along the flooding gradient was shown by the significant correlation between the first DCA axis and the duration of flooding (Table 5.6). In contrast to the Blackwood River, there was no significant effect of flooding on species richness or the number of exotic species, whereas cover of shrubs significantly decreased with increasing frequency of flooding (Table 5.6, Figure 5.3). Perennial grasses were the dominant life form m Site 05 on the Ord River (see Chapter 4), and percentage cover increased with duration of flooding. On the Ord River, seedling numbers of the major trees in the vegetation plots was not significantly correlated with the flooding variables. Tree size showed the opposite trend to that seen on the Blackwood River, with tree diameter significantly decreasing with increased flooding frequency (Table 5.6, Figure 5.3). 101 a) Frequency v DCA axis1 b) Duration v Species richness 4 40 Ill (!) 35 u .!!! 3.5 Cl) >< c.. Ill -« Ill 30 () 0 c 3 z 25 2.5 -t-'1.-.-..-.-,-,-,_,....,..,.""T0-,--,.-;:-.---.-"1,...... , 20~~TM~~no~~~TM~ 0 0.5 1 1.5 2 0 10 20 30 40 50 60 No. events per year Duration of flooding (days) c) Duration v cover of exotics d) Duration v cover of ann. herbs Ill 40 15 .!!! u (!) g. 30 ... ~ 10 u0 0 10 20 30 40 50 60 0 1 0 20 30 40 50 60 Duration of flooding (days) Duration of flooding (days) e) Frequency v cover of shrubs 50 (/) '§... 40 • J: (/) 30 .... Cl) ~ 20 u ~ 10 0 0 0.5 1 1.5 2 No. events per year Figure 5.2: Significant relationships between duration of flooding (b,c & d) and number of flooding events per year (a & e) with floristic attributes of vegetation plots along transects on the Blackwood River site. 102 a) Duration v DCAaxis1 b) Duration v cover of peren. grasses 6 80 ~ 70 .... 5 Ill ~ 60 1/) ... ')( 4 Cl m 50 «< <( ·;: 40 g3 c Cl) w 30 2 c. eft 20 1 10~~~~~~~~~~~~ 0 2 4 6 8 10 12 0 2 4 6 8 10 12 Duration of flooding (days) Duration of flooding (days) c) Frequency v cover of shrubs d) Frequency v tree diameter 100 50 ~ 80 40 ...:I J:: ...1/) 60 ._,'[ 30 • Cl) > J: 8 40 ~ 20 0~ 20 • 10 0 0 0 1 2 3 4 5 0 1 2 3 4 5 No. events per year No. of events per year Figure 5.3: Significant relationships between duration of flooding (a & b) and number of flooding events per year (c & d) and floristic attributes of vegetation plots along transects on the Ord River site. 103 Table 5.6: Results of step-wise multiple regression from vegetation plots of floristic variables with duration of flood events in days (x) and number of events per year (z). Dependent variable (X) 1R2 p 2Model Blackwood River DCA axis1 0.816 0.003 y = -0.79+3.935z Species richness 0.494 0.023 y =-0.189x+35.4 % cover exotic spp 0.774 0.0054 y =-1.307+0.254x+11.191z % cover annual herbs 0.570 0.0116 y = 1.512+0.162x % cover perennial grasses 0.121 0.251 % cover perennial herbs 0.152 0.562 % cover shrubs 0.597 0.0088 y = -23.80lz+45.619 No. seedlings 0.292 0.170 tree DBH 0.357 0.102 y =25.606+ 1.899x Ord River DCA axis1 0.697 0.014 y = -0.462x+6.043 Species richness 0.105 0.677 % cover exotic spp 0.015 0.949 % cover annual herbs 0.061 0.802 % cover perennial grasses 0.564 0.0123 y =6.214+4.665x % cover perennial herbs 0.150 0.392 % cover shrubs 0.758 0.001 y =-13.107z+82.436 No. seedlings 0.106 0.407 tree DBH 0.758 0.0001 =-8.539z+39.429 R value is or the ind&endent variable that gives a significant result 2 Model includes the in ependent variables that give a significant result. Discussion This Chapter highlights the strong relationship between the riparian vegetation and stream hydrology for sites on the contrasting Blackwood and Ord rivers. Given this linkage between hydrology and riparian vegetation characteristics, there appears to be wide applicability of this method for identifying the link between riparian vegetation and the flow regime. Historical flow records for a particular river can be related to patterns of vegetation development, such as reproductive phenology and seedling establishment, as well as plant community patterns in riparian zones. Therefore, analysis of natural flow regimes can be used in conjunction with other abiotic and biotic factors to predict the pattern of vegetation dynamics for the riparian zone. For example, on the Murray River in eastern Australia, reduced duration and depth of winter flows and increased summer flows through the dry season increased survival of E. camaldulensis seedlings (Bren 1992). The relationship between stream hydrology and the 104 vegetation has been recorded elsewhere for littoral plants (Walker et al. 1994) and for the population dynamics of Eucalyptus camaldulensis seedlings on the Murray River (Bren 1992). Natural flow regime The natural flow variability of the Blackwood River indicated that it is similar to the Murray /Darling River which experiences moderate variability, while the Ord River has very high variability, similar to the arid zone rivers of inland Australia such as the Diamantina River and Cooper Creek (Walker et al. 1995), (Chapter 3). The flow regime in semi-arid regions of Australia is characterised by long periods of no flow with intermittent very large flows (Puckridge et al. 1998). The Ord River showed strong predictability of flows, mainly due to low, constant, dry season monthly flows, with substantially greater variability in wet season flows. Seasonality thus contributed relatively little to the overall predictability. A similar flow pattern has been reported for streams in the Australian wet tropics (Arthington et al. 1994; Pusey et al. 1995). The highly predictable seasonality and low variability of flow on the Blackwood River is similar to the flow regime reported for other south-western Australian streams and rivers (Davies et al. 1998). High variability of flows in summer, due to high intensity short duration storms derived from tropical airflows (Gentilli 1972), can lead to episodic recruitment events as floodwaters disperse seed, provide moist seedbeds and create gaps in the vegetation. An example of such an event occurred in January 1982 when rain from a cyclonic depression caused an extreme flood event which resulted in extensive recruitment, particularly of the dominant overstorey species (E. rudis and M. rhaphiophylla). These species may display varying degrees of serotiny so that seed is generally always available if conditions are conducive to germination. Flows in the Ord River are strongly seasonal, as a consequence of climatic patterns which are mainly influenced by monsoonal rainfall in the wet season (December to March) (Gentilli 1972). Flow variability is high for all months. Flows are usually of short duration, even in the wet season, being the result of high intensity rainfall events with limited groundwater storage to 105 sustain higher base flows (see Figure 3.6). The period of seedfall of E. camaldulensis and M. leucadendra on the Ord River may be related to the reduced flows occurring in March and April. This will be explored in Chapter 7. The episodic recruitment of these species is also likely to be related to the high variability of the flow regime, with the conditions favourable for survival of seedlings being very variable. In addition, the high energy, strong flows that frequently occur can be destructive for establishing seedlings and saplings. Large floods can also be destructive to existing mature trees (Duncan 1993; Piegay & Bravard 1997). For example, during a large flood m the Mississippi River, many riparian trees were killed, however extensive sediment deposition provided substrate for subsequent seedling recruitment (Spink & Rogers 1996). Discharge and vegetation relationship There was a strong correlation between floristic data in the vegetation plots and both duration and frequency of flooding. This clearly shows the relationship of these flow variables in determining the characteristics of the riparian vegetation and the zonation of the vegetation along a gradient with distance from the low flow water level for both rivers. However, because of the dominance of different vegetation life forms on the two rivers and the different flow regimes, response to flooding is likely to be quite different. For both rivers there was an increase in disturbance-tolerant ruderal species with increased duration and frequency of flooding. These were annual herbs and annual grasses on the Blackwood River and perennial grasses m the Ord River. On the Ord River, perennial grasses are considered tolerant to flooding and are usually short-lived and are able to quickly colonise new sediments. In contrast, woody shrubs tend to be terrestrial species such as Acacia spp. which are not tolerant of waterlogging and, as a consequence, are killed by long duration of flooding or frequent flooding. In addition, these species cannot survive on unstable sediments created by frequent flooding. On both rivers cover of shrubs declined with increased flooding. This life form can be fairly long-lived and they are usually stress tolerators which do not persist under frequent disturbance. On the Blackwood River, species richness was reduced with increased duration of flooding and may reflect a gradient in tolerance to the physiological stress of waterlogging. The cover of exotic species (mainly annual species) was greater with increased duration and frequency in flooding and 106 suggests that high flooding frequency (and therefore high levels of disturbance) enhances the spread of these species which tend to be ruderal species adapted to high levels of disturbance (Grime 1979). Spread of exotics is further enhanced by ·other disturbances such as fire and livestock grazing which are common disturbances on the Blackwood River (see Chapter 6). The number of tree seedlings showed no relationship with inundation, indicating that germination of seeds and early establishment of seedlings may be more influenced by other factors including microsite, soil moisture and soil texture. Size class structure of trees and flooding showed no relationship on the Blackwood River. On the Ord River smaller sized (i.e. younger) trees occurred in areas more frequently flooded and large mature trees generally occurred at higher elevations with low flooding frequency. On the Blackwood River, trees (particularly E. rudis) can establish and mature anywhere throughout the profile, including at low elevations near the water's edge, indicating that newly established trees are not unduly affected by floodwaters. Smaller size classes tended to occur in areas less prone to flooding and perhaps in areas where presumably competition from older trees is less intense. The spatial distribution of tree size classes within the riparian zone will be examined in detail in Chapter 9. The relationship between natural flow regime and riparian vegetation outlined in this chapter, for these rivers, indicates different models of vegetation dynamics for each river. On the Ord River, the large wet season floods capable of scouring soil and destroying vegetation are highly episodic and in the intervening years smaller floods allow the development of the existing trees and the recruitment of younger age classes. This prevents the establishment of stable states and therefore the ecosystem is characterised by long periods of transition between short-lived stable states. In contrast, lower energy, more predictable seasonal flooding on the Blackwood River allows mature stands of trees to develop throughout the river profile. This disturbance regime results in long periods of stable states with short periods of transition. This aspect will explored in more detail in following chapters. 107 Aspects of the methodology of the "holistic method" adapted here was useful for identifying the natural flood disturbance regime that occurred at these sites and their impacts on vegetation dynamics of the riparian vegetation. The addition of· vegetation data can strengthen the applicability of the holistic method for determining ecologically significant river flows for setting criteria for environmental water requirements in regulated river systems. Previously, this has been determined with reference to riparian vegetation but without quantitative data (i.e. Davies et al. 1998). The techniques described here can be used to compare the response of different vegetation components of the riparian vegetation to different fluvial regimes. This technique can be expanded and applied in the management of riparian zones and in planning rehabilitation programs, such as for deciding the appropriate species to be planted in different landscape positions within the riparian zone. It may also be useful for improving the ecological knowledge base for setting environmental flows in regulated systems. Clearly, this technique is useful in the management of the vegetation in the riparian zone, not only for natural systems but also for regulated systems. 108 CHAPTER 6 EFFECTS OF LIVESTOCK GRAZING AND FIRE DISTURBANCE ON RIPARIAN VEGETATION Introduction One of the most serious impacts on the riparian zone is due to livestock. Livestock, usually cattle or sheep, can have direct impacts such as destruction and trampling of vegetation. They can also cause erosion (by trampling stream banks), compaction of the soil, increased nutrient levels, both in-stream and in the soil, and also increased in-stream turbidity (Fleischner 1994). On most rivers in the agricultural and pastoral zones of Australia, livestock have free access to the riparian zone for summer watering and grazing. In the Kimberley region of north-western Australia, cattle overgrazing of the native vegetation has caused major erosion and siltation problems in many rivers (Williams et al. 1996). With the loss of native species and reduction in cover, an increase in nutrient levels and disturbance of the soil surface, livestock can also promote invasion by weeds (usually annual ruderal species) and bring about changes in vegetation structure (Kauffman et al. 1983; Fleischner 1994). In many semi-arid areas, the riparian zone is highly susceptible to grazing as livestock tend to congregate in this zone which is the most productive and may be the only source of water (Wilson 1990; Armour et al. 1994). Grazing by domestic livestock can dramatically change the structure of the vegetation by preventing recruitment of trees and shrubs, thus transforming woodlands into grasslands (Chesterfield 1986; Gibson & Kirkpatrick 1989; Cheal 1993). In some semi-arid regions of Australia the opposite problem has emerged, with livestock promoting invasion by unpalatable woody shrubs and thus adversely modifying native grasslands (Hodgkinson & Harrington 1985; Williams 1990). In some 109 pastoral areas of Australia, such as chenopod shrublands, palatable species may become locally extinct under continued grazing pressure from domestic stock (Wilson 1990). Identification of the proportion of different life form groups among plants within a community can be useful in providing a means of predicting the community response to continued stress and disturbance such as from grazing or fire. Individual plant species response to grazing varies according to growth form and reproductive strategy. For example, annual species are more tolerant of disturbance, due to their fast growth rates and early and prolific seed set, than perennial species which tend to be comparatively slow growing and usually require several years to reach reproductive maturity (Grime 1974). Other groups, such as perennial grasses, avoid heavy grazing pressure by having above-ground growth occurring in winter and spring, and dying back to an underground storage organ over summer or the dry season (Pate & Dixon 1981) when grazing pressures are usually heaviest. This situation may occur in riparian areas that are continuously grazed. These functional strategies of plants have been observed in riverine lentic ecosystems in which competition and stress tolerance were found to be the most important traits rather than disturbance (Hills et al. 1994) Fencing to exclude stock, or to at least manage grazing, is a fundamental step towards achieving rehabilitation of degraded riparian sites. Whether grazing needs to be excluded completely will depend on the type of community and the level of existing degradation (Armour et al. 1994). In environments that have had a long history of grazing and where the vegetation has adapted to this form of stress, the exclusion of livestock may result in changes to the vegetation structure such as invasion by woody plants and reduction in species diversity (Williams 1990; Milchunas & Lauenroth 1993). Experiments on grazing exclusion in riparian vegetation have shown a reduction in species richness and an increase in plant cover (Kauffman et al. 1983). Predicting which particular species are most affected by livestock grazing and which species are likely to return after stock exclusion is important for managing the rehabilitation of degraded riparian areas. 110 Altered fire regime is another impact within riparian zones that may have detrimental effects on regeneration and particularly in combination with livestock grazing. Many plants in the riparian zone may require a disturbance such as strong rivet flows or fire to provide gaps for recruitment (Kellman & Tackaberry 1993) and to stimulate seed release (Lamont et al. 1991) and germination (Shea et al. 1979). In managed areas, prescribed fires for fuel-reduction or to stimulate new growth of edible species may be too frequent, resulting in the loss of many species and the increase in weeds (Busch & Smith 1993), or fire may be too infrequent to create gaps to stimulate recruitment of native riparian species. A large proportion of the riparian zone of the Blackwood River and virtually all of the Ord River are subjected to uncontrolled livestock grazing and fire. Therefore a study of the recruitment of riparian vegetation of these rivers needs to investigate the effects of these disturbances. Vegetation analysis in Chapter 4 indicated that a major determinant of floristic composition is livestock grazing, particularly for the Blackwood River sites. These differences will be explored in greater detail in this Chapter, along with preliminary results of the recovery of vegetation where grazing has been excluded. This Chapter will examine aspects of livestock and fire disturbance in the riparian zone. These include surveying what effects livestock grazing has had on the floristics of sites on the Blackwood and Ord Rivers, particularly in relation to species richness and diversity; structural changes in terms of life form; and seedling establishment and continued recruitment of particular species. This chapter will provide a more detailed analysis of the effects of livestock grazing m the riparian vegetation that were described Chapter 4. In addition, using grazing exclusion plots, this Chapter will describe short term changes in the vegetation in the absence of grazing and will present some data on the effects of fire on riparian vegetation on these rivers. 111 Methods Sites surveyed for comparison of grazed and ungrazed riparian zones are described in Chapters 3 & 4 and included Sites B1a & b, B2a & band B3a & b on the Blackwood River and Sites 04 & OS on the Ord River. Blackwood River Site B1 was located on the lower section of the middle catchment with Site B1a in an area of intact jarrah (Eucalyptus marginata) forest upslope of the riparian zone. Site B1b is some S kilometres upstream in an area where land upslope of the riparian zone has been cleared for pasture and is grazed by cattle which have free access to the riparian zone. Site B2a was located in a remnant of around 10 hectares of the original jarrah/wandoo woodland surrounded by cleared agricultural land. Site B2b was on the opposite bank of the river, in an area of pasture cleared down to the riparian zone and which is grazed by sheep. Site B3 was in the upper catchment in the area most affected by salinity and consisted of a remnant of native vegetation of around 30 hectares adjacent to the river (Site B3a) and of an area of cleared land 1 kilometre upstream adjacent to the riparian zone which is grazed by sheep (Site B3b). Stocking rates in the grazed sites can vary considerably but usually range from O.S - S sheep per hectare, with greatest grazing pressure in summer when annual pastures die off and there is a lack of green feed in the surrounding agricultural land. In areas cleared for agriculture, a strip of riparian vegetation along the river was generally left with at least the overstorey intact. For the Ord River, areas of the riparian zone that are not grazed by cattle are extremely rare and so for this comparison a site within the Ord River regeneration area (Site OS) was selected as a non-grazed site. This site has been free from cattle grazing for the past 30 years, although it was heavily grazed in the past (de Salis 1993). Throughout this Chapter this site will be referred to as ungrazed, although this is not strictly the case. It was compared to Site 04 on the Negri River (a major tributary of the Ord River) around 8 kilometres downstream which is currently being grazed at around 1 animal per 10 hectares, although cattle densities are probably much higher around the river during the dry season (Petheram & Kok 1991). 112 At each site, transects perpendicular to the river were established, the length of each varying depending on the width of the riparian zone. Along each transect, 10 metre x 10 metre quadrats were established at variable distances located to coincide with the bands of vegetation, giving a total of 10 quadrats at each site. Within each plot, all plants rooted within the plot were identified and an estimate of percentage cover was made, where cover is the vertical crown or shoot projection per species. In addition, along the 10 metre wide belt transects, size class (as diameter at breast height (DBH)) distribution of overstorey species was recorded. Vegetation surveys were completed in July and August 1997 on the Ord River and in October and November 1997 on the Blackwood River. Response to grazing of species common to both grazed and ungrazed sites is compared using a response index (Williams 1969). This is the ratio of cover of a species per unit area in grazed quadrats to the cover of a species per unit area in ungrazed quadrats. Exclosure plots were established at three grazed riparian sites on the Blackwood River (Site B4, Site B5 and Site B3b, see Figure 3.2). These sites were adjacent to cleared agricultural land and had been subject to livestock grazing for at least the last 25 years. At each site five 10 metre x 10 metre quadrats were pegged out and a sheep-proof ring-lock fence was erected around each plot. Adjacent to each fenced plot an open 10 metre x 10 metre plot was also pegged. While fencing excluded domestic livestock, rabbits and native herbivores could gain access. All plants rooted within each plot were identified and an estimate of percentage cover was made. The initial survey of all exclosure and adjacent open plots was completed in September and October 1996 for Site B4 and Site B3b and October 1997 for Site B5, with subsequent surveys carried out in October 1997 and 1998. On the Ord River livestock exclosure plots were set up in April 1998 at Site 01 (in the area downstream of the Diversion Dam) and were re-surveyed in April 1999. Three 20 metre x 20 metre plots were fenced with three strands of barbed wire to exclude cattle and adjacent to each of these exclosure plots a 20 metre x 20 metre open plot was also pegged out. Fencing was designed to exclude cattle but allow access to native herbivores. 113 For each survey at all sites all species within a plot were recorded and their percentage cover within the plot was estimated. Numbers of seedlings of overstorey species was also recorded within each plot. To identify the effects of grazing on. plant functional types, species were grouped according to life form. Attributes within each plot were averaged for a site so that grazed and ungrazed sites could be compared. A Mann-Whitney non-parametric U test was performed for species richness and total vegetation cover between means for the exclosure plots and adjacent open plots to test for a significant change after 12 months. At Site B2a (ungrazed) on the Blackwood River a wildfire occurred in April 1997 and the transects on this site were re-measured in April 1999 (i.e. two years post-fire) to assess the recovery of and effect of this fire on the riparian vegetation. At Site B4 on the Blackwood River a controlled bum was carried out in May 1996 in part of the heavily grazed riparian zone and five 10 metre x 10 metre exclosure and open plots were established in this area to assess the effect of fire on natural regeneration and to compare changes with the plots in the unbumt area. In addition, to assess the effect of fire on the release of seed from the dominant overstorey species at this site, Melaleuca rhaphiophylla, thirty branches from ten trees at each of the burnt and unburnt sites were examined and the proportion of open fruit was recorded. Results Comparison of grazed and ungrazed riparian areas For two of the three paired sites on the Blackwood River, species richness and diversity was significantly higher at the ungrazed sites than at grazed sites (Table 6.1). Percentage of litter within a plot was also significantly greater in ungrazed plots whereas percentage of bare ground and number of tree seedlings was not. In terms of species, native perennial herbs, and particularly 114 shrubs, showed significantly greater numbers in the ungrazed sites, comprising an average of 26% of species and 31% of cover in the ungrazed sites with no shrubs present in the grazed sites on the Blackwood River (Table 6.2). Cover of perennial herbs was significantly greater at the grazed Site B3b due to the dominance of a single unpalatable sedge species, Gahnia trifida. There was significantly greater foliage cover of exotic annual herbs and exotic grasses in the grazed site at Site B1 and of exotic grasses at Site B3. There was generally a high proportion of cover of exotic species in the ungrazed locations at all sites. Table 6.1 : Comparison of floristic attributes, percentage bare ground and percentage litter cover as well as number of tree seedlings between ungrazed and grazed locations at three sites on the Blackwood River. Figures are means of 10 plots at each site. P values are for Mann-Whitney U test for ungrazed and grazed comparisons .. Site B1 Site B2 Site B3 a b p a b p a b p un~razed ~razed un~razed ~razed ungrazed ~razed Species richness 29.3 12.1 0.004 16.7 6.8 0.008 14.7 14 0.691 Species diversity 2.5 1.3 0.006 2.0 1.0 0.004 1.8 1.6 0.262 %bare ground 38.7 19.1 0.540 23.3 26.6 0.681 18.7 21.6 0.513 % litter cover 27.5 4.1 0.005 12.8 6.5 0.025 48.3 4.3 0.008 tree seedlings 4.3 15.3 0.127 17.3 18.5 0.632 6.8 2.5 0.210 For the Ord River comparison there was no significant difference in any of the floristic attributes or life form groups between the grazed (Site 04) and ungrazed (Site 05) sites (Tables 6.3 & 6.4). An exception was the significantly greater cover of exotic species at the ungrazed site which is most likely due to the deliberate introduction of species such as Cenchrus ciliaris, C. setiger and Aerva javanica to stabilise highly degraded areas at this site (Site 05) (de Salis 1993). There was also a significantly greater proportion of bare ground at the grazed Ord River site (Table 6.3). 115 Table 6.2: Comparison of plant life form groups between ungrazed and grazed locations at three sites on the Blackwood River by number of species and percentage foliage cover. Figures are means of 10 plots at each site. P values are for Mann-Whitney U test for ungrazed and grazed comparisons. Life Site Bl Site B2 Site B3 Form1 a a p a b p a b p ungrazed grazed ungrazed grazed un__S!_azed grazed species species species nah 4.7 2 0.01 1 0.3 0.37 3 1 0.05 eah 4.5 4.7 0.86 4.3 2.5 0.10 2.7 2.5 0.83 eg 1.8 2.8 0.39 3.2 3.2 - 3.2 4.5 0.03 nph 6.5 1.3 0.001 3.5 0 - 1.0 2.7 0.02 ns 10.2 0 - 4 0 - 2 0 - % foliage cover %cover %cover nah 6.3 4.3 0.44 0.6 0.1 0.38 0.5 0.9 0.35 eah 5.1 18.1 0.001 11.7 2.1 0.15 7.5 16.1 0.02 eg 9.4 28.3 0.001 19.7 27.3 0.47 18.3 19.7 0.81 nph 13.4 5.3 0.02 9.4 0 - 6.3 16.0 0.01 ns 10.2 0 - 17.6 0 - 22.5 0 - 1Life Forms: nah=native annual herbs; eah=exohc armual herbs; eg=exotic grasses; nph=native perennial herbs; ns=native shrubs. Grazing response index gives an indication of the response of particular species to grazing. This showed the strong positive response to grazing of many species of exotic armual herbs and grasses at the Blackwood River sites, although some of these species, such as the exotic armual herb common to disturbed areas, Hypochaeris glabra, showed a negative response (Table 6.5). Native perennial grasses had a fairly neutral or slightly negative response to grazing, with species benefiting from a reduction in cover of other perennial species at grazed sites but also being susceptible to heavy grazing. At the Ord River sites, species representing a variety of life forms responded positively to grazing (Table 6.5), including the exotic vine Passiflora foetida, a widespread weed of rivers and creeks in the Kimberley region (Wheeler et al. 1992). 116 Table 6.3 : Comparison of floristic attributes, percentage bare ground and percentage litter cover as well as number of tree seedlings between ungrazed and grazed locations on the Ord River. Figures are means of 10 plots at each site. P values are for Mann-Whitney U test for ungrazed and grazed comparisons. · Site 05 Site 04 ungrazed grazed p Species Richness 7.2 5.7 0.23 Species Diversity 2.0 2.0 0.46 %bare ground 19.0 56.5 0.01 % litter cover 9.7 9.2 0.32 No. of tree seedlings 49.0 28.0 0.14 Table 6.4: Comparison of plant life form groups between ungrazed and grazed locations on the Ord River by number of species and perc~ntage foliage cover. Figures are means of 10 plots at each site. P values are for Mann-Whitney U test for ungrazed and grazed comparisons. Site 05 Site 04 Site 05 Site 04 un~razed ~razed p ungrazed ~razed p Life form species % foliage cover exotic species 3.4 1.9 0.06 48.3 13.5 0.001 annual herbs 0.3 0.1 0.92 0.53 0.05 0.84 perennial grasses 2.1 1.8 0.46 46.7 13.6 0.001 perennial herbs 0.5 0.3 0.52 0.17 0.31 0.21 shrubs 1.4 1.2 0.78 1.6 0.57 0.24 trees 2.0 2.1 0.89 28.1 22.0 0.47 lianas 0.5 0.2 0.75 1.5 2.4 0.12 Comparison of size class distribution of overstorey species for grazed and ungrazed sites on the Blackwood River reveals the absence of trees in the smallest size class (i.e. juveniles and saplings) at the grazed sites (Figure 6.1a). This may indicate that little recruitment into the population has taken place for some time. In contrast, the greatest proportion of trees at the ungrazed sites are in the smallest size class, as would be expected as large numbers of recruits enter the population with munbers reducing over time through natural attrition. On the Ord River, size class distribution was similar for both grazed and ungrazed sites, with the greatest proportion in the smallest size class and numbers reducing with size class (Figure 6.1b). 118 Table 6.5: Grazing response index (cover in grazed/cover in ungrazed) for species common to grazed and ungrazed sites on the Blackwood and Ord rivers. Blackwood River Ord River Grazing Grazing 1 Species (life form ) response Species response index2 index Avena fatua (ag)* 23.5 Passiflora foetida (v)* 16 Aira careophyllea (ag)* 7.1 Eragrostis tenellula (ag) 5.2 Romulea rosea (ah)* 6.9 Sida fibulifera (s) 5.1 Cynodon dactylon (pg)* 4.4 Heteropogon contortus (pg) 4.6 Vulpia myuros (ag)* 4.3 Achyranthes aspera (ph) 2.5 Arctotheca calendula (ah)* 3.5 Parkinsonia aculeata (s)* 1.6 Lolium rigidum (ag)* 2.7 Cynodon dactylon (pg)* 1.0 Gahnia trifida (ph) 2.4 Datura inoxia (ah)* 1.0 Bromus diandrus (ag)* 2.1 Crotalaria novae-ltollandiae (s) 0.7 Ptilotus mangelsii (ph) 1.5 Cyperus lwloschoenus (ph) 0.6 Briza minor (ag)* 1.2, Dantlwnia caespitosa (pg) 1.0 Helipterum cotula (ah) 1.0 Anagallis arvensis (ah)* 0.8 Neuracltne alopecuroidea (pg) 0.7 Austrostipa elegantissima (pg) 0.7 Hypochaeris glabra (ah)* 0.5 Galium murale (ah)* 0.4 Oxalis corniculata (ah)* 0.13 Cotttla turbinata (ah)* 0.05 1 life forms include : ah =annual herb, ag =annual grass, ph =perennial herb, pg =perennial grass, s =shrub, v =vine. 2 grazing response index= mean cover at grazed site/mean cover at ungrazed site * exotic species Grazing exclusion from grazed riparian land After three years 'relief' from grazing, vegetation in exclosure plots at the heavily grazed sites on the Blackwood River have shown only small changes. The major difference has been a slight increase in species and cover of exotic species and also some native species (Figure 6.2). Changes have been variable across the sites as the majority of species are annuals and therefore are highly influenced by yearly variations in climatic conditions, particularly timing and amount of rainfall. Only at Site B3b has there been a noticeable increase in cover of native species in exclosure plots relative to open plots due to the growth of the perennial herb Gahnia trifida, with a relative growth rate in exclosure plots of 0.32 m2 yr-1 compared to 0.17 m2 yr-1 in the open plots. Changes in nuriiber of species and cover of the different life form groups in the exclosure 119 plots relative to changes in adjacent open plots over the study period show similar variability between sites (Figure 6.3). Native perennial grasses and herbs increased in both number of species and cover at all sites. Indicating that they have benefited from relief from grazing. Exotic annual herbs and grasses had generally increased but yearly variability would make discerning any trends difficult at this early stage. At the only site where shrubs were originally present (Site B4) these had also increased in number and cover, including the legume species Acacia acuminata and Jacksonia furcellata. The abundance of seedlings of overstorey species was similar in exclosure and open plots at two of the sites but numbers were much greater in the exclosure plot at Site B3b (Figure 6.4). New seedlings appeared each year in both exclosure and open plots but older seedlings were rarely seen in the open plots. There was some survival of seedlings into subsequent years in the exdosure plots (Figure 6.4). Exclosure plots at Site 01 on the Ord River showed no significant increase in the number of species but a significant increase in cover of the vegetation after 12 months. Mean number of species per 400m2 changed from 15.7±0.9 to 18.3±0.7 (p == 0.415) and mean total vegetation cover increased from 110.5±8 to 164±17 (p == 0.003). In contrast, in the adjacent open plots species richness fell slightly, but not statistically significantly, after 12 months, with mean number of species changing from 18.3±3 to 16.6±2 (p= 0.211) but vegetation cover significantly decreased from 133±10.4 to 96.4±3 (p = 0.004). At the initial survey and after 12 months there was m significant difference in the mean number of species between exclosure and open plots (p=0.324) and (p=0.169). While vegetation cover was significantly greater in the open plots initially (p=0.012), after 12 months cover in the exclosure plots was significantly greater (p=0.008). In terms of life form groups, number of species in each group have not changed significantly in either exclosure or open plots (Figure 6.5a). For cover, annual herbs, perennial herbs and lianas had shown the greatest increase in the exclosure plots, while perennial grasses have declined in the open plots (Figure 6.5b). 120 ...... m 11111 fenced m exotic fenced native .... D· ... open - exotic ----0---- open- native a) Site 83b m 18 100 m 16 90 Cl) 80 c 14 :I.. .c 12 Cl) 70 0 > 60 10 0 a: 0 50 m 8 -~- Cl) ~/ ~ 40 o"' .,.,. 0 '(j 6 52------C)~,~~~~ 30 D c.Cl.) 4 20 ~.. en 2 10 0 0 1996 1997 1998 1996 1997 1998 b) Site 84 en 18 100 m 16 T 90 Cl) 80 c 14 ll.. .c Cl.) 70 0 12 > ....~ ,,.''.J_, ...L 60 ' a: 10 0 ' ' T 0 50 ' . ' ' m 8 ~ ' Cl) ·······• 0 40 ' . '(j 6 ·················!············· :!: 30 r:r Cl) c.. 4 .... o~~~------o 20 en 2 ()~------~ ~ 10 0 0 1996 1997 1998 1996 1997 1998 c) Site 85 m 18 100 m T 90 Cl.) 16 c 14 . :I.. 80 .c Cl) 12 > 70 0 ....~ 0 60 a: 10 0 T 50 M 8 ~ Cl) 0 40 i p ...... ______! S2 '(j 6 30 ~ Cl) c.. 4 20 e...... en 2 10 ------0 0 0 1996 1997 1998 1996 1997 1998 Figure 6.2: Change in number of species and percentage cover of vegetation in plots where livestock grazing has been excluded, compared with adjacent open plots where grazing has continued, at three riparian sites on the Blackwood River. Values are obtained from the means (±SEM) of five plots at each site. a) Site B3b 121 Species Cover trees native shrubs native perennial herbs native grasses native annual herbs exotic grasses exotic annual herbs ~3 -2 -1 0 1 2 3 4 5 -1 0 -5 0 5 10 15 20 change in no. of species change % cover b) Site 84 Species Cover trees native shrubs native perennial herbs native grasses native annual herbs exotic grasses exotic annual herbs -3 -2 -1 0 1 2 3 4 5 -1 0 -5 0 5 10 15 20 change in no. of species change in % cover c) Site Species Cover trees native shrubs native perennial herbs native grasses native annual herbs exotic grasses exotic annual herbs •3 •2 m 1 0 1 2 3 4 5 •1 0 •5 0 5 10 15 20 change in no. of species change in % cover Figure 6.3: Change in number of species and percentage cover of different life form groups in exclosure plots relative to adjacent grazed plots after 3 years at Site B3b and Site B4 and 2 years at Site B5 on the Blackwood River. Values are obtained from the means of five plots at each site. 122 70 (/) 60 :t: ;:] :I.. 50 0 (!,) :I.. 40 0 -:I.. (!,) 30 .a E ;:] 20 z 10 0 U)f'-..(1()m mm fenced open fenced open fenced open Site 84 Site 85 Site 83b 0 seedling 1 year old 2 years old 3 years old [] 4 years old Figure 6.4: Occurrence of tree seedlings (Eucalyptus rudis & Melaleuca rhaphiophylla at Site B4 & BS, Casuarina obesa at Site B3b) over a three year period in fenced exclosure plots and adjacent unfenced grazed plots at three riparian zone locations subjected to heavy livestock grazing on the Blackwood River. 123 a) b) 20 tn Cl.) 15 '(j Cl.) c. tn -0 10 a.. Cl.) .c E 5 ::lc: 1998 1999 1998 1999 1998 1999 1998 1999 fenced open fenced open D annual herbs EJ perennial grasses perennial herbs shrubs ~ trees • vines Figure 6.5: Change in (a) number of species and (b) percentage cover of different life form groups in fenced and adjacent open plots over 12 months at riparian site (01) subjected to cattle grazing on the Ord River. 124 Effects of fire on riparian vegetation In fenced plots at Site B4 which were burnt in 1996 cover .of exotic herbs increased from 37% to 52% (p=0.003) and in tmburnt plots decreased from 27% to 24% (p=0.314) after 3 years. Cover of exotic grasses also increased from 17% to 66% in burnt plots (p=0.014) and from 40% to 71% (p=0.023) in unburnt plots. Thus fire had either no effect or a slightly negative effect on reducing the cover of these exotic life forms which were dominant at this site prior to burning. A similar result was seen in the numbers of species of these life forins. For native plants, although not significant, there was a decrease in cover (10.6% to 3.6%, p=0.08) and species richness increased slightly (4 to 5.5, p=0.46). Number of species of native perennial herbs has increased in the burnt plots and cover of perennial grasses has also increased, while there has been a reduction in cover of annual herbs. In llllhurnt plots, over the same period, cover of perennial herbs and shrubs has also increased (Figure 6.6). Overall, for native perennial species, five species have been gained and one lost in the burnt plots after three years and six species have been gained and two lost in the unburnt plots. The fire at this site appears to have initially stimulated the germination of seedlings of the main overstorey species (Melaleuca rhaphiophylla and Eucalyptus rudis) (Figure 6.7). However after three years the number of seedlings and young saplings is similar for the burnt and llllhurnt plots. Inspection of mature seed capsules on branches of Melaleuca rhaphiophylla showed no significant difference in the proportion of open capsules between the burnt (47.9% of capsules open) and llllhurnt plots (48.7% of capsules open) (p = 0.96). This would indicate that fire has not stimulated a greater release of seed. Twenty four months after a wildfire in an ungrazed riparian site on the Blackwood River (Site B2a) species numbers have changed little from pre-burn levels measured in 1997 (Figure 6.8). Cover of exotic species has increased significantly after the fire, from a plot mean(±SEM) of 31±6% to 72±5% (p<0.0001). The greatest increases were in the exotic grasses, particularly Ehrharta calycina and E. longiflora. The native annual grass Austrostipa compressa was also prominent post-burn (2.5±0.5%), but was not recorded at this site pre-burn. Cover of native perennial herbs and shrubs decreased from 29±5.4% pre-burn to 19.4±3% post-burn (p=0.084). 125 10 9 14 tn 12 ·c:;CD 8 7 c.CD 10 tn 6 8 0 5 -11- CD 4 6 J:'l 3 E 4 :s 2 c 1 2 0 o~~~~~--.-~~~~~ !1.. !1.. :$-~ ~~~ ~ 'Q .;> 1996 0 native annual herbs [3 native grasses G trees shrubs native perennial herbs Figure 6.6: Change in number of species and cover of different native life form groups in burnt and unburnt fenced plots in a heavily grazed riparian site (Site B4) on the Blackwood River. 140 tn :!:! :s 120 L.. 0 CD 100 11- 0 80 -L.. 60 .cCD E 40 :s c 20 0 1996 1997 1998 1996 1997 1998 unburnt burnt [). seedling 1 year old 2 years old 3 years old Figure 6.7: Occurrence of tree seedlings (Eucalyptus rudis & Melaleuca rhaphiophylla) in fenced, burnt and fenced, unburnt plots over a three year period at Site B4 on the Blackwood River. 126 a) Mean No. Species b) Mean cover 100 90 80 I- 70 ~ 60 0 0 50 '#. 40 30 20 10 0~~~~~-----~~~~- 1:ne-burn post-burn pre-burn post-burn D exotic annual herbs § exotic grasses native annual herbs native perennial herbs shrubs Figure 6.8: Change in (a) mean number of species and (b) mean cover of different life form groups before (pre-burn) and 24 months after (post-burn) a wildfire at Site B2a on the Blackwood River. 127 Many rommon species such as Acacia saligna, Hypocalymma angustifolium, Melaleuca rhaphiophylla and M. incana had been completely defoliated by the fire but the majority of individuals had resprouted from underground lignotubers qr epicormic buds. In addition, many seedlings of native perennial species were observed, including Eucalyptus rudis (10 seedlings), Melaleuca rhaphiophylla (54), Acacia saligna (122), Jacksonia sternbergiana (21), Hypocalyma angustifolium (10) and Dodonea viscosa (5). Discussion Grazing effects on riparian vegetation One obvious effect of intensive herbivory in a native plant community is a reduction of species diversity through the elimination of palatable species (Whelan 1989). If intense grazing continues over a long period the grazed area will tend towards becoming a grassland because of the competitive advantage in the life cycle of grasses (Miles 1979). This trend has been documented in sub-alpine woodlands in the eastern states of Australia (Bryant 1971; Wimbush & Costin 1979; Gibson & Kirkpatrick 1989). Disturbances such as grazing and soil disturbance have led to loss of native species, with partial replacement by exotics in temperate Australian grasslands (Mcintyre & Lavorel 1994). On the Blackwood River livestock grazing in the riparian zone has resulted in species richness and diversity being reduced as many native perennial species have been lost and replaced by fewer exotic annual species. The exotic species recorded at the study sites are mainly grasses and annual herbs most of which are ruderal species with characteristics which allow them to thrive in highly disturbed environments. These attributes include seed dormancy, short life histories and early flowering (Grime 1979). The invasion by exotic species may also be enhanced by the reduction in cover of perennial species creating gaps which ruderal species are best able to exploit. Native perennial shrubs and herbs have suffered the most significant loss of species. These species are generally slow-growing stress tolerators which are particularly susceptible to 128 grazing pressure in the late summer and early auturrm when annual pasture species have died leaving perennial plants in the riparian zone to provide the only source of green feed for livestock and the river as the only source of water. Th~s loss of perennial species has also resulted in a simplification of the structure of the riparian vegetation, an impact observed in other environments in Australia (Gibson & Kirkpatrick 1989; Cheal 1993). In riparian zones in the USA, livestock grazing has been shown to reduce species richness and to alter vegetation structure by the loss of woody perennial species in particular (Kauffman et al. 1983; Fleischner 1994). In environments with a long history of grazing, species diversity usually exhibits a bell shaped response (Grime 1974), with highest species diversity at intermediate levels of disturbance (such as grazing) and species diversity reduced in the absence of grazing (McNaughton 1985; Williams 1990; Milchunas & Laurenroth 1995). However, environments in south-western Australia, although they are adapted to some grazing by herbivores such as kangaroos, are poorly adapted to heavy grazing by large, hard-hoofed herbivores such as sheep and cattle. Large numbers of exotic grasses and annual herbs were seen in ungrazed riparian areas. This may be due to the high level of natural disturbance of the riparian zone and the availability of moisture which favours these ruderal species. These conditions suit the ruderals capacity for fast growth rates and ability to rapidly exploit resources while they are available. This and other forms of disturbance, such as edge effects, increased fire frequency and fertiliser drift combine to affect remnants of native vegetation in an agricultural matrix (Saunders et al. 1991). Transportation of seed of ruderal species may also be facilitated by water-borne dispersal along the river. At Site B3 the riparian vegetation is also affected by rising levels of salinity, where many salt tolerant exotics such as Atriplex prostrata and Hordeum leporinum can colonise. These results have shown that grazing does not inhibit germination of seed of overstorey species. However, from analysis of size classes and results from the exclosure plots it is clear that these new germinants are generally not surviving beyond the first 12 months. Grazing and competition for light and moisture from weeds and detrimental effects on soil structure, such as compaction, 129 are likely to be the most important factors in preventing survival of tree seedlings (De Steven 1991; Pettit et al. 1998). This lack of recruitment is a serious threat to the composition and structure of the riparian vegetation because of the important role of trees in the functioning of the riparian zone (Bunn 1993). On the Ord River there was little difference between grazed and ungrazed sites, which may reflect the long history of grazing of this environment and the dominance of the understorey by Poaceae spp. which are generally well adapted to grazing (Wilson 1990). The grassy understorey predominant in the east Kimberley has evolved as a result of the seasonally dry climate and the frequent burning regime (Beard 1979). Grazing animals other than cattle, including feral horses, donkeys, goats and pigs may also have a considerable impact on the riparian zone vegetation. Significantly reduced cover of perennial grasses and a larger proportion of area of bare ground are the only indicators of greater grazing pressure at the grazed site (Site 04). The fact that the ungrazed site (Site 05) was heavily grazed in the past may contribute to the lack of difference between the grazed and ungrazed sites. High levels of bare ground at the grazed site may indicate that livestock are causing problems by increasing the erosion potential in the riparian zone through destruction and trampling, a common problem of over-grazing (Wilson 1990; Armour et al. 1994). Seedlings of overstorey species were recorded for both grazed and ungrazed Ord River sites and the size class distribution indicates that recruitment into both populations is taking place. Grazing therefore does not appear to affect tree recruitment. This may indicate that abiotic factors such as river hydrology and geomorphology are more influential than livestock grazing on overstorey recruitment on the Ord River sites and mask the effects of grazing on the vegetation (Friedel et al. 1993). The influence of river fluvial processes on recruitment patterns of the vegetation will be explored in later chapters. There may also have not been enough sites sampled to discern any effect of grazing. Further, sampling may also need to be stratified to isolate the grazing effect from the influence of other factors. Other possible explanations for the lack of recruitment of overstorey species m the grazed sites on the Blackwood River while recruitment was adequate on the Ord River, is the greater grazing intensity on the Blackwood River. However, grazing intensity in the 130 riparian zone on the Ord River is higher than in the surrounding vegetation as cattle congregate near the river, particularly in the dry season. There may also be more alternative sources of fodder on the Ord River, such as the abundant perennial grasses. The palatability of the overstorey species of Eucalyptus and Melaleuca are unlikely to be different for the two rivers. The processes by which livestock grazing can prevent the recruitment of riparian overstorey species is an important aspect of regeneration in the riparian zone and requires further investigation to clarify these issues. There is wide variation in individual species response to grazing. On the Blackwood River native shrubs are obviously the most susceptible to grazing as they are almost completely absent from the heavily grazed sites. Common pasture weeds such as the annual grasses Avena fatua and Aira careophyllea and the exotic annual forb Romulea rosea all responded positively to grazing and would be regarded as increaser species (see Hacker 1987). These annuals are ruderal species, well adapted to exploiting disturbed conditions (Grime et al. 1986). They complete their life cycle in less than one year and therefore are usually absent in the dry sununer months when grazing pressure in the riparian zone is most intense as livestock feed has disappeared from surrounding annual pastures. Other annual exotics such as Hypochaeris glabra showed a negative response to grazing. This species has been reported as being highly palatable to stock and therefore is a decreaser species (Williams 1969), however it is well adapted to colonising disturbed sites and under favourable conditions can occur as a short-lived perennial species (Marchant et al. 1987), which with relief from heavy grazing, can rapidly colonise and dominate a site. Native perennial grasses such as Dantlwnia caespitosa and Stipa elegantissma show little response to grazing. They are generally not palatable to stock and the above ground parts of the plants die back in summer thus avoiding the season of highest grazing pressure. In fact, D. caespitosa has been reported as an increaser species with livestock grazing in woodlands in western New South Wales (Williams 1969). Species which showed a positive response to grazing on the Ord River include the increaser perennial grasses Heteropogon contortus and Eragrostis tenellula which both tend to be 131 unpalatable to stock when mature (Petheram & Kok 1991). The buffel grasses (Cenchrus ciliaris and C. setiger) showed a negative response to grazing. These species were planted for rehabilitation of degraded rangeland at Site 05 (de Salis 1993), where they have extensively invaded the riparian zone. They are highly valued for livestock feed and for their ability to establish on degraded rangeland (Petheram & Kok 1991; de Salis 1993). Native perennial species such as the shrub Crotalaria novae-hollandiae and the sedge Cyperus holosclwenus showed a negative response to grazing. These perennial species, particularly the shrubs, are the most susceptible to livestock grazing, with green foliage available all year round and fairly low stature rendering the whole above-ground part of the plant accessible to livestock, including the perenniating organ which is usually above or at ground level. C. novae-hollandiae is reported to be a common species of disturbed sites and not very palatable to stock but will be grazed in the absence of more palatable species (Petheram & Kok 1991). Effects of relief from grazing The results of exdosure experiments on the Blackwood River showed little improvement after three years, with only minor increases in the occurrence and cover of native species. Increase in these species may be difficult in competition with the greater abundance of exotic grasses and annual herbs which have increased in cover in the absence of grazing. Exclosure experiments in a riparian zone in the United States have shown an increase in competition between plants with rapid vegetative growth of strong competitors (Kauffman et al. 1983). The study sites in the riparian zone in the present study are highly degraded and perhaps do not contain sufficient stores of vegetative and sexual reproductive material of native species, such as underground storage organs or a soil seedbank, to allow natural regeneration. These sites may require intervention in the form of weed removal and reseeding or replanting to re-establish native species. Sources of seed available through the seedbank and through dispersal will be discussed in subsequent chapters. Ord River exclosure plots showed a marked increase in vegetation cover and therefore biomass over a 12 month period (one growing season). Over the same period, adjacent open plots showed a 132 decrease in vegetation cover, indicating that these changes in cover are not due to yearly variation in climatic conditions, such as extended wet season or substantial rainfall events in the dry season. However, 12 months is an insufficient period to fully assess the recovery of the vegetation with relief from livestock grazing, and therefore livestock impact. These preliminary results do indicate, however, that there is a grazing effect on the vegetation. In a long term study of cattle grazing in the riparian zone on a stream in north central USA, shrubs and graminoid species showed the greatest increase in cover with relief from cattle grazing in exclosure plots, and cattle were estimated to utilise 65% of annual growth of the vegetation in the grazed plots (Shultz & Leininger 1990). On the Ord River, seedlings of tree and shrub species such as Barringtonia acutangula, Sesbania formosa, Parkinsonia aculeata and Acacia farnesiana were uncommon but some did occur in both exclosure and open plots so that there appears to be m effect of grazing on germination of seedlings. This concurs with observations in the surveys of the grazed and ungrazed sites. Cover of perennial grasses remained the same and cover of perennial herbs increased in the exclosure plots while both reduced in the open plots, suggesting that these life forms are the most affected by livestock grazing. Although cover of these life forms is reduced in the grazed plots this does not necessarily indicate that the area is being degraded or that species are being lost, as species richness between plots showed little difference. Exclosure plots on the Ord River do not show the same initial level of degradation (such as invasion of exotic plants and loss of native species) that were seen on the Blackwood River sites and, as a consequence, recovery with grazing relief may be more rapid. High growth rates experienced in the tropics, due to warm temperatures and adequate moisture, may also speed up the recovery of biomass. These results are short term and a much longer period is required to look at the vegetation dynamics and successional processes of these sites. Pioneer species such as annuals may eventually be replaced by another suite of species depending on future disturbance and management (Onans & Parsons 1980; Wahren et al. 1994). The riparian zone may need some episodic disturbance such as a flood and/ or particular climatic conditions for a successful recruitment event. 133 Fire Recovery of the riparian vegetation from fire at an ungrazed site (Site B2a) on the Blackwood River highlights the importance of resprouting in the recovery process for this vegetation. Many of the native woody species at this site are facultative seeder/sprouters which allows them to recover leaf area quickly after fire through resprouting from epicormic buds or from an underground storage organ such as a lignotuber (James 1984) and also to re-establish through germinating seed. Resprouting after defoliation is a common strategy in fire-prone environments in Australia (Gill 1981) and in the jarrah forest of south western Australia (Bell & Koch 1980) and the facultative seeder/sprouter strategy is considered to be the most resilient strategy in environments where predictable low frequency disturbance such as fire is common (Keeley 1986). Weed species became more prominent in a grazed Blackwood River site (Site B4) after burning. This was also seen in the burnt exclosure plots, with little effect on native species. However, there may be a detrimental effect as fire enhances the invasion of exotic grasses and annual herbs which, because of their high flammability, in turn makes the vegetation more fire prone (Milberg & Lamont 1995). Exotic grass species such as Ehrharta spp., which is common in degraded areas of the riparian zone on the Blackwood River, have been recorded elsewhere as increasing fire frequency (Baird 1977). The combination of increased fire frequency and rising salinity has lead to the replacement of native riparian woody species by an invasive weed, Tamarix ramosissma in the south-western United States (Busch 1995). The combination of disturbances such as flooding and increased fire frequency in the riparian zone can favour ruderal species such as annuals (Grime 1979) which are generally exotic species and strong competitors and which can prevent the re-establishment of native perennial species (Hobbs 1988). Burning may enhance the germination of seed of overstorey species and improve short term survival by creating areas of bare ground suitable for seed germination (Wellington & Noble 1985b) but once the biomass of annual species recover these may quickly reduce survival of seedlings. Counter-intuitively, the fire did not appear to stimulate greater seed release from 134 Melaleuca rhaphiophylla fruit. However, the trees in the unburnt area were not in good health and were senescent with many dying branches, so that a hot dry late summer/autumn caused the opening of the seed capsules anyway. Unfortunately,. as assessment of fire effects was made opportunistically, no observation was made of the pre-burn status of the M. rhaphiophylla capsules in the area that was burnt. Although no direct examples of the effects of fire on riparian vegetation were measured on the Ord River, data from experimental fires on a wet tropical river at Kakadu in the Northern Territory (Douglas 1999) indicated that time of year was a critical aspect of the effect of fire on the riparian vegetation. Annual late dry season burning resulted in significantly less diverse and abundant riparian communities and recovery was slower than after burning early in the season. Vegetative resprouting from epicormic buds and lignotubers was the principal method of biomass recovery for the perennial vegetation (Douglas 1999). In the dry tropics there is a shift towards greater grass cover following late dry season fires (Bowman et al. 1988) and this probably extends into the riparian zone. In fact, because of the greater abundance of mesic species and rainforest species, the vegetation in the riparian zone is much more susceptible to fire than the surrounding upland vegetation, with greater probability of a loss of species richness. For example, Douglas (1999) recorded no vine species in riparian areas that were burnt annually compared to many species of vine in riparian areas that had not been burnt. The occurrence of vines at all the study sites on the Ord River (see Chapter 4) would indicate that these areas have not been subject to fire recently. This Chapter has detailed aspects of disturbance from livestock grazing and fire on riparian vegetation, particularly on life form groups and recruitment from seed of overstorey species. It has shown that on the Blackwood River livestock grazing has significantly changed the structure and floristics of the riparian vegetation and is inhibiting the recruitment of the overstorey species (principally Eucalyptus rudis and Melaleuca rhaphiophylla). Exclosure plots indicated that recovery of native perennial species, with grazing relief, may be slow and require a disturbance event to reduce competition from annual exotic species and stimulate recruitment. 135 On the Ord River, effects of livestock grazing were not as obvious and may be masked by other forms of disturbance or stress. Long term monitoring of exclosure plots will provide more conclusive evidence on livestock grazing effects. This Chap.ter also implicates the importance of the role of sources of seed in the resilience of riparian vegetation to these disturbances. Factors influencing survival of seedlings and the possible role of fluvial processes in recruitment of the vegetation will be examined in more detail in the following chapters. 136 CHAPTER 7 SOURCES OF SEED FOR RECRUITMENT OF RIPARIAN COMMUNITIES Introduction Mechanisms significant for recruitment are important in contributing to maintenance and the resilience to disturbance of the riparian vegetation. These mechanisms include aspects of the phenology of reproduction and growth, development and release of propagules, dispersal of propagules and the storage of mature seed that will germinate when conditions become suitable. The riparian zone can be a high disturbance environment through flooding, but also, as discussed in the previous Chapter, through livestock grazing and fire. Strategies for regeneration that are successful in this environment will help to shape the vegetation communities it supports. This Chapter assesses the relative importance of these mechanisms by examining the contents of the seedbank in the soil, the storage of seed on trees, mechanisms for propagule dispersal, the importance of dispersal by water and the survival and longevity of seed. This is achieved through a survey of the content of the soil seedbank, the dispersal mechanisms found in the riparian vegetation, the abundance of propagules in flood debris and a focus on the reproductive phenology and fate of seed of the two most common overstorey species identified in Chapter 4. These were, Eucalyptus camaldulensis and Melaleuca leucadendra on the Ord River and E. rudis and M. rhaphiophylla on the Blackwood River. Eucalyptus camaldulensis is a dominant riverine species on the Ord River and throughout continental Australia, except for the south-west where it is replaced by E. rudis which is a closely related species that is known to hybridise with E. camaldulensis (Boland et al. 1989). E. rudis is the most common large tree that occurs in the riparian zone of south-west rivers, including the Blackwood River. Melaleuca leucadendra on the Ord River and M. rhaphiophylla on the Blackwood River have 'similar ecological traits and are common components in the overstorey or 137 midstorey, having high tolerance to inundation and waterlogged conditions. By studying these functionally equivalent riparian tree species in two contrasting river systems and climatic zones, the influence of different physical environments on seedlip.g recruitment in the riparian zone can be understood. For recruitment to take place plant propagules of native species must generally come from nearby undisturbed vegetation in the canopy seed store or from seed stored in the soil seedbank. Regeneration of native vegetation from seed stored in the soil is therefore very important (Onans & Parsons 1980). The seedbank will have a bearing on species composition after a disturbance such as grazing or fire (Roberts 1986), allowing recovery from catastrophic events (such as wildfire), severe grazing or clearing (Grime 1979). In lentic wetlands, the seedbank is also very important in determining the vegetation community (Thompson 1992) and assessment of the composition of the soil seedbank can be used to predict the outcome of succession after a disturbance (van der Valk 1981). The role of the seedbank in the riparian zone is unclear and little work has been done on this, however the highly disturbed nature of the riparian zone and the frequent movement of soil sediments would suggest that the role may be minimal (Malanson 1993). In the riparian zone, the soil seedbank is therefore likely to be important only for annual or ephemeral species that can germinate and complete their life cycle within one season, thereby avoiding seasonal flooding. The seedbank may be more abundant at higher elevations from the river where soil is less likely to be scoured away or in protected wetland areas that act like lotic systems. For seeds to enter the long term seedbank they need some form of dormancy and be able to maintain viability for a long period of time, ensuring that germination only occurs under favourable conditions (Schneider & Sharitz 1986; Thompson 1987). For the Myrtacaeous overstorey species studied on these rivers survival and longevity of seed may not be sufficient for them to enter the soil seedbank. Dispersal of propagules is an important mechanism for successful colonisation and establishment of species and is likely to influence the vegetation structure of communities (Brown 1992). Successful colonisation can be dependent upon species' ability to disperse into a particular 138 environment (Barrat-Segretain 1996). Identifying the array and establishing the relative importance of different plant diaspore dispersal modes in a plant community can help identify the relative advantage of a dispersal type in a particular habitat (Willson 1992). Recolonisation and spatial patterns of succession can also be related to the relative importance of the different modes of dispersal. Plant dispersal unit in combination with life form and reproductive strategy has been used to predict species composition after disturbance (Mclntyre et al. 1995). In riparian habitats, longitudinal gradients of dispersal mode have been identified (Weaver et al. 1925), with wind dispersal being more prevalent in the headwaters and dispersal by animals being more important in the lower reaches. An obvious means of secondary dispersal for riparian plants is via water (Malanson 1993). Species may release seeds into flowing water so that they can be dispersed further downstream away from parent trees and deposited on newly exposed sediment when waters recede (Naiman et al. 1998). For this to occur, seeds need to be able to float and maintain viability (Nilsson et al. 1991b). In addition, reproductive phenology, particularly seed fall, needs to coincide with seasonal changes in river hydrology. Timing of seed production and seed fall is also very important so that seed is released at times of most favourable conditions. This includes, after floodwaters have receded and fresh soil is exposed (White 1979), when soil moisture levels are high (Roberts 1993) and when there is peak abundance of seed vectors (Willson 1992). Seed may be released from the fruit on the tree or remain within the fruit when it dehisces to be released later as the fruit dries out or is broken up by physical or chemical deterioration. During floods, woody plant debris large and small is carried downstream. This debris tends to accumulate around obstacles or at the high flow mark along the river and may contain an important source of propagules for recruitment of riparian plants in the form of fruit and seeds (Naiman et al. 1998). This Chapter investigates the soil seedbank in the riparian zone at study sites on the Blackwood and Ord rivers and whether diversity and abundance changes with lateral distance from the river. It will also assess seed predation by ants and longevity of seed of E. camaldulensis and M. leucadendra on the Ord River and E. rudis and M. rhaphiophylla on the Blackwood River. This 139 will indicate whether these are likely to be factors in the ability of these species to become part of the soil seedbank and to therefore be available for regeneration when favourable conditions arise. The spectrum and abtmdance of different dispersal modes enconntered in the riparian vegetation on these two rivers will also be documented. Flood debris is examined as an important source of plant propagules and for its role in the dispersal of these propagules, as will the floating ability of particular representative species and therefore their potential to use river flows as a dispersal mechanism. The reproductive phenology of the study species will also be described. Methods Soil seedbank On the Blackwood River at Sites B1a, B2a, and B3a soil samples were collected in November 1997 along the two vegetation transects at distances from the river (low water level) of 0-10 metres, 10-20 metres, 20-30 metres, 30-40 metres and 40-50 metres, giving 10 samples per site and a total of 30 samples. On the Ord River samples were collected in July 1997 at Sites 01, 04 & 05 along the vegetation transects at distances from the river (dry season water level) of 0-10 metres, 10-20 metres, 20-30 metres, 40-50 metres and 70-80 metres, giving 10 samples per site and a total of 30 samples. At each sampling point, ten 5 cm x 5 cm cores of soil were taken to a depth of 10 cm m a radius of 0.5 metre. These samples were then bulked to give a soil sample of 2500 cm3 at each sampling point. Samples were transferred to a glasshouse in Perth where seedling trays of 35 cm x 40 cm x 5 cm were lined with hessian and a layer of vermiculite, over which a soil sample was thinly spread. Soil in each tray was divided in half with one half receiving tap water and the other water infused with smoke from burning of plant material which has been shown to improve germination of some seeds (Brown 1993; Dixon et al. 1995). Trays were watered daily in ambient conditions for both light and temperature. Each sample was given an initial treatment 1 of a broad spectrum fnngicide (Benlate 1g L" ) and watered with a liquid fertiliser initially and later as required. Germination was recorded weekly over a 16 week period and specimens of some 140 individuals were transplanted into pots and grown until identification was possible. Soil in the trays was then turned over, to release any seed that may have been too deeply buried to allow germination, and monitoring continued until germination in. the trays ceased. Propagule dispersal mode All species of plants identified in vegetation plots surveyed in Chapter 4 were categorised according to their principal mode of diaspore dispersal (i.e. Willson et al. 1990; Malanson 1993). This was based mainly on examining the morphology and structure of the seed or fruit and on field observation (Willson et al. 1990). It should be noted that many species may have an important secondary mode of diaspore dispersal, such as by water in the riparian zone for those diaspores with the ability to float (Nilsson et al. 1991b). Dispersal categories are based m those suggested by Malanson (1993) and include hydrochores (dispersed by water), including diaspores of aquatic plants with structures which enhance floating ability (cyclochores); anemochores (dispersed by wind) which can include diaspores with wings (pterochores), very small hard seeds (sporochores) or diaspores with hair-like or plumose appendages (pogonochores); and zoochores (dispersed by animals), diaspores of which can be transported internally for fleshy fruited species (sarcochores) or externally for diaspores with spiny or glandular appendages (desmochores) and which also include plants with an elaiosome and which are thus transported by ants (myrmechores) (Berg 1975). Where the principal form of dispersal was not obvious from the structure of the diaspore or had not been observed, it was considered unknown. Flood debris samples At each site on both rivers, five samples of 2500 crrf were collected of flood debris that had accumulated, usually at the high water mark from the previous years wet season or winter high flows or where debris had accumulated around an obstruction. Samples were taken in the vicinity of the vegetation transects at each site. Samples were air dried in oven at 50° C for 12 hours and then sieved in a 2 mm mesh sieve. The larger material was examined and the presence of all fruit and seed was, where possible, identified and recorded. The sample portion less than 2 141 mm in size was mixed with coarse sand, spread onto trays and treated as for the soil seedbank samples to record germination of seed within the debris samples. Differences between sample means of soil and debris samples were analysed using paireq t tests (Zar 1984). The Czekanowski Index (Kent & Coker 1992) was used to measure the floristic similarity between soil and debris samples. The coefficent values range between 0 (complete dissimilarity) and 1 (total similarity) and are based on the joint occurrence of a species using abundance data. Seed floating Seed and fruit of commonly occurring species were collected from the study sites on the Blackwood and Ord rivers for a seed floating experiment. These species included, from the Blackwood River, the trees Banksia seminuda, Exocarpos sparteus and Dodonea viscosa, the shrub Acacia saligna, the exotic herb Rumex crispus and the sedges Lepidosperma gladiatum, Juncus pallidus and Carex appressa. On the Ord River, the species included the trees Ficus coronulata, Terminalia platyphylla, T. hadleyana, Cathormion umbellatum and Pandanus spiralis, the shrub Abrus precatorius and the exotic vine Passiflora foetida. Five replicates of 20 seeds or fruit of each species were placed into 100 ml beakers of deionised water which were stirred daily. Percentage of seeds floating was recorded for each species over a one month period. For the two major overstorey species on the Blackwood River (Eucalyptus rudis and Melaleuca rhaphiophylla) and the Ord River (Eucalyptus camaldulensis and Melaleuca leucadendra), five replicates of 50 seeds each were floated over a one month period and any seeds that germinated during this time were recorded. Reproductive phenology of target species Phenology of the two dominant overstorey species on each river was monitored over a two year period. On the Blackwood River, 10 seed traps were placed at Sites B1a & b and B2a & b under the canopy of Eucalyptus rudis trees where estimated canopy cover above the traps was greater than 60%. Traps consisted of a cone-shaped calico sack with an opening of 44 ern diameter (area 2 of 0.15 m ) suspended on steel posts one metre from the ground. Contents of traps were examined every two months for the presence of seed and other reproductive material such as opercula and 142 aborted flowers and fruit. To measure seed fall for Melaleuca rhaphiophylla, closed mature fruit on branches of ten trees at each site were tagged and these fruit were examined every two months for open capsules. During each visit, along vegetation transects, percentage of trees flowering and/or fruiting as well as whether this was light, medium or heavy was recorded. The stage of development of fruit or flowers was also recorded as well as the presence of fruit from previous seasons growth. On the Ord River reproductive status of all trees along the vegetation transects were recorded during each visit. Seed traps were not used on the Ord River due to the difficulties of access during the wet season. Instead, branches of E. camaldulensis and M. leucadendra with flower buds and/or new un-open fruit were tagged in July 1996 and the number of open fruit was recorded at subsequent visits in March 1997, December 1997, April 1998 and April 1999. The timing of flowering and seed release of these Ord River species was confirmed with Department of Conservation and Land Management staff in Kununurra and with a local seed collector (C. Dupe, personal communication). Seed longevity The longevity in the soil of seed of E. rudis and M. rhaphiophylla on the Blackwood River was tested by burying nine lots of 50 seeds at a depth of 1 cm at Site Blb. The location of each burial site was clearly marked and after 1, 6 and 12 months, three lots of seed were retrieved by taking a soil core sufficiently large and deep to encompass the burial site. These soil cores were then spread thinly on a bed of filter paper and vermiculite saturated with deionised water and placed in growth cabinets at 20°C. These were watered daily and germination of seed was recorded until no further germination took place. Similar cores nearby were also taken at the same time and treated to promote seed germination to assess the natural occurrence of seed. Control replicates of seed of each species that had not been buried was also germinated at this time and germination results for buried seeds are expressed as a percentage of numbers of control seeds germinating. On the Ord River, this procedure was repeated for seed of E. camaldulensis and M. leucadendra, except that seed was retrieved after six months and 12 months only. 143 Seed predation Rates of seed removal of the four study species by ants was assessed by recorcling the loss of seed from ant feeding stations (i.e. Yates et al. 1995). Feeding st~tions consisted of a petri dish with a lid on a plywood board with three access holes cut into the side of the petri dish. Ten seeds were placed in the middle of the petri dishes and removal of seed was monitored every two hours over an eight hour period of daylight for two days. Each feeding station consisted of two petri dishes, one of which ants were prevented from entering by placing a sticky barrier arotmd the entrances. This was done to assess the loss of seed other than from predation, such as by wind. Ten pairs of feeding stations were deployed at one site on each river and the experiment was repeated in November 1997 and March 1998 at Site B1 on the Blackwood River and July 1996 and April 1998 at Site 01 on the Ord River. Results Soil seedbank Number of species and abundance of germinants in soil samples showed no significant difference ( t test, p> 0.1) between the two watering treatments (smoke water and tap water) and results presented here are therefore for combined watering treatments. On the Blackwood River a total of 45 species from 23 families germinated from soil samples (Table 7.1), with an average of 1318 ± 2 535 seeds m· • Species which occurred most often included the exotic annual herbs Hypochaeris glabra, Juncus bufonius and J. capitatus and the exotic annual grasses Briza minor and Lolium rigidum. Species with greatest number of germinants also included the exotic annual herbs Romulea rosea, Rumex acetosella and Galium murale and the exotic annual grass Aira caryophyllea. The greatest proportion of species were annual herbs, making up 76% of species and 78% of germinants, with native perennials including grasses, shrubs and trees making up less than 2% of the germinating seed 144 Table 7.1: Species germinating from soil and debris stored seed collected in November 1997 at 3 sites on the Blackwood River. *exotic species Species Family Soil samples Debris samples (n=30) (n=15) No. Freq.% No. Freq.% Exotic annual herbs *Anagallis arvensis Primulaceae 78 57 26 80 *Arctotheca calendula Asteraceae 6 3 *Atriplex prostrata Chenopodaceae 162 62 *Cotula turbinata Asteraceae 4 3 4 3 *Crassula alata Crassulaceae 82 29 *Crassula glomerata Crassulaceae 22 43 *Epilobium sp. Onagraceae 5 3 *Erodium botrYts Geraniaceae 2 3 2 3 *Galium mura e Rubiaceae 20 30 4 3 *Hypochaeris glabra Asteraceae 44 56 12 3 *Juncus bufonzus Juncaceae 296 72 462 100 *Juncus capitatus Juncaceae 120 100 108 100 *Mentha pulegium Laminaceae 20 12 16 6 *Monopsis simplex Lobeliaceae 56 43 4 3 *Oxalis corniculata Oxalidaceae 6 43 *Petrorhagia velutina Caryophyllaceae 8 3 *Romulea rosea Iridaceae 14 30 *Rumex acetosella Polygonaceae 10 6 4 3 *Rumex crispus Polygonaceae 16 30 38 80 *Sagina apelata Caryophyllaceae. 64 6 4 9 *~ergularia rubra Caryophyllaceae 4 3 * rif.olium campestre Fabaceae 4 3 32 20 *Trifolium subterraneum Fabaceae 34 17 Native annual herbs Alternathera nodiflora Amaranthaceae 64 13 53 60 annual herb Asteraceae 8 3 ffehelia brizula Centrolepidaceae 8 6 elipterum cotula Asteraceae 29 9 Helipterum laevum Asteraceae 50 42 12 60 Isolehsis australensis C&Eeraceae 20 12 Lobe ia alata Lo eliaceae 4 3 23 18 Podolepsis /essonii Asteraceae 6 3 Poranthera microphylla Euphorbiaceae 8 3 5 6 Quinetia urvelia Asteraceae 2 3 Schoenus annual Cyperaceae 1 3 Walhenbergia priessii Campanulaceae 2 3 Exotic grasses *Aira careophyllea Poaceae 78 30 20 17 *Avena fatua Poaceae 24 9 *Briza minor Poaceae 98 67 16 9 *Ehrharta longiflora Poaceae 20 6 122 30 *Lolium rigidum Poaceae 105 57 64 40 *Holcus lanatus Poaceae 28 6 12 6 Native perennial herbs Comesperma virgatum Polygalaceae 1 3 Lepidosperma sp. Cyperaceae 1 3 Native Shrubs Halosarcia lepidosperma Chenopodaceae 8 6 Native Trees Eucareptus rudis Myrtaceae 2 3 24 60 Mela euca rhaphiophylla Myrtaceae 4 3 71 67 unidentified 15 5 145 Table 7.2: Species germinating from soil and debris stored seed collected in July 1997 at 3 sites m the Ord River. *exotic species Species Family Soil samples Debris samples (n=30) (n=15) No. Freq.% No. Freq.% annual herbs Blumea diffusa Asteraceae 70 8 Heliotropium sp. Asteraceae 10 25 8 10 Asteraceae sp. annual Astereaceae 4 17 1 10 Asteraceae sp.2 Astereaceae 1 8 Cyperus sp. annual Cyperaceae 36 42 34 50 Cyperus sp.2 Cyperaceae 2 8 22 30 Eu£horbia sp. Euphorbiaceae 81 42 In igofera sp. Papilionaceae 18 25 *HJ{fatis suaveolens Laminaceae 8 6 *St a acuta Malvaceae 23 19 Portulaca sp. Portulaceae 12 7 Calandrina quadrivalvis Portulaceae 2 16 10 20 *Solanum sp. Solanaceae 1 8 4 10 Annual herb 3 17 3 10 Annual herb2 2 7 annual grasses Eragrostis tenellula Poaceae 20 25 5 10 Poaceae sp. annual Poaceae 26 50 Sporobolus sp. Poaceae 13 17 13 10 Brachyacne convergens Poaceae 8 15 5 10 perennial grasses Paspalidium jubiflorum Poaceae 54 31 ce,nodon dactn,lon Poaceae 75 43 11 20 * enchrus ci iaris Poaceae 110 41 Heteropogon contortus Poaceae 1 10 Poaceae sp.1 Poaceae 9 25 Poaceae sp.2 Poaceae 1 8 Poaceae sf.3 Poaceae 3 8 perennia herbs Achyranthes aspera Amaranthaceae 43 50 8 60 Cyperus sp. Cyperaceae 6 6 3 10 Bulbostylts barbata Cyperaceae 6 32 Fimbristylis sp. Cyperaceae 2 10 *Passiflora foetida Passifloraceae 13 100 shrubs Acacia sp. Mimosaceae 1 6 2 10 Croatalaria sp. Papilionaceae 2 7 2 20 *Calotropis procera Asclepiadaceae 3 10 trees Eucalvrtus sp. Myrtaceae 2 10 Termmalia czatyphylla Combretaceae 11 20 Terminalia ursarina Combretaceae 2 10 Melaleuca leucadendra Myrtaceae 43 67 68 80 Eucalfeptus camaldulensis Myrtaceae 20 32 19 80 Nauc ea orientalis Rubiaceae 2 10 unidentified 16 4 146 (Figure 7.1a). There was a high level of variability among soil samples, with the number of germinants ranging from 54 to 330 per sample and number of species ranging from 7 to 20. Variation in abundance of germinants with distance from the Blackwood River is shown in Figure 7.2 a&b. The relationship between abundance of the soil seedbank and distance from the river is shown in Table 7.3 with no significant correlation with any life form group, although total abundance of viable seed in the soil appears to increase further from the river. For the Ord River, a total of 32 species from 17 families were recorded in the soil samples, with an average of 652 ± 291 seeds m-2 (Table 7.2). Common species included the annual herb Euphorbia sp., the perennial herb Achyranthes aspera and the tree Melaleuca leucadendra, with the most abundant species including Euphorbia sp and the perennial grasses Cynodon dactylon, Cenchrus sp. and Paspalidium sp .. Perennial grasses and annual herbs were the dominant life forms in terms of species and abundance on the Ord River, although there was large variation between samples (Figure 7.2). Woody perennials (trees and shmbs) also made up a more substantial proportion of germinants in terms of species and abundance than was seen in the Blackwood River samples. Variability between samples was also high, with abundance ranging from 10 to 178 per sample and species from 3 to 12. There was no significant relationship between distance from the river and total soil seed bank abundance (Figure 7.2 c&d, Table 7.3). However, numbers of germinants of grasses increased with distance from the river while number of germinants of woody perennials decreased with distance from the river (Table 7.3). 147 90 a) Blackwood River Life Forms Seedbank mean % 80 70 D Species .... 60 I: Number (I) 50 ,_(,) (I) 40 D. 30 20 10 0 exotic native exotic native annual annual grasses perennials herbs herbs b) Ord River- Life Forms Seedbank mean % annual annual perennial perennial woody herbs grasses grasses herbs perennials Figure 7.1: Mean percentage (±SD) of each life form by species and number of germinants in soil samples averaged from 30 samples at three sites on (a) the Blackwood River and (b) the Ord River. 148 a) b) 30 450 400 25 (j) 2 350 Cl.l c (.) 20 Cl'S 300 Cl.l c c. E 250 (j) 15 (j; 200 Cl'S Cl 0 10 150 1- .! 100 5 0 1- 50 0 5 1 0 15 20 30 35 50 5 1 0 15 20 30 35 50 Distance from river (m) Distance from river (m) c) d) 30 300 25 (j) Q) 20 (.) Cl.l g. 15 :§ 10 .! 100 0 0 1- 1- 5 50 0 5 1 0 15 25 35 40 55 60 0 5 10 15 25 35 40 55 60 Distance from river (m) Distance from river (m) Figure 7.2: Mean (±SE) total number of species and total number of germinants in soil seedbank samples with distance from the river channel for the Blackwood River (a) & (b) and the Ord River(c) & (d). 149 Dispersal mechanisms The spectra of diaspore dispersal types for the Blackwood River showed the dominance of wind dispersal both in terms of number of species and cover (Figures 7.3 a&b). The number of plant species for which the major form of diaspore dispersal was not known, that is, had no obvious morphological structure for dispersal, was also high. The same trend was seen for individual sites, but with water dispersal significant in terms of cover at Sites B1b, B2a and B3a, where semi-aquatic species, such as emergent macrophytes, occurred. Only one species, the exotic rambling shrub Rubus discolor (Blackberry), was thought to be primarily dispersed by birds. This species is also probably effectively dispersed by water as its distribution may indicate. Secondary dispersal by water is also likely for many species and, based on diaspore ability to float, a mean of 63% of species across the sites could also be effectively dispersed by water. Table 7.3: Relationship between distance from the river of a soil sample and the abundance of germinants in the sample for total species and for the various life form groups from 30 samples on the Blackwood and Ord rivers. Blackwood River Ord River r value p value r value p value total germinants 0.727 0.0643 total germinants 0.052 0.9840 exotic annual herbs 0.685 0.0890 annual herbs - 0.468 0.1463 native annual herbs - 0.335 0.4625 grasses 0.685 0.0090 exotic grasses 0.332 0.4675 woody perennials - 0.651 0.0190 woody perennials 0.180 0.6991 150 a) Blackwood R. ~ Seed Dispersal b) Blackwood R. m Seed Dispersal Mean No. Species Mean % Cover 25 70 (/) 60 (/) 20 Cl) r:: I.. 50 .r:: Cl) .2 15 ~ 40 a: 0 ~ 30 ~ 10 0 (,) 20 ~ 5 Cl) 10 o~~~~~~~~~~~~ , .... r:: , ... ..-. r:: Q) Q) .... r:: r:: ~ r:: r:: X Si: - Q) Si: ... - Q) «< :;:..- 0 «< :;:...... ,- 0 ~ - iij r:: ~ iij r:: :s: iij ~ :s: ~ - E m r:: .5 r:: '2 .5 ::::1 r:: .5 ;:) r:: < r:: < < < c) Ord R. m Seed Dispersal d) Ord R. Q Seed Dispersal Mean No. Species Mean % Cover 25 70 (/) (/) 20 60 (I) r:: ... 50 .r:: (I) .2 15 > 0 40 a: 0 (/) 10 30 Cl) '# (,) 20 ~ 5 (/) 10 o~~~~~~~~~~~~ , .... r:: , ... r:: Q) r:: -r:: ~ Si: r:: Q) r:: ~ :;:.. Q) - Q) Si: «< - 0 'l'ii :;:...... - 0 ~ - iij r:: ~ r:: :s: iij ~ :s: m -iij ~ E r:: E c '2 .5 ;:) '2 ::::1 r:: .5c < < < < Figure 7.3: Number of species and % cover of the major seed dispersal types on the Blackwood River (a & b) and Ord River (c & d). Values are means (±SEM) for six sites on the Blackwood River and five sites on the OrdRiver. 151 Wind dispersal was the dominant dispersal mechanism of plants on the Ord River, particularly in terms of cover (Figure 7.3 c&d). Species with an 'unknoWn' dispersal mechanism were the second most common but there was generally a greater spread of dispersal types among species than measured on the Blackwood River. Species with dispersal by frugivorous animals were the second most common in terms of cover for the Ord sites. These included the fleshy fruited Ficus spp. which are well represented at all sites and are fed upon by flying foxes and birds, but similar to many species can also be effectively dispersed· by water. The dominance of wind dispersal is apparent for all sites along the river, particularly at the upstream sites (Sites 03, 04, 05). At the downstream sites, water dispersal as a primary mechanism is well represented. This may be because of the greater number of emergent macrophytes on these sites of permanent water. Similar to the Blackwood River sites, many species, based on their ability to float, can also be dispersed by water, with a mean of 69% of species across all of the sites. Flood debris samples Results of germination tests for flood debris samples on the Blackwood River showed that all but one of the species found in the debris samples were also present in the soil samples (Table 7.1). The exception was the annual herb Atriplex prostrata. Due to large differences in the abundance of species, Czekanowski Similarity Index was low (0.45). When soil samples and debris samples were compared for total number of species and abundance there was no significant difference (p= 0.645 and p=0.711 respectively). For the different life forms, abundance of trees was significantly greater in the debris samples (19±13) (mean±SE) than in the soil samples (1.5±1) (p=0.0009). There was no significant difference between debris and soil samples for any other life forms. Visual examination of flood debris samples identified eight species of woody perennials, with Corymbia calophylla the only species not generally found in the riparian zone (Table 7.4). The majority of species were identified from their fruit which were mainly woody capsules. The two most common riparian species Eucalyptus rudis and Melaleuca rhaphiophylla, were numerically dominant in the debris samples, with large numbers of open fruit. Un-open capsules were dried to 152 release seeds which were counted giving an average of 30.2±10.3 seeds per capsule for E. rudis with 61% viability and 20±8.1 seeds per capsule for M. rhaphiophylla with 12% viability. Table 7.4: Fruit and seed of woody perennial species identified in flood debris samples at three sites on the Blackwood River. Results are tabulated from five 2500 cm3 bulked samples taken from each site in December 1997. Site Habitat Species Plant material Bla riparian Eucalyptus rudis 289 open fruit; 10 seeds upland Corymbia calophylla 19 closed fruit riparian Melaleuca rhaphiophylla 149 open fruit riparian Melaleuca s p 30 open fruit riparian Banksia seminuda 8 fruiting cone fragments B2a riparian Eucalyptus rudis 142 open fruit; 52 unopen fruit riparian Melaleitca rhaphiophylla 159 open fruit B3a riparian Eucaluptus rudis 16 open fruit riparian Casuarina obesa 41 open cones flood plain Eucalyptus occidentalis 3 open fruit flood plain Melaleuca laterifolia 21 open fruit flood plain Melaleuca uncinata 19 open fruit In flood debris samples on the Ord River, 68% of species that germinated were also recorded in the soil samples (Table 7.2), with a similarity index between the debris and soil of 0.22. Comparison of soil and debris samples showed no significant difference for total germinants or for annual herbs or grasses (p>0.05 in both cases). The exception was for woody perennial species which had a significantly greater number of germinants in the debris samples (37±3) compared with soil samples (10±2.8) (p=0.003). Examination of debris samples showed 16 species, 11 of which were common riparian zone species, the other five species are normally found further out on the floodplain (Table 7.5). Species were identified from a mixture of fruit and seeds, with the most abundant species being the common riparian trees at the study sites, Eucalyptus camaldulensis and Melaleuca leucadendra. All the woody capsules from these species were open and had therefore already released seed. 153 Table 7.5: Fruit and seed of woody perennial species identified in flood debris samples at three sites on the Ord River. Results are tabulated from five 2500 cm3 bulked samples taken from each site in Aprill997. Site Habitat Species Plant material 05 riparian Eucalyptus camaldulensis 173 open fruit riparian Melaleuca leucadendra 101 open fruit riparian Ficus coronulata 6 fruit floodplain Acacia sp 1 open fruit pod riparian Lophostemon grandiflorus 10 unopened fruit riparian Nauclea oriental is 3 fruit riparian Terminalia platyphylla 47 fruit riparian Pandanus spiralis 3 seeds flood plain Eucalyptus sp 'bloodwood' 10 open fruit flood plain Eucalyptus chlorophyll a 6 open fruit flood plain Eucalyptus houseana 18 open fruit 04 riparian Eucalyptus camaldulensis 167 open fruit riparian Nauclea orientalis 20 fruit riparian Fiats coronulata 30 fruit riparian Terminalia platyphylla 34 fruit riparian Pandanus spiralis 1 seed riparian Barringtonia acutangula 6 fruit flood plain Hakea sp 1 fruit flood plain Eucalyptus sp 'bloodwood 5 open fruit riparian Melaleuca leucadendra 84 open fruit riparian Lophostemon grandiflorus 24 closed fruit flood plain Eucalyptus gymnotales 122 closed fruits flood plain Eucalyptus argillacea 2 open fruit 01 riparian Melaleuca leucadendra 118 open fruit riparian Eucalyptus camaldulensis 177 open fruit; 10 seed riparian Terminalia platyphylla 11 fruit riparian Lophostemon grandiflorus 5 fruit floodplain Eucalyptus microtheca 6 open fruit flood plain Eucalyptus sp 'bloodwood' 6 open fruit riparian Pandanus spiralis 11 seeds riparian Barringtonia acutangula 3 fruit riparian Nauclea oriental is 2 seeds riparian Ficus coronulata 6 fruit hills lopes I flood plain Adansonia gregorii 1 fruit Seed floating The results of seed floating trials of representative species common in the riparian zone of each river indicated that seeds of the majority of species tested had some floating ability (Table 7.6, Figure 7.4). For the Blackwood River, seed of most species floated for at least five days and up to 30 days for Dodonea viscosa, Juncus pallida and Lepidosperma gladiatum, with seed of some species germinating while floating and then sinking. Some species did not float at all, including 154 Acacia saligna and Exocarpos sparteus. In the case of Banksia seminuda, although seeds did not float at all, the fruiting cone was buoyant. Of the Ord River species, the legume Abrus precatorius did not float at all and Cathormion umbellatu.m seeds did not float themselves but seeds were released in intact pods which were buoyant. The pods started to break up after 20 days and the seeds subsequently sank. Only the seeds of Pandanus spiralis were all still floating after 30 days. For the most common overstorey species on each river seed of all species began to germinate within six to nine days, while floating (Figure 7.4). Germinated seed subsequently sank. Average floating times ranged from 5.6 days for E. rudis to 10.4 days for E. camaldulensis. All seeds remained afloat for at least four days for all species except E. rudis which began to lose buoyancy after one day. Reproductive phenology Seasonal reproductive phenology for Eucalyptus camaldulensis and Melaleuca leucadendra en the Ord River and E. rudis and M. rhaphiophylla on the Blackwood River is summarised for a 12 month period and compared with mean monthly river discharge (Figures 7.5 a&b). For both Eucalyptus species seedfall appeared to be timed to coincide with that period of the year when river water levels were receding, that is, March/ April for E. camaldulensis and September to November for E. rudis. For E. camaldulensis flowering and seed fall occur in a single annual cycle, with fruit maturing and releasing seed and little seed appearing to remain in the canopy. In contrast, E. rudis flowers from August to November with fruit and seed developing over summer. Greatest seed release from these matured capsules is not until the following early summer period. Seed is therefore retained on the tree so that some seed may be released throughout the year with a peak in the early summer period (Figure 7.6). A similar contrast is seen for Melaleuca species with M. leucadendra on the Ord River releasing all seed from February to April, in the period after flowering and fruit development from the previous August to January. M. rhaphiophylla on the Blackwood River flowers in September to November. Fruit development occured over summer and autumn and the woody fruit and seed are 155 retained on the plant for several years. Seed release appeared to be opportunistic after the desiccation of the fruit. Table 7.6: Seed floating ability of representative common species from the Blackwood and Ord rivers. Values are means of five replicates of 20 seeds. Percentage seed floating Da~s 0 5 10 20 30 Blackwood River ]uncus pallidus 100 81 75 58 55 1 seed germinated Lepidosperma gladiatum 100 65 22 19 6 Acacia saligna 0 0 0 0 0 Dodonea viscosa fruit 100 40 0 0 0 Dodonea viscosa seed 100 58 36 36 17 2 seed germinated Rumex crispus fruit 100 61 35 33 0 Rumex crispus seed 100 100 83 74 72 Carex appressa 100 82 73 73 71 8 seed germinated Banksia seminuda 0 0 0 0 0 Exocarpos sparteus 0 0 0 0 0 Ord River Ficus coronulata fruit 100 90 80 20 10 Passiflora foetida fruit 100 43 31 30 6 Passiflora foetida seed 100 83 25 0 0 Pandanus spiralis 100 100 100 100 100 Terminalia platyphylla 100 31 0 0 0 Terminalia hadleyana 100 83 66 66 8 Cathormion umbellatum fruit 100 100 100 41 0 Cathormion umbellatum seed 0 0 0 0 0 Abrus precatorius 0 0 0 0 0 a) Eucalyptus camaldulensis "0 100 100 0----<>--- Cl) 0-<>-- tCl) ----Cl) 11) -~0 Cl) 80 20 -~0 80 20 0) - 0) -c t: I: :;:::; 0 c 60 mean: 10.4::t3 40 0 60 40 :;:::: :;:::; ~ ea median: 11 0 ~ ~ -c 0 40 mode: 10,11 60 c u: 40 60 u: E range: 7-20 E,_ ,_ -??. Cl) 20 80 Cl) 20 80 -0~ G G - - 0 100 0 100 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 c )Eucalyptus rudis d) Melaleuca rhaphiophylla "0 100 0-<>-- •"0 100 o---o------Cl) mean: 5.6::t3 Cl) Cl) ~ Cl) ~ en 80 median: 6 20 -0 00 80 20 -0 mode: 5,6 0) - C) - c t: c 60 range: 2-10 40 0 c 60 40 0 .... :;:::: :;:::: ea ea I'll ltl 0 40 60 t: -0 40 60 c u. 1.1.. ,_E E 20 80 Cl) 20 80. ~ ~ ~ -0 G -0 ~ - 0 100 - 0 100 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 No. of Days No. of Days Figure 7.4: Floating duration of seed of two major overstorey species on the Ord River (a & b) and the Blackwood River (c & d) and the percentage of seed that germinated while floating. Values are the mean percentage of five replicates of 50 seeds. Mean, median, mode and range are for the duration m days that seed remained floating. >-' (JJ Cl'\ 157 (a) Ord River Melaleuca Jeucodendra Seedfall Flower development . Fruit development <::t ...... )::> ~ ______-~------flo..~v_emJg~ __ -):> Eucalyptus camaldu/ensis Flower development Fruit development 5000 .,._____ S:...:e;.:..ed::.:.fa=ll~~~ ...... <~~~~t- ..-.- ..-.- ..-1~~ ..' Q) ...Cl 4000 .cIll (,) Ill i5 3000 :c;::.. 1: ...... ,2000 0 0 ::i: Q) (/) ~(;;1000 ::i:...,_Q) E Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec (b) Blackwood River . Melaleuca rhaphiophylla Fruit development Flower development . <::!: ______-)> lill!lo ~ower developmenl.. Flowering Eucalyptus rudis 1111111 200 ...... ''<~~~<1!11111---.::::S::::.:ee::.::d::.::fa::.:.ll_-1~~ Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 7.5: Relationship between phenology of selected riparian species with monthly river flow levels for the (a) Ord River and the (b) Blackwood River. 158 180 160 140 -gen 120 Cl) 100 en -0 80 ci z 60 40 20 0 ... ;:::...... > c ... > c m 'S c. 0 m m 0 m m '"':I Cl) '"':I :lE '"':I I z I I z I :lE :lE I I I .,!. (.) (.) c enI .c .... .c ... :::J (.) Cl) Cl) (.) Cl) Cl) c. Cl 0 u. <( '"':I :::J 0 c u. c <( 1996 1997 1998 Figure 7.6: Mean monthly number of seeds(± SE) of Eucalyptus rudis in seedfall traps averaged over 15 traps at three sites. 159 Fate of seed On the Blackwood River, after seed burial for one month 7~±10% (mean±SE) of buried seeds of E. rudis and 79±9% of M. rhaphiophylla seeds germinated. After six months of burial 19±12% of E. rudis and 0% of M. rhaphiophylla seed germinated and by 12 months no further seed of either species germinated. For Ord River species, after six months of burial 4±1.5% of E .camaldulensis seed and nil M. leucadendra seed germinated. No germination of these species was recorded in the non-seeded soil cores on either river. In ant feeding stations, after eight hours, all seed of E. rudis and M. rhaphiophylla had been removed from 60% of the feeding stations and no seed had been removed from the other 40% of stations. This indicated that seed is efficiently harvested once it has been located. All seed of E. camaldulensis had been removed from 40% of feeding stations after eight hours and from 90% of stations after 24 hours. All seed of M. leucadendra had been removed from 20% of feeding stations after 24 hours. Discussion The results show the potential for regeneration from seed of riparian species on the Blackwood and Ord rivers. They indicated that in the riparian zone regeneration of vegetation from soil seedbanks is probably important for annual species of herbs and grasses but of only minor significance for perennial species. For perennial species, particularly overstorey species, direct seedfall from existing vegetation is important and the enhanced dispersal by floating downstream and up-slope with flood debris is a consequential recruitment mechanism. Reproductive phenology of the four species monitored in this study appears to be well adapted to the hydrological regimes on the respective rivers. Riparian soil seedbank Recruitment from a soil seedbank is limited for perennial species and is probably only important for ephemeral species. These ephemeral species tend to be ruderals which are adapted to high 160 levels of disturbance and possess mechanisms such as seed dormancy which allow them to persist in the soil seedbank (Grime et al. 1986; Thompson 1987). Overstorey species of the Ord and Blackwood rivers were not present in any number in the soibeedbank in this study. This was also reported for a riverine swamp in the eastern United States (Schneider & Sharitz 1986). Seed predation and limited longevity of seed may be the main reasons for this, at least for the species looked at in the study. Ants have been reported to harvest large proportions of Eucalyptus seed in forests (Ashton 1979), in woodland (Yates et al. 1995) and in mallee communities (Wellington & Noble 1985a). Ant activity is usually seasonal (Andersen 1989) and the effects of seed predation may have less impact during times of massive seedfall (mast seeding), when predator satiation can occur Ganzen 1971). Seed of E. salmonophloia has been reported to have limited viability in the soil (Yates et al. 1995), as was seen here, and therefore limited ability to enter the soil seedbank. In evolutionary terms, the cause for the dominance of annual species in the soil seedbank may be, whereas there are other opportunities for regeneration of long-lived perennials, the seed of annual herbs and grasses provide the only opportunity for regeneration. Dormancy of seed ensures germination only under favourable conditions (Sclmeider & Shari tz 1986; Thompson 1987). The limited number of perennial species represented in the seedbank indicated that regeneration of these species does not come from the seedbank alone, if at all. Numbers of seed found in this study (1318 m·2 for the Blackwood River sites and 652m2 for the Ord River sites) is within the 2 same range of other studies of a riverine swamp (600- 3500 m· ) (Schneider & Sharitz 1986) and 2 jarrah forest (767 m· ) (Vlahos & Bell 1986). The elevated values are mainly due to the high numbers of annual species which are generally exotic weeds on the Blackwood River. This correlates with the high numbers of these species in the vegetation (see Chapter 4) which would enable the build-up of large numbers of seed in the soil. The transient nature of the soil, particularly close to the river, may explain the gradient of increased seed numbers in the soil seedbank away from the river. For the Ord River, the greater numbers of grass seed with distance from the river reflects the dominance of grass in the understorey vegetation as the conditions become more terrestrial and therefore less influenced by the river. The greater number 161 of tree seeds may indicate the importance of river flow for the dispersal of tree seed. Samples of the soil seedbank were taken at only one time during the year and these results therefore may reflect both the permanent and transient seedbank. To di~tiriguish between the permanent and transient seedbank, samples would need to be taken at various times throughout the year (Gross 1990; Thompson 1992). The time of year that samples in the present study were taken would most likely coincide with the greatest number of seeds in the soil, particularly the overstorey species which have a limited viability. Results showed a high degree of variability of numbers of seeds in samples, highlighting the need to take many small samples to adequately survey the soil seedbank (Brock et al. 1994). The nature of seed dispersal would suggest a clumped distribution of seed stored in the soil (Thompson 1987). Many species can reproduce vegetatively through rhizomes, lignotubers and stolons, etc (James 1984), which allow them to expand in an area or recover from disturbance. For the overstorey species examined here, although they can recover from defoliation or partial destruction of above-ground biomass through vegetative resprouting, colonisation of a new area usually requires establishment from seed. However, some species (e.g. Salix spp.) possess vegetative propagules that can be effectively dispersed downstream in river flows. An important function of vegetative propagation for species is therefore to allow persistence at a site after a disturbance, through rapid expansion and colonisation of new substrate (van der Valk 1992). The rapid recovery of established resprouters gives them a competitive advantage over seedlings and consequently resprouting species may dominate the post-disturbance habitat (Malanson & O'Leary 1982). Many species that occur in riparian habitats in North America, such as Salix spp. and Populus spp., possess the ability to resprout and produce large quantities of seeds Gohnson 1994), as do some of the dominant riparian trees on the study rivers, such as Eucalyptus camaldulensis and E. rudis. Dispersal mechanisms The river is important for dispersal of diaspores, despite wind dispersal being the most common mode of dispersal on both rivers. Wind-dispersed seeds are not only those with obvious 162 morphological features, such as the winged seeds of Casuarina obesa (Blackwood River) and Lysiphyllum cunninghamii (Ord River) or pappus of many Asteraceae species, but also small seeds that can be effectively blown reasonable distances (Malanson 1993), such as Eucalyptus and Melaleuca species. Most species with a primary form of dispersal such as wind or by animals are also capable of secondary dispersal by water (Sauer 1988; Thebaud & Debussche 1991). The distance of dispersal depending on floating capacity, although non-floating seed can be 'rafted' on other debris (Nilsson & Grelsson 1990). Water dispersal may be particularly important for rapid movement of seed downstream during floods and over much greater distances than would be achieved by wind dispersal alone. Water dispersal can also move seed up-slope and tend to concentrate seed in pools and depressions where moisture levels will be relatively high. Few species have diaspores with obvious structures for water dispersal, examples include the introduced vine Cardiospermum halicacabum on the Ord River and Atriplex prostrata on the Blackwood River. In fact, species with particular structures for water dispersal are considered rare (Willson et al. 1990). On a riparian meadow site, water dispersal was considered the only source of seeds not already present (Skogltmd 1990). Species with fleshy fruit or inflorescence such as Ficus spp. on the Ord River or Rubus discolor on the Blackwood River are buoyant in water and can be effectively dispersed in this way. Indeed, the inflorescence of Ficus spp. have been observed to be eaten by fish and to have their seed dispersed in this manner (Malanson 1993). Many other species can have features that would also aid water dispersal such as hairs (Herrera 1991) or wings, as in Dodonea viscosa and Casuarina obesa. Many plants simply have light seeds with a large surface area to volume ratio, such as Eucalyptus and Melaleuca spp .. On the Ord River, there was a greater variation of dispersal types, with many species having diaspores dispersed by animals, both internally and externally with external dispersal by vertebrates being common in riparian zones in disturbed and grazed habitats (Sorensen 1986). Riparian species rarely rely entirely on the river for the spread of diaspores. However one species that appears to is Cathormiom umbellatum, which only occurs fringing the river and has buoyant pods which dehisce, usually into or near the water, with large non-buoyant seed intact and only releasing the seed after decomposition of the pod. Barringtonia acutangula is also 163 restricted to the water's edge and has large buoyant fruit. No examples of species relying on the river for dispersal were obvious on the Blackwood River. Therefore, although water dispersal can be important for the rapid movement of propagules downstream and over long distances (Nilsson et al. 1993), it is not essential for the majority of species. No longitudinal trend in dispersal modes was seen on the study rivers as was observed by Weaver et al. (1925) for a Nebraska river in the USA. This may be due to river morphological processes and species physiological traits also influencing species distribution. Dispersal by water in debris and seed floating Flood debris is shown here to be an important source of propagules, at least for some of the perennial species on both rivers, with a significant amount of seed in debris samples and significantly more than occurred in the soil samples. Flood debris is seen to have an important ecological role in the colonisation and establishment of riparian vegetation (Naiman et al. 1998) through the accumulation of sediment and the transport and concentration of propagules. The occurrence of seed in debris also indicates the importance of secondary dispersal of seed by water. This has been reported for other riparian habitats (Schneider & Sharitz 1988; Johansson et al. 1996). Although seed floating ability is probably advantageous it does not have to be for very long as debris movement downstream in a flooded river is rapid (Nilsson et al. 1997) and seed may be carried on more buoyant debris (Nilsson & Grelsson 1990) or in buoyant fruits. Visual examination of the flood debris samples showed that the majority of species present are common riparian inhabitants, although several terrestrial species were also present. These species would enter the riparian system through overland flow. Floating seed of the studied overstorey species germinates readily while floating and then lose buoyancy as the hypocotyl develops. This suggests that for successful establishment of seedlings, seeds need to be washed onto near shore sediments or left along debris lines when water is rapidly receding, soon after seed fall. Reproductive phenology For the four species examined here on the Blackwood and Ord rivers there is a good correlation between reproductive .phenology and hydrology. In particular, seed fall appears to coincide 164 with the end of the rainy season when river water levels start to recede, exposing moist bare sediment ideal for germination of seed. Reproductive phenology of the four species reflects their adaptation to their environment. Maturation of flowers and seed takes place within a couple of months in many tropical eucalypts (Eldridge et al. 1993; Brooker & Kleinig 1994), as opposed to more than two years in some southern eucalypts (Ashton 1975). The period of greatest seedfall for E. camaldulensis, M. leucadendra and E. rudis is timed to coincide with receding floodwaters when moist sediments favourable for seed germination and survival are exposed. This adaptation has been reported for riparian species of Populus and Salix in the central United States (Bradley & Smith 1986; Johnson 1994) and in the south-eastern United States (Schneider & Sharitz 1988). Indeed, phenological events in many species are thought to be timed to abiotic conditions (Estabrook et al. 1982). The Ord River species release the majority of seed over a short period of time at the end of the wet season which provides their only opportunity for recruitment for the year. Other northern Australian Eucalyptus spp. have been reported to display a similar reproductive phenology (Bowman et al. 1991; Fensham 1992; Setterfield & Williams 1996). The small seed of these species do not survive long in the soil and therefore do not become incorporated into the soil seedbank. Therefore post-dispersal conditions may limit recruitment. E. rudis on the Blackwood River retains mature seed in the canopy for at least 12 months. Although there is a time of peak seed release, some seed is released all year round so there is potential for germination of seed whenever conditions are favourable, and for continuous recruitment into the stable community. For many temperate eucalypts seed maturation can take several months and mature seeds are retained in the capsules while a continuous low level of seed release occurs, which is accelerated by dry conditions (Cunningham 1957; Wellington & Noble 1985a; Davies & Myerscough 1991). Interestingly, reproductive phenology of E. camaldulensis occurring on the Murray River in south-eastern Australia is similar to that documented here for E. rudis on the Blackwood River (Dexter et al. 1986). These rivers have similar fluvial regimes and climatic conditions. M. rhaphiophylla retains mature fruit on the tree and appears to only release seed with the desiccation of the woody fruit. This can happen after disturbance such as flooding or during strong winds when branches are commonly broken off and carried downstream onto point bars. This may result in the release of seed when conditions are most favourable for 165 successful germination, such as onto bare sediments when floodwaters are receding and away from competition of a understorey of sedges and dense canopy cover. Large floods will also create gaps in the vegetation. Fire may also result in the release of seed of Melaleuca spp. in wetlands (Froend et al. 1993). These results of reproductive phenology are summarised from over only three years of observation and are therefore tentative. There is generally wide variation in annual fruit and seed production in Eucalyptus species caused by such limitations as amount of resources available for flower and fruit development, seasonal weather conditions, predation and fire (Ashton 1975; House 1997). It has been reported that years of heavy seed production are commonly preceded by several years of low seed production (Burrows & Burrows 1992). Therefore, to confirm the results presented here requires long term monitoring. The overstorey species on the Blackwood River (E. rudis and M. rhaphiophylla) exhibited some degree of a canopy seedbank (serotiny), particularly M. rhaphiophylla which appears to retain seed on the plant for a number of years. For the Ord River species, in contrast, only E. camaldulensis showed limited serotiny. A large proportion of species in the Myrtaceae family which have woody fruits retain large stores of seed in the fruit for several seasons, releasing seed only after desiccation through fire or death of a limb (Beardsell et al. 1993). Eucalyptus spp. in general show a varying degree of serotiny. However eucalypt seed is rarely held in the canopy for more than three or four years (House 1997). This Chapter has shown the mechanisms which are important for the storage and dispersal of seed in the riparian zone. Using the four common overstorey species, that is E. rudis and M. rhaphiophylla (Blackwood River) and E. camaldulensis and M. leucadendra (Ord River) the following two chapters will explore the conditions which are necessary to allow successful germination of seed, the establishment of seedlings and their subsequent recruitment into the population. 166 CHAPTERS SEEDLING RECRUITMENT OF SELECTED RIPARIAN SPECIES Introduction After considering sources of seed (i.e. seedbank, flood debris and dispersal) in the previous Chapter, this Chapter focuses on the factors that influence the establishment of seedlings of riparian species and therefore their regeneration niche (see Grubb 1977). These factors include the effects of temperature on the germination of seed and of microsite scale environmental conditions on young seedlings as well as the resilience and growth of young seedlings under environmental extremes of flood and drought. In addition, because of the influence of rising salinity levels on riparian plants on the Blackwood River, especially in the upper reaches, the effect of salinity on germination of seed and survival of seedlings will also be examined. The two most common overstorey species from each river system (identified in Chapter 4 and described in Chapter 7) have been chosen for this study. That is, Eucalyptus rudis and Melaleuca rhaphiophylla on the Blackwood River and E. camaldulensis and M. leucadendra on the Ord River. Temperature and moisture play a critical role in providing the stimulus for seed germination and this is linked to climatic conditions, with many plants evolving to germinate when conditions are most favourable for survival (Mott & Groves 1981). Seed dormancy has evolved to delay germination until environmental conditions are such that they provide the best chance of survival (Thompson 1992). In many cases germination is cued to temperature and therefore for most species there is an optimal temperature above and below which germination is delayed but not prevented (Bell e(al. 1993). For eucalypts in southern Australia constant temperature (range 167 15° - 25° C) gives enhanced germination than alternating temperatures and most species have a germination requirement for light (Boland et al. 1980). Species from northern Australia may have higher temperature requirements for germination (around 30° C). Species from environments where soil moisture is available longer into the dry season, such as the riparian zone or wetlands, may also have higher optimal germination temperatures (Bell & Williams 1997). Species of Eucalyptus and Melaleuca generally have some species-specific tolerance to saline conditions during germination but are generally much more sensitive than halophytes (van der Moezel & Bell 1987). Rising salinity levels in the Blackwood River may inhibit germination of these important species in the riparian zone. Successful establishment of seedlings following germination can be greatly influenced by position in the landscape (Bell & Williams 1997). Microsites which can retain moisture and maintain high humidity provide the right conditions for successful germination and early establishment of seedlings (Battaglia & Reid 1993; Pollock et al. 1998). The occurrence of elevated microsites which provide such conditions can affect the distribution of seedlings in flooded riparian wetlands (Titus 1990; Bren 1992). Soil type is also important in terms of compaction and bulk density, moisture retention and nutrient availability. Litter, even small amounts, has been shown to reduce seedling survival rates (Fowler 1988) and the interaction of flooding and litter accumulation can affect plant establishment in the riparian zone (Nilsson & Grelsson 1990). Competition, particularly with grasses, has been reported as influencing growth and survival of seedlings (Withers 1978; Noble 1984; Davidson & Reid 1989) and livestock grazing has been suggested to improve E. camaldulensis seedling survival by reducing competition with grasses (Dexter 1978). The previous Chapter showed that timing of seedfall for three of the four study species occurred to coincide with receding river water levels and it was postulated that this was to allow seed to fall onto moist sediments which are conducive to survival of germinating seed. This Chapter will examine the fate of germinating seed and seedlings in this and other areas of the riparian zone. 168 In the riparian zone of the Ord and Blackwood rivers, high flows occur regularly almost every year. Therefore establishing seedlings will be subjected, depending on their location, elevation and level of flooding, to extended periods of waterlogging, inundation or total submergence. The flooding tolerance of plant species varies widely and is important in determining species composition and distribution (Kramer & Kozlowski 1979). Flooding effects include suppression of transpiration and photosynthesis and a subsequent reduction in growth (Emst 1990; Blom & Voesnek 1996). Responses to flooding include formation of adventitious roots, stem elongation and formation of aerenchyma cells in the roots of flood tolerant species (Blom & Voesnek 1996). Establishing the degree of flood tolerance of the selected species in this study will indicate their potential distribution on these rivers. This Chapter examines the ecological factors that influence seed germination and the survival and growth of seedlings in the particular riparian environments of the Ord and Blackwood rivers. It examines what are the optimum temperatures required for the selected species on each river and the rate of germination of these species at various temperatures that may be experienced throughout the year. It investigates conditions that are likely to allow successful early establishment of seedlings and, for the selected species, tolerance to various degrees of flooding. The previous Chapter examined the sources of seed of riparian species which enable their recruitment into the riparian zone and this chapter takes the next step to examine the conditions necessary for the germination of this seed and the establishment and growth of resulting seedlings. Methods Seed germination Seed used in germination trials was supplied by the Department of Conservation and Land Management seed store (Manjimup) for Blackwood River species (Eucalyptus rudis, Melaleuca rhaphiophylla and M. cuticularis) and by a commercial seed supplier (C. Dupe, Kununurra) for 169 the Ord River species (Eucalyptus camaldulensis ssp. obtusa and Melaleuca leucadendra). For the experiment on the effects of salinity on germination, seed collections of E. rudis were from two localities, one from an area with freshwater conditions (Narinup) and the other from a saline environment (Narrogin). Seed of M. cuticularis was also tested as this species tends to replace M. rhaphiophylla in saline areas. For the temperature effects experiment, seed of E. rudis was from the Collie River in the catchment adjacent to the Blackwood River catchment where water quality is similar to the Blackwood River in being marginal to brackish. In the trial on temperature effects on germination, seed from the four target species were tested (E. camaldulensis and M. leucadendra from the Ord River, E. rudis and M. rhaphiophylla for the Blackwood River). Twenty replicates of 50 seeds of each of the four species were placed in 9 an petri dishes that contained vermiculite and a Whatman No. 4 filter paper which were wetted 1 up with 25 ml deionised water and given a light spray of fungicide (Benlate 1g L- ). A replicate of each species on the petri dishes was placed in a moistened plastic bag which was sealed to reduce evaporation and five replicates of each species were placed in growth chambers at 15° C, 20° C, 25° C and 30° C. Chambers were in 12 hours continuous light (2500 lux) and 12 hours dark and were monitored daily over 20 days. A regime of continuous temperature was chosen as studies have shown that for many species alternating temperatures have shown no positive effect and the diurnal temperature range is adequately represented by the mean (Mott & Groves 1981). Seed was considered to have germinated when the radicle had emerged and as germinants were counted they were removed. Seed that did not germinate was dissected and visually examined and almost all of this material did not contain a developed embryo inside the seed coat. This indicated that germination levels reflect, reasonably well, seed viability for each species. Total germination after 20 days for each treatment and species was presented and the effect of temperature was tested using a one-way analysis of variance and Fischer's LSD post hoc test (Zar 1984). Initial germination speed (1/Tmt) for each species and temperature was calculated as the time in days for at least one seed to germinate from any replicate. Median germination speed (1/T50%) was the time for at least 50% of viable seed to germinate (Lush et al. 1984). ------ 170 The germination trial on the effects of salinity on germination was restricted to the Blackwood River species including two ecotypes of E. rudis, M. rhaphiophylla and M. cuticularis. The same procedure was used as for the temperature trial except that growth chambers were set at a constant 25° C and the twenty replicates were watered with one of four levels of salt 1 1 1 1 concentration: 0 mScm- (i.e. deionised water only), 10 mScm- , 20 mScm- or 30 mScm- • Saline solutions of the desired concentrations were achieved by adding NaCl into a beaker of deionised water until the desired electrical conductivity was obtained. Seedling establishment To examine where within the riparian zone seedlings are most likely to establish, seedling plots were set up and environmental parameters that may be important for establishment of seedlings were recorded. At Sites 04 and 05 on the upper Ord River (in the unregulated section) and on the middle section at Sites B1 and B2 on the Blackwood River (see Figures 3.1 & 3.2) six 200 metre line transects were set up perpendicular to the river. One metre square seedling quadrats were laid out adjacent to the transects where seedlings occurred. Within each quadrat heights of all one year old seedlings were recorded. For each quadrat a number of environmental parameters were recorded: surface soil type using field texture analysis (McDonald & Isbell 1984); four categories of microtopography where plots occurred, including depression, flat area, slope and rise; surface soil moisture status with four categories including dry, moist, waterlogged and inundated; elevation above river base flow level measured along the transect with an automatic level and staff; a visual estimate of% bare ground and% litter; distance of seedling plot from river base flow level; and distance to nearest seed tree. Frequency of seedling plots for each environmental parameter was collated. For quantitative parameters (elevation, % litter, distance to water and distance to nearest seed tree) correlation analysis against seedling density was performed. For each categorical parameter (soil type, microtopography and surface soil moisture status) a Krusall-Wallis test (non-parametric ANOVA analogue) was conducted for seedling density to test if there were significant differences between categories for each parameter and seedling density or number of quadrats. A non-parametric 'Tukey-type' multiple 171 comparison test was carried out to separate differences between category means (Zar 1984). As there were unequal numbers of quadrats in the categories for each parameter, the five seedling density scores around the median score were used for the subsequent analysis. In addition to these plots, to measure the survival and growth of seedlings, transects were set up in areas close to the river where germination had recently taken place. On the Blackwood River, seedling plots were located opportunistically in an area at Site B2a where a large germination event had taken place along the river edge in exposed moist sediment from the receding winter high water levels in November 1996. At this site three belt transects were set up perpendicular to the river from the water level to the top of the bank. The transects were 10 metres long and one metre wide and were divided into 10 contiguous 1 m2 plots. In each plot all seedlings of E. rudis and M. rhaphiophylla were counted and seedling height was measured. Elevation along the transects was measured, using an automatic level and staff, as well as distance from the water. At 0 m, 4 m and 8 m soil samples were taken to measure gravimetric soil moisture percentage and electrical conductivity using a 1:5 soil deionised water solution (Klute 1986). Transects were measured every two months for a period of 48 weeks. On the Ord River, seedling transects were located on a sediment island and on a depositional area of the river on the inside of a bend. Transects were located at Site 01 and Site 04. Belt transects 10 m long and 1 m wide were divided into 10 contiguous 1 m·2 plots. In each plot a 11 seedlings of M. leucadendra were counted and seedling height was measured. Elevation along the transects was fairly flat and at 0 m, 4 m and 8 m from the water soil samples were taken to measure gravimetric soil moisture content(%). From flood debris marks in the nearby vegetation and from recorded river discharge levels for the measurement period, seedling plots were inundated to a level of at least one metre at both sites for approximately three months during the wet season (December to March). Transects were measured four times, in August 1996, April 1997, December 1997 and April1998. 172 Effe<:ts of flooding A flooding experiment was set up to test the effect on survival and growth of various levels of flooding on seedlings of the four target species. One hundred and forty four seedlings of each species were raised in a commercial nursery for six months. They were transferred to 380 cm3 volume pots with a commercial potting mix, sealed at the top with a clay plug and allowed to settle in for three weeks prior to commencement of the experiment. Forty eight seedlings of each species were assigned to one of four treatments: 1) control- pots were free-draining and watered daily; 2) waterlogged - pots were sat in water to ~ 1 cm below the top of the pot so that soil was permanently waterlogged; 3) flooded - pots were placed in water to a depth of 20 cm so that approximately the lower one third of each seedling was under water; 4) submerged - pots were placed in water to a depth of 55 cm so that seedlings were completely under water. Three plastic tubs (600 mm x 500 mm x 600 mm) were used for each treatment giving three replicates of 12 seedlings of each species per treatment. Tubs were arranged on three levels so that water circulated through all the tubs and was then pumped back up to the top tub. Tubs were also orientated so that the duration of sunlight and shade in all tubs was similar. At the commencement of the experiment, and after 4, 8 and 12 weeks, one seedling of each species was removed from each tub. Plant moisture status was assessed by measuring shoot xylem pressure potential using a Scholander pressure bomb (Scholander et al. 1964) as an indication of the physiological effects of flooding. Seedlings were oven dried at 50° C for 24 hours and roots and shoot was weighed to give a measurement of dry weight total biomass and also root:shoot ratio. Top height of all seedlings was also measured, every two weeks, from soil surface to the top of the plant and the occurrence of any adventitious roots was recorded. Relative growth rates were calculated over the 12 week period as: RGR = logeW2 -loge W1 t2 - tl where W1 and W2 are the total dry weights at the beginning and end of the sampling period and t1 and t2 are the dates of sampling (Kramer & Kozlowski 1979). This was done for each species in each treatment to allow for the variable growth patterns of the different species (Bidwell 1974). Differences between treatment means were tested using a Kruskal-Wallis non~parametric 173 ANOVA analogue and the 'Tukey type' non-parametric multiple comparison (Zar 1984). water temperatures in the tubs ranged from 14° C in the morning to 22° C by mid-afternoon at the beginningoftheexperiment (4th September 1997) to 17°C to 26° Cat the end of 12 weeks. Day time dissolved oxygen levels in the tubs were measured weekly with a portable dissolved oxygen probe averaging 8.4 ± 2.4 mg 1-1 for the duration of the experiment. Results Seed germination Total germination of seeds was significantly affected by temperature for only Eucalyptus rudis (p=0.01), with a decrease in total germination at 15° C (Figure 8.1). However, rate of germination was affected by temperature for all species (Figure 8.2). For E. rudis, initial germination was rapid (one day) at 20,25 and 30°C and for the higher temperatures at least 50% of seed germinated within this time. For the other species there was a delay fl'Orn initial lo median germination at all temperatures. Of the temperatures examined, 25° and 30° C were the optimum temperature for rapid germination in all species. The effects on germination of increasing levels of salt concentration for Blackwood River species showed a significant reduction in total germination with increasing salinity of water in the petri dish (Figure 8.3). For all species, germination was not affected at 10 mScm-1 but a significant reduction in germination occurred with each subsequent increase in salt concentration. There was no significant increase in salinity tolerance with seed from saline environments for E. rudis or M. cuticularis. Germination rates were also affected by increasing salinity (Figure 8.4), with time to 1 initial germination increasing with increased salinity even at 10 mScm- • 174 100 90 80 .... 70 !: (!) 60 ...Q (!) 50 ll. 40 30 20 10 0 Eucalyptus Eucalyptus Melaleuca Melaleuca rudis camaldulensis leucadendra rhaphiophylla Figure 8.1: Total germination percentage of seed subjected to four constant temperature treatments in growth cabinets. Species included E. rudis and M. rhaphiophylla from the Blackwood River and E. camaldulensis and M. leucadendra from the Ord River. Figures are the means(± SEM) of five replicates. 175 Ill 1/Tint ...... 1/TSO% a) E. camaldulensis b) M,. leucadendra 1 1 0.9 '"C 0.9 '"C Cl) Cl) Cl) 0.8 Cl) 0.8 0.. a. 0.7 11) 0.7 11) 0.6 0.6 c c 0 0.5 ...... 0 0.5 «< 0.4 •.. ····· «< 0.4 .. ..· c -c 0.3 - 0.3 ...... E 0.2 ...E 0.2 Cl) •...... -············ Cl) ...... ·········· C) 0.1 C) 0.1 0 0 10 15 20 25 30 10 15 20 25 30 Temperature (oC) Temperature (oC) c) E. rudis d) M. rhaphiophylla 1 1 0.9 '"C 0.9 ...... '"C Cl) Cl) 0.8 Cl) 0.8 ./ Cl) a. 0.7 ... a. 0.7 11) ...... 1/J 0.6 0.6 c c 0 0.5 •••• .2 0.5 .·· _ «< 0.4 ... ·· «< 0.4 .. ··· -c .. c ...... ·········· 0.3 ...... - 0.3 ....E 0.2 ...E 0.2 Cl) Cl) C) 0.1 C) 0.1 0 0 10 15 20 25 30 10 15 20 25 30 Temperature (oC) Temperature (oC) Figure 8.2: Speed of germination of the four selected species at different temperatures with initial germination speed (1/T int) whereTint is the time taken in days for at least one seed to germinate and T50% is time taken for 50% of the total seeds to germinate. 176 100 90 80 .... 70 ~ 60 u :D 50 ll. 40 30 20 10 0 Eucalyptus Eucalyptus Melaleuca Melaleuca rudis rudis cuticularis rhaphiophylla (Nannup) (Narrogil1) 0 0 mS/cm lil 10 mS/cm [21 20 mS/cm 30 mS/cm Figure 8.3: Total germination percentage of seed watered with four different salt concentrations for species occuring on the Blackwood River including two ecotypes of E. rudis from freshwater area (Nannup) and saline area (Narrogin). Figures are the means(± SEM) of five replicates. Within each species, treatment means with the same letter are not significantly different (Fisher LSD, p = 0.05). 177 11 1/Tint ...... 1/T50% a) E. rudis (Nannup) b) E. rudis (Narrogin) 1 1 0.9 "C 0.9 \ "C 0) 0.8 0) 0.8 0) \ ...\ 0) c.. 0.7 c.. 0.7 Ill Ill 0.6 \ .•.. 0.6 1: 1: 0 0.5 •···... 0 0.5 .... ·. .... 1\'1 0.4 ··... 1\'1 0.4 •···••····•····· 1: 1: ... 0.3 ~ ...... 0.3 ...E 0.2 ...E 0.2 ··...... 0) ...... 0) C) 0.1 C) 0.1 0 0 0 10 20 30 0 10 20 30 Salt Concentration (mS/cm) Salt Concentration (mS/cm) c) M. rhaphiophylla d) M. cuticularis 0.5 0.5 "C "C :R 0.4 :R 0.4 c.. c.. Ill Ill 0.3 0.3 1: 1: 0 ....0 m o.2 ...... 1\'1 0.2 ...... 1: ...... 1: ...... E... 0.1 ·············· ...E 0.1 (I) (I) C) C) 0 0 10 20 30 0 10 20 30 Salt Concentration (mS/cm) Salt Concentration (mS/cm) Figure 8.4: Speed of germination of four selected species at different salinities (mS/ cm) with initial germination speed (1/T int) whereTint is the time taken in days for at least one seed to germinate and T50% is time taken for 50% of the total seeds to germinate. 178 Seedling establishment The majority of the Ord River seedling plots had loarri to sandy soils (Figure 8.5a). When seedling density was compared for the different soil types, numbers of seedlings per m2 was significantly higher in loam, loamy sand and sandy soils than in the clay and gravel soil types (H=15.87, p= 0.003). Frequency of plots was similar for the different microtopography sites with slightly more plots occurring on the river flat area but analysis of seedling density showed that plots in depressions, slopes and rises had significantly more seedlings per m-2 (H=13.28, p=0.004). Moist and dry sites had the greatest frequency of seedling plots but waterlogged sites had significantly greater density of seedlings than the other moisture-related parameters (H=9.74, p=0.02). There was no significant co~relation between seedling density and distance to nearest seed tree, with a large group of plots in close proximity and another group at some distance (>200 m) from a seed tree. There was also no significant correlation between seedling density and percentage litter or percentage bare ground. Frequency of plots with 0-20% litter and 0-20% bare ground was much greater than for the other cover classes (Figure 8.5). There was a strong negative relationship between seedling density and elevation (r= 0.875, p=O.OOl), with most seedlings occurring at lower elevations (Figure 8.6a). There was, however, not a strong relationship between elevation and distance from the river (r=0.284, p=O.lO). The majority of Blackwood River seedling plots were loamy sand and sandy soils (Figure 8.5b ). When seedling density was compared for the different soil types, numbers of seedlings per m2 was significantly higher in loam, loamy sand and sandy soils than in the clay soil and no plots occurred in gravel soil (H=10.99, p= 0.011). Frequency of plots for microtopography showed more than 50% of plots occurring on the river flat, but analysis of seedling density showed river flat, depression and rise had significantly greater seedlings per m2 than plots on a slope (H==11.66, p=0.0086). Moist sites had the greatest frequency of seedling plots with wet and moist sites having a significantly greater density of seedlings than did dry sites (H=10.47, p=0.0053). There was a significant correlation between seedling density and distance to nearest seed tree (r==0.470, p= 0.048) with most plots in close proximity to a seed tree and no plots >30 metres from a seed 179 tree. There was no significant correlation between seedling density and percentage litter or percentage bare ground. Like the Ord River plots, frequency of plots with 0-20% litter and 0-20% bare ground was much greater than for the other cover classes (Figure 8.5b). There was a strong negative relationship between seedling density and elevation (r= 0.758, p=0.001}, with most seedlings occurring at lower elevations (Figure 8.6b). There was also a significant relationship between elevation and distance from the river (r=0.815, p=O.OOl). After initial germination of seed of E. rudis and M. rhaphiophylla in November 1996 seedling numbers declined along the seedling transects at Site 3 on the Blackwood River (Fig1,1re 8.7&8}. For E. rudis, seedling numbers were greatest in the middle of the transect where conditions were moist but not waterlogged. Greatest loss of seedlings was over weeks 16 to 32, the time of highest salinity and lowest moisture (Figure 8.9). There were few seedlings at seven to eight metres and seed that did germinate rapidly died. There was steady growth of seedlings over the period, with rapid growth from week 40 to 48 with warmer temperatures and adequate moisture, except in the 0-2 m quadrats which were inundated so growth was negligible and some death occurred. M. rhaphiophylla had higher survival and better growth in plots closest to the water, although more seedlings initially established in the middle plots. The first quadrats (0-1 m) were underwater for 3.5 months (July to October) up to a depth of 0.5 m while the second quadrats (1-2 m) were inundated for two months (August -September). Moisture content was initially high before dropping to lowest levels in March-April after 16 weeks, then rising steadily with winter rains and rising river levels (Figure 8.9). Surface soil electrical conductivity along the transect peaked in May after initial rains following the summer drought period (when salts would accumulate at the soil surface through evaporation) but subsequently reduced after more substantial rains flushed out the salt. 180 (a) Ord River 100 90 80 70 Ill 60 .2- a. 50 0 - 40 (ft 30 20 10 0 Soil type Micro Moisture % bare % litter topography ground (b) Blackwood River 100 90 80 70 2 .2 60 a. 50 -0 40 30 20 10 o~uw~~~--~~~~--~~.Ua---~UM~~~~~~~~-- Micro Soil type topography Moisture % bare % litter ground Figure 8.5: Percentage of lm x lm seedling plots with particular environmental attributes for (a) the Ord River and (b) the Blackwood River 181 (b) Blackwood River y = 22.357 * 10 -O.lSOx r = 0.755 E 40 tr Ill 30 0 Do 20 0 10 z0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Elevation (m) Figure 8.6: Relationship between elevation from the river and number of seedlings in lm x lm plots for (a) the Ord River and (b) the Blackwood River. (a) No. of seedlings water level 8 16 24 -..-- • 4 ! water level N 70 weeks• • ~: : : • 60 ! : 40 week$ E 50 ! water leV:e' ~ 40 ! 32 week~ c 30 y: ! ~ 20 i i Cl) ' : Cl) 10 : : ..... Cl) 0 i :...... •...... , ...... ;••······················· Q~~~NO~! O~~~NO~i O~~~NO~i Q~~~NO~i C~~~NO~~ O~~~NO~iO~~~NO~i O~~~NO~i O~~~NO~ . -NM~~:ieeks -NM~~: -NM~~: -NM~~: -NM~~ -E 50 0 - 40 ..t: -0) 30 Cl) :I: 20 c ea 10 Cl) _nm:u:a:n:l :_nnr.ti=II:H:J : a 0 O~~~NO~!O~~~NOWi O~~~NO~!o~~~NO~! O~~~No~f~~~~~6~16~~~~6~]6~~~~6~f6~~~~6~ -NM~~i -NM~~! ~NM~~i -NM~~i -NM~~i ~~M~~~ -NM~~~ -NM~~~ -NM~~! ------~--~------J------~----~------__j______?. __ ~ ______J______~ __ m______~[~~~-~A__ _m ______j______~ ___m______j______§___ ~ ______t ______?. __ ~------'------ltm ______j Distance from water Figure 8.7: Number of seedlings and growth of Eucalyptus rudis over a twelve month period in lm x lm quadrats on exposed sediment on the river foreshore at Site B2bon the Blackwood River. Monitoring from November 1996 (0 weeks) to November 1997 (48 weeks). 1-' R3 .. .,".''·;! 1 (a) No. of seedlings --- water level C\1 80 l 8 ' 1 6' 2 4, 4!8 . ·e 70 i weeks 1 [ water level' Ui 60 t» 50 weeks c: = 40 -g 30 : : w 20 .... Cl) hlh 10 O OCIOCD" 0 m I 1 m i 2 m j 3 m ~eeks4 m I 5 m I 6 m I 7 m I 8 m I ------··u··········----i. ______j______{...... t. ______i______~------L ...... L------j Distance from water Figure 8.8: Number of seedlings and growth of Melaleuca rhaphiophylla over a twelve month period in lm x lm quadrats on exposed sediment on the river foreshore at Site B2b on the Blackwood River. Monitoring from November 1996 (0 weeks) to November 1997 (48 weeks). es>-' 184 Ill 2 metres ...... 4 metres ----•···· 8 metres a) 30 -"#. 25 -ij 20 -s:: 8 15 Cl) loo .a 10 .!:!! ~ 5 0 8 16 24 32 40 48 Weeks b) -E cl} 5 E - 4 . . ..1 ... 0 8 16 24 32 40 48 Weeks Figure 8.9: a) Mean (±SE) gravimetric soil moisture content and b) soil electrical conductivity along seedling transects on exposed sediment on the river foreshore at Site B2b on the Blackwood River. Soil samples were taken at 2, 4 & 8 metres along the transects. Monitoring from November 1996 (0 weeks) to November 1997 (48 weeks). r 185 a) Site 01 18 3.5 ········0···'"'"' E 16 3 - & 14 E 0 ········· 2.5 - :1.. 12 -.s:: Cl) C) c. 10 2 ·a; 0 C) :I: s:: 8 1.5 ~ -g- 6 Cl) en 4 seedlings seedlings flooded flooded 0.5 2 ~ ~ 0 0 AS 0 N D! J F M A M J J AS 0 N D !J F M A 1996 1997 ' 1998 l; l; b) Site 04 4 2.5 -111-- ... 2 ...... 0 ...... E 12 . :r..················· 2 C" 0 10 ...... ~ ...... ~··· ... ··· 2 8 1.5 6 1 4 seedlings seedlings 0.5 2 flooded flooded ~ ~ o~~-r~~-r~~-r~~-r~~~~r+ 0 AS 0 N Dj J F M A M J J AS 0 N D J F M A 1996 ! 1997 1998 Figure 8.10 : Mean density and growth of seedlings of Melaleuca leucadendra in ten lm x lm quadrats on sediment islands at two sites on the Ord River over a period of 80 weeks. 186 Monitoring of seedling plots on sediment islands on the Ord River at two sites indicated the greatest loss of seedlings was in the wet season when seedl!J1gs were inundated for two to three months and completely submerged for some of that time (Figure 8.10). There was also some loss of seedlings in the dry season with soil moisture content of 8.4% ± 0.7 in the mid dry season (August 1996) and 3.8% ± 0.4 in the late dry season (December 1997) for both sites combined. During the April 1997 and April 1998 measurements, sediment in the seedling plots was waterlogged. Growth of seedlings was steady over the study period with seedlings increasing in height by 2 m over a twenty month period at Site 01 and 1.4 m at Site 04. Relative growth rates were 0.044 an cm yr- 1 for the dry season (April - November 1997) and 0.141 cm cm yr-1 for the wet season (December 1997- April1998) at Site 01 ari.d 0.0194 cm cm yr-1 for the dry season and 0.171 cm an yr-1 for the wet season at Site 04. Effects of flooding After 12 weeks of the flooding experiment, survival of seedlings of Melaleuca leucadendra and M. rhaphiophylla was 100% for all treatments including total submersion. Eucalyptus camaldulensis seedlings had 100% survival in control and waterlogged treatments with 60%±12 for flooded and 73%±15 for submerged treatment. E. rudis had 100% survival in the control and waterlogged treatments, 71 %±6 for flooded and 60%±9 for submerged treatments. Growth of seedlings showed the detrimental effects of flooding and submersion on the Eucalyptus spp. with significantly reduced mean dry weight of seedlings in flooded and submerged treatments after eight weeks (p= 0.03, Ec; p=0.063, Er) and 12 weeks (p=0.032, Ec; p=0.021, Er) (Figure 8.11). Growth of the Melaleuca spp. seedlings in the totally submerged treatment was significantly reduced after eight weeks (p= 0.033, Ml; p=0.031, Mr) and 12 weeks (p=0.022, Ml; p=0.015, Mr). Dry weights of seedlings in the flooded and waterlogged treatments were similar to those of seedlings in the control treatment. Measurement of seedling top height showed similar trends for each treatment as dry weights except for M. rhaphiophylla where there was no difference in mean seedling height for any treatment after 12 weeks (Figure 8.12). This indicated the 187 elongation of the stems of this species in trying to grpw out of the water. Some seedlings of this species had grown out of the water in the submerged treatment by the ninth week. Mean relative growth rates over the 12 week period also demonstrated ti:e superior growth of the Melaleuca spp. and the lack of a negative effect on growth rate of partial submersion in the flooded treatment (Figure 8.13). For the Eucalyptus spp. there was a clear separation of growth rates with treatment, although, as for dry weights, there was no difference between control and waterlogging treatments. Further adaptation of the Melaleuca spp. to flooding was evidenced by the formation of adventitious roots on the stems of 70% of seedlings by the fourth week. Root to shoot ratio dry weights did not differ significantly for any treatment for any of the species and at 12 weeks varied from 2.07 to 3.95 across all species. Xylem pressure potential was significantly more negative than the control in the waterlogging, flooded and submerged treatments for E. camaldulensis (p=0.023) and in the submerged treatment for E. rudis (p=0.014) after 8 weeks (Figure 8.14). For M. leucadendra (p=0.098) and M. rhaphiophylla (p=0.0675) there was no significant difference between treatments at 8 weeks but after 12 weeks xylem pressure of seedlings of M. rhaphiophylla in the flooded and submerged treatments was significantly more negative (p=0.025). 11 control ...... waterlogged 188 ----•---- flooded ---la--- submerged a) E. camaldulensis b) M. leucadendra 8 8 7 - 7 -9 6 6 .1: 5 5 C') 4 4 ·~ 3 3 ~ 2 2 c 1 1 0,_----,-----~----~ 0~----,-----~----~ 0 4 8 12 0 4 8 12 Weeks Weeks c) E. rudis d) M. rhaphiophylla 8 8 7 7 C') ...... ··:~; - 6 6 •······· ,~' .l!: • -J:: 5 5 -C') ...... (1) 4 4 3:: 3 .. ·· 3 .;.·· ... 2 2 1 1 0 04-----~----~----~ 0 4 8 12 0 4 8 12 Weeks Weeks Figure 8.11 : Mean dry weight of seedlings subjected to one of four flooding treatments over a twelve week period. Means (±SEM) are of three replicates for each treatment. 189 11111 control ...... waterlogged ··--•···- flooded ---lr·- submerged a) E. camaldulensis b) M. leucadendra 70 70 -E 65 65 0 60 60 .i ...... -.c:: 55 55 -m 50 50 'Ji./ ·a; .c:: 45 45 :::&·········:!: 40 40 ..·· .~ m ..· .. ·· ·--1:: 35 35 -g 30 .• / ·1 30 Q) :;::::··;·~··--.:f:::.·t·:_ en 25 25 20 20 15 15 0 2 4 6 8 10 12 0 2 4 6 8 10 12 Weeks Weeks c) E. rudis d) M. rhaphiophylla 70 70 65 65 60 60 .c:: 55 55 -m 50 50 ·a; .c:: 45 45 m 40 40 1:: 35 35 ·--"C 30 Q) 30 Q) 25 (/) 25 20 20 15 15,_-o--.--.--.--.--.- 0 2 4 6 8 10 12 0 2 4 6 8 10 12 Weeks Weeks Figure 8.12 : Mean top height of seedlings subjected to one of four flooding treatments over a twelve week period. Means (±SEM) are of three replicates for each treatment. 190 Ill control • waterlogged • flooded .. submerged 0.25 p = 0.017 p = 0.015 p = ·o.oog p = 0.019 0.2 Cl)- -Cl) «<~ ... Cl) 0.15 i .t:ID-::s: 1:3: I ON 0.1 ! 0,.,...... I Cl) Cl) 0.05 > c. :;::; ! «< • I -Cl 0 Cl) • • a:~ -0.05 i I -0.1 Eucalyptus Melaleuca Eucalyptus Melaleuca camaldulensis leucadendra rudis rhaphiophylla Figure 8.13: Relative growth rate of seedlings of each species for each of the four flooding treatments, over a 12 week period. Means (±SEM) are from three replicates and p values are from Kruskall-Wallis non-parametric ANOVA of treatment means for each species. 191 Ill control ...... waterlogged --·-•--·- flooded ---AI--- submerged a) E. camaldulensis b) M. leucadendra 0 0 ~~ ~- Cl) -500 -500 '• L.. ·~ ... en::::J "·······~-~.~-::~- -...... - en ··•················i f -1000 -1000 c. E ~ -1500 -1500 >- >< -2000 0 6 12 0 6 12 Weeks Weeks c) rudis d) M. rhaphiophylla 0 0 -c.Ctl .!11:: G) -500 -500 -L.. en::::J en ~ -1000 -1000 c. E .2:! -1500 -1500 ><>- -2000 0 6 12 0 6 12 Weeks Weeks Figure 8.14: Xylem pressure of seedlings of each species for each of the four flooding treatments over a 12 week period. Means (±SE) are of three replicates for each species in each treatment. 192 Discussion Field observation and experimental results presented here show the adaptation of the study species to their riparian environment. Temperature provides a cue to break innate dormancy of seed so that seeds are likely to germinate at the time of year with the greatest chance of seedling survival (Mott & Groves 1981; Bell et al. 1993; Britton & Brock 1994). As demonstrated in the previous Chapter this also correlates well with timing of maximum seedfall and river hydrology. Optimum temperature for germination of Ord and Blackwood river species was similar, despite very different climatic and hydrological conditions. This may be related to the common features of the riparian environments where moisture is readily available (Bell & Williams 1997). Position of seedlings in the riparian landscape is also strongly related to environmental conditions that provide adequate moisture but which protect seedlings from flooding and relates to dispersal of seed within the riparian zone (see Chapter 7). Of the study species, the Melaleuca species showed a greater degree of flooding tolerance than the Eucalyptus species, which is reflected in the position in the riparian landscape in which they occur. This provides an insight, at least for the species examined here, into the distribution of plants along the elevation gradient in the riparian zone described in Chapter 4. Seed germination Optimal germination temperatures for all four species tested was 25- 30° C in terms of both total germination and germination speed. For the Ord River species (E. camaldulensis and M. leucadendra) this corresponds with mean diurnal temperatures experienced from March to May (see Chapter 3) which is also the time of receding high river flows and maximum seed fall (Chapter 6). These ambient temperatures also occur in October to November but successful recruitment is unlikely then because of insufficient seed storage, low moisture availability and the oncoming wet season which results in prolonged inundation of small seedlings and/or river flows scouring out seedling beds. Eucalyptus species from northern Australia are considered to have germination temperature optima of around 30° C (Bell & Williams 1997). Variation in 193 germination requirements of moisture and temperature of different provenances of E. camaldulensis is considered to be a consequence of adaptation to the environment in which they occur (Gibson & Bachelard 1987). Germination temperature optima for Blackwood River species (E. rudis and M. rhaphiophylla) is higher than would be expected for species occurring in the Mediterranean climate of south-western Australia (Bell et al. 1993; Bell & Williams 1997). However results presented here for E. rudis agree with those presented in a study on temperature effects on germination of native species used for mining rehabilitation (Bell & Bellairs 1992). It has been suggested that as E. rudis is a species of streams and wetland areas, germination temperatures are higher as soil moisture will be available longer into summer in these habitats (Bell & Williams 1997). This may also apply to the germination of M. rhaphiophylla. Germination of seed is also enhanced by high humidity (Battaglia & Reid 1993), which is a common feature of riparian environments and which may allow germination to continue later into the dry season. In addition, delaying germination to later in the year when temperatures are warmer allows germinating seed to avoid the period of flooding and waterlogged soils. Optimum germination temperature for E. camaldulensis on the Murray River in Victoria is reported as 35° C and this has been related to delaying germination until after winter floods (Dexter 1978). The very rapid germination of E. rudis at 25 and 30° C may provide a competitive advantage (Mott & Groves 1981) for this species in being able to quickly establish when conditions are optimal. The temperatures tested in this Chapter are limited and a greater range of temperatures would provide a more complete picture of the germination range for each species as it relates to temperature. In addition, it may be useful to create germination conditions that would be experienced by the seed in situ. This may include alternating light and dark conditions and corresponding changes in temperature. A regime of alternating wetting and drying periods may also affect germination patterns of seeds and may be useful to determine whether seeds can remain viable after imbibition followed by drying out (Gibson & Bachelard 1987). Testing germination under these conditions may shed some light on the viability of buried seed which was tested under field conditions in the previous Chapter. 194 Germinating seed of the Blackwood River species showed some tolerance at low levels of salinity 1 (10 mScm- ) but total germination and germination speed was severely affected at the higher salinity levels. Seed of Melaleuca cuticularis and the· ecotype of E. rudis from saline environments showed no greater tolerance to higher water salinity. Tolerance to salinity at germination has been found not to correlate with the tolerance of seedlings (Pearce-Pinto et al. 1990) or of adult plants (Morris 1984). lmbibition of seeds may only take place when rainfall sufficiently dilutes surface and soil water salinity levels to more favourable external osmotic potentials (Bell & Williams 1997). Seed germination tolerance to salinity of riparian species en the Blackwood River may become important if these species are reliant on water from the river to initiate germination. Blackwood River salinity levels have been steadily rising at a rate of 52 mgl-1 yr·1 (Schofield et al. 1988) to present levels of 2069 mgl-1 (five year moving average) at Site B1 on the lower Blackwood with higher levels for the upstream sections of the river (Water and Rivers Commission unpublished data). Seedling establishment For both rivers, the environmental conditions that appear to coincide with seedling establishment are similar. Most seedlings occur in river sediment of larger particle size of sandy to loam soils. Seedlings are also more dense at lower elevations and in microtopographic sites that are moisture- gaining but also where there is a slight rise. These rises are important for seedling establishment in creating safe sites to prevent waterlogging (Bren & Gibbs 1986; Skoglund & Verwijst 1989; Titus 1990). This highlights the importance of fluvial processes in recruitment. Moisture-gaining sites are important for sustaining seedlings through the dry part of the year. Greater emergence of seedlings of E. delegatensis in depressions was attributed to lower vapour pressure deficits in these microsites (Battaglia & Reid 1993). Litter levels and areas of bare sediment do not appear to be important for establishment of developing seedlings. These seedling plots were of older seedlings (;:::1 year old) and litter levels and area of bare ground may however be important for the successful initial establishment of germinating seed (Nilsson & Grelsson 1990; Greenway 1994; Johnson 1994). Seedling plots on the Ord River occurred at large distances from mature trees and therefore probably from an alternative source of seed. 195 This may indicate that seed can be dispersed by flood waters. The presence of numerous open fruit in the flood debris lines indicate that seed is being transported in this way. Thus in large flows seed can be transported not only longitudinally down river, but also laterally away from the channel to areas which afford some protection from subsequent flooding after initial dispersal en floodwaters. In fact, for the Ord River, successful long term establishment may require a period without large floods. On the Blackwood River seedling density was closely related to distance to seed trees and no seedlings were found further than 30 metres from a potential seed tree. For E. camaldulensis on the Murray River in south-eastern Australia a similar close association between seedlings and seed trees was found (Dexter et al. 1986). This suggests that seeds are generally dispersed by wind with little dispersal on floodwaters. Germination trials on seeds of E. rudis indicate that seeds germinate rapidly after wetting (one to two days) and the majority of floating seed germinate within four to five days (Chapter 7), so that they may not have time to be dispersed by water. However, floating seed can be transported long distances very rapidly en floodwaters (Nilsson et al. 1993) so that seed do not need to remain viable for long to be effectively dispersed in this manner. Seed and fruit of E. rudis was found in debris samples on the Blackwood River (Table 7.5) with good viability of seed. It is most important for seed to be deposited in a safe site where conditions are conducive for survival and growth. In seedling plots along the transect on the Blackwood River the greatest survival and growth of E. rudis was in the mid-slope position. Where seedlings were below this, flooding conditions caused high mortality and restricted growth and in the upper-most plots conditions were too dry with mortality and poor growth most likely caused by drought conditions in summer. For M. rhaphiophylla the zone of optimal survival and growth was lower on the transect. This reflects the greater tolerance to waterlogging of this species and also reflects the respective positions of the adults of these two species in the landscape with M. rhaphiophylla occupying the lower landscape position on the fringe of the high water level and E. rudis generally occupying the bank top and on to the floodplain. Greatest loss of seedlings for both species was at the dry time of year (late summer and autumn) when soil moisture levels are low and soil salinity is high. This indicates that surviving the summer drought may be more important in survival of seedlings 196 than flooding in winter, which inhibits growth but to which plants are fairly tolerant, particularly M. rhaphiophylla. Replenishment of soil moisture is probably important for riparian seedlings in arid wne rivers. Roberts (1993). found a good correlation between recruitment and flood events which replenish soil water for E. coolabah seedlings on a semi-arid ephemeral creek in eastern Australia. M. leucadendra seedling plots on deposition islands on the Ord River showed the greatest loss of seedlings in the wet season when they are completely submerged for up to two months. Although, owing to the 'flashy' nature of river flows on the Ord River, which may cause levels to rise and fall during this period, seedlings would have periods when at least some portion of the plant is exposed to the air so that photosynthesis, and therefore growth, can continue. This is supported by growth rates in the wet season are higher than in the dry season when lack of water may be limiting but is not great enough to cause widespread drought-death of seedlings. Effects of flooding The results of the flooding experiment correspond to field observations of the different positions in the landscape occupied by the Melaleuca and Eucalyptus species with the former occupying the lower landscape positions subject to greater frequency of inundation. Greater tolerance, at the seedling stage, to inundation has been used to explain the differences in distribution in the riparian zone of two rainforest species in Victoria (Melick 1990). The flooding experiment indicated that both Melaleuca species can survive even complete submergence for several months. Growth, however, is affected by total submergence as transpiration and photosynthesis shut down, severely affecting the energy budget of the plant (Ernst 1990). The Melaleuca species showed signs of adaptation to flooded conditions, such as elongation of the stem and the formation of adventitious roots. Elongation of the stem (controlled by the plant hormone ethylene) is important in flood tolerance as it allows plants to reach the water surface and maintain photosynthesis (Blom et al. 1990). This, coupled with the formation of aerenchyma tissue and the growth of adventitious roots allows better oxygen diffusion. Elongation of the stem or petiole is a common adaptation of flood tolerant aquatic plants (Blom & Voesnek 1996). The 197 formation of adventitious roots and stomatal closure were considered important factors in the flooding tolerance of Melaleuca quinquenervia seedlings, a species of northern Australia (Senna Comes & Kozlowski 1980). For Melaleuca halmatrorum seedlings, survival was poor after six to nine weeks of total submergence, although elongation of the stems increased survival rates (Denton & Ganf 1994), indicating some flood tolerance in this species. Survival and growth of the Eucalyptus species was significantly affected by partial and total submersion. This may indicate an inability of these species to produce the physiological adaptations seen in the Melaleuca species. All of the species tested, including the eucalypts, did not show any detrimental effects of waterlogging, indicating adaptation to their environment where persistence in proximity to the river would require some tolerance to waterlogging. Waterlogging has been demonstrated experimentally itot to effect the growth of E. camaldulensis seedlings (Akilan et al. 1997) nor leaf water potential (Pereira & Kozlowski 1977). Age of seedlings would also clearly affect tolerance to flooding. Seedlings tested here were more than six months old and younger seedlings may not show the same tolerance. However, flooding clearly affected the Eucalyptus species which suggests that these species require access to the air through the soil surface for survival and growth. Although in the flooding experiment some attempt was made to circulate water between the seedlings by pumping water through the tanks, providing less detrimental conditions than stagnant water (Kozlowski 1984), this was at a fairly slow rate and would not very well simulate the high flow rates that would be experienced in floodwaters. The physical effect of high flow rates on seedlings is likely to be substantial and therefore requires further examination to contribute to understanding of the mechanisms affecting survival and growth of seedlings in the riparian zone. The previous Chapter examined the sources of seed in the riparian zone, including the development of seed and its release, the dispersal of seed in flood debris and the storage of seed in the soil seedbank as well as seed predation and long term viability. This Chapter has expanded this in investigating the optimum germination temperature of seed from the selected species and the environmental factors that affect establishment and growth of seedlings including soil type, microsite conditions and adaptation to flooding. The next chapter will look 198 at the factors affecting recruitment in the riparian zone by dealing with the population structure and spatial distribution of the selected species. 199 CHAPTER 9 POPULATION STRUCTURE AND SPATIAL DISTRIBUTION OF SELECTED SPECIES IN THE RIPARIAN ZONE. Introduction The previous chapter examined the ecological factors influencing the germination, survival and growth of seedlings of E. rudis and M. rhaphiophylla on the Blackwood River and E. camaldulensis and M. leucadendra on the Ord River. This Chapter will consider the age class structure of these species as well as their spatial distribution within the riparian zone. This will be done by inferring age by the size class structure of each species along transects on different river reaches. Recording the age class structure of a sample of a population can infer survivorship through time of cohorts of individuals (Hutchings 1997). This can be used to assess the population stability and successional trends or vegetation dynamics of these species (Veblen 1986). The spatial structure of the populations at these sites will also be related to the long term hydrologic records for these rivers. Fluvial regime and river geomorphology are major influences on the spatial and temporal structure of riparian vegetation Gohnson 1994; Barnes 1997; Johnson 1997). The formation of sandbars, levee banks and river terraces as well as river bends, pools and riffle zones all make for a heterogeneous environment for the recruitment and establishment of riparian plants (Fonda 1974; Bell & del Moral 1977). Added to this is the regime of regular or irregular flooding. Large scale disturbances, such as high water flows and flooding, are likely to be important factors in the dynamics of the riparian community. Recruitment, establishment and survival of riparian 200 trees is thought to be closely tied to the hydrological regime (Barnes 1985; Bradley & Smith 1986; Johnson 1994). Flood-induced disturbance may instigate the transition from stable states of plant composition and population structure. It can be hyp<;>thesised therefore that river systems with frequent disturbance from high energy flood events will support riparian communities with long periods of transition and rather shorter stable states. That is, communities which are in a state of arrested succession (Webb et al. 1972) and which, in terms of individual populations of trees, rarely reach maturity and remain in a juvenile stage of development. The present state of an ecosystem is the result of a particular history of disturbance, climatic factors and/ or management. The regeneration process and longevity of the vegetation can also be critical in determining the stage and rate of transition states (Hobbs 1994). State and transition models have been suggested as a tool for rangeland management (Westoby et al. 1989; George et al. 1992), as an alternative to linear succession, to allow for different vegetation community composition (states), depending on the transition pattern. Non-equilibrium vegetation dynamics may be described by discrete states and definable transitions between states. Transitions are usually instigated by natural events and may occur quickly or over an extended period and change to a persistent state depending on events during the transition (George et al. 1992). This model may provide an appropriate framework for management of degraded riparian zones and assist in interpreting the response of riparian vegetation to disturbance in terms of possible states and transitions of riparian vegetation. The state and transition model generally describes non equilibrium vegetation dynamics in communities but may be usefully applied to populations of particular species as in this case. With chronic to frequent flooding disturbances riparian forests are young, as vegetation patch development or individual populations are continually reset and remain in stand initiation or immature stages of development (Wissmar & Swanson 1990). In environments with low magnitude, seasonally predictable floods the vegetation or populations are likely to be more stable and more developed. This chapter compares the size class distribution and spatial arrangement of the selected study species that occur in the riparian zone on the two river systems. This will give an indication of 201 the health of these populations in terms of the ability of the population to regenerate as the maintenance of a population is an expression of the population's resilience. For example, if small size classes are absent and there is a large number of trees· in large size classes this may indicate that there is a lack of recruitment and the population is senescent (Hutchings 1997). In contrast, if there is a majority of trees in smaller size classes and few large individuals it may suggest a building population after a disturbance (Szaro 1990). In addition, the presence of groups of even aged individuals (cohorts) or a more even distribution of sizes will provide some evidence of whether recruitment patterns are episodic or continual (Harper 1977). The position these species occupy in the riparian zone will be considered as well as those areas which are favoured for seedling recruitment and establishment of mature trees. Vegetation dynamics will also be discussed in terms of a state and transition model for populations of these riparian species, in relation to the fluvial regime and the implications for riparian vegetation management. The information presented in this chapter on the development of the mature plant into the population provides the next step in understanding regeneration processes of these species in the riparian zone. This follows on from the previous Chapters which looked at the replacement stages of plants including production of seed, dispersal, germination and the establishment of seedlings. These elements incorporate the five factors which are considered important in the differentiation of the regeneration niche (Grubb 1977). Methods This Chapter uses the two most common overstorey species from each river system as described in the previous chapter. Eucalyptus camaldulensis occurs as a dominant riverine species on the Ord River. E. rudis is the only large tree that occurs in the riparian zone of south-west rivers including the Blackwood River. Melaleuca leucadendra on the Ord River and M. rhaphiophylla on the Blackwood River are also similar species and common overstorey or midstorey species with a high tolerance of waterlogged conditions. 202 At the five sites on the Ord River and Sites B1 and B2 on the Blackwood River (see Figure 3.1 & 2), three 10 m x 100-200 m belt transects were set up perpendicular to the river. The length of each transect varied with the width of the riparian zone ·at each site. Each transect was divided into 10 metre sections to give a series of contiguous 10 m x 10 m plots. Within each plot the diameter at breast height (1.5 m) (DBH) of all trees of the study species in the transect were measured. In addition, the distance (m) of each tree from the river dry season base flow level was measured. Relative elevation (m) above dry season river base flow level was taken along each transect at 5 m intervals using an automatic level and staff. Elevations were given as height above the lowest point along the transect, which was invariably the low water mark in the river channel. For all sites combined for each species on each river elevation was correlated with DBH to test if there was a pattern between tree size and position in the riparian landscape. Diameters of individuals of each species at each site were arranged in ascending order to visually identify cohorts of trees which would indicate past recruitment events. From this data trees were also arranged into size classes and the frequency of each size class was plotted along the transects along with elevation for each site on both rivers. 3 1 Historical flow records (1958-1995) of monthly river discharge (m s· ) for the Blackwood River were obtained from the Water and Rivers Commission gauging station at Darradup, located near Site B1a. This station covers approximately the upper two thirds of the Blackwood River catchment and includes the two study sites on the Blackwood River. Historical flow records 3 1 (1971-1993) of monthly river discharge (m s· ) for the Ord river sites were obtained from Water and Rivers Commission gauging stations at Old Ord River Station for Site 05 and Mistake Creek station on the Negri River for Site 04. No records were available for Site 03 on the Behm River or for Site 02 which is on Lake Kununurra since the river was dammed in 1963. Average monthly flow levels for the Ord River below the diversion dam (Site 01) were obtained from the modelling of flow calculated by the Water and Rivers Commission ( J. Ruprecht, unpublished data). 203 Aerial photographs (1:20 000) were used to map the approximate distribution of the study species in the riparian zone at each site on the Ord and Blackwood Rivers. Data from the transects was used to obtain tree density estimates and species distribution. Median tree diameter and density per hectare was derived from data measured in the 100 m2 plots. An indication of the relationship between stem diameter and tree age was developed by taking a transverse section of a sample of stems up to 10 cm in diameter, polishing the cut surface and counting the growth rings which were assumed to be laid down annually. For stems greater than 10 cm in diameter an increment borer (Sunnto , Finland) was used to obtain a stem core which was mounted and polished and examined under a stereo microscope to reveal the annual growth rings. A sample of stems of different sizes and in different locations was taken at each site for E. rudis and M. rhaphiophylla on the Blackwood River and E. camaldulensis and M. leucadendra on the Ord River. M. rhaphiophylla growth rings were not obvious so this species was not considered for age estimates as the relationship between size and tree age could not be reliably estimated. Estimates of annual growth ring count were correlated with stem diameter to test the relationship between stem diameter and age so that stem diameter could be used as a surrogate of age. This method provides only a relative indication of the age of plants and is not a dendrochronological study. This is because of the inherent problems of distinguishing annual growth rings in species from variable climates or where there are not distinct seasonal climatic patterns (Ogden 1978). Riparian species especially may be insensitive to seasonal drought, or flooding may affect growth, and rings may not necessarily be laid down annually (Schweingruber 1992). In a shtdy of tropical eucalypts annual rings could be distinguished from each other by groups of even growth zones (Mucha 1979). 204 Results Tree population structure A summary of the total number of trees of Eucalyptus rudis and Melaleuca rhaphiophylla in the different size classes (DBH) for the Blackwood River sites is given in Figure 9.1. E. rudis was found in greater abundance for all sites and in all size classes, showing it to be the dominant tree at these sites. As highlighted in Chapter 6, the effects of livestock grazing can be seen in the lack of trees of smaller sizes at the sites where stock have free access to the riparian zone (Sites Blb & B2b). This is particularly evident for E. rudis, with no stems less than 10 cm diameter being recorded for these sites and with a large proportion of older mature trees indicating an over-mature population. On the ungrazed sites there is a large proportion of stems in the smaller size classes indicating that recruitment of these species is taking place. The spread of size classes, with the greatest proportion of smaller sizes and fewer large older trees, indicates a dynamic population. Where individual stems of each species at each site were arranged by increasing diameter no obvious cohorts of similar sized trees were recognisable (Figures 9.2 & 9.3). There may be some evidence of size cohorts, particularly for E. rudis at Sites Bla & Blb but there are many stems which fit between the groups. This may be due to continual recruitment of individuals into the population or may indicate the presence of suppressed or exceptional individuals. Summary of size classes of Eucalyptus camaldulensis and Melaleuca leucadendra for all five sites on the Ord River showed the greatest number of stems in the smaller size classes and therefore younger trees (Figure 9.4). 205 riJ Eucalyptus rudis D Melaleuca rhaphiophylla a) Site 81 a b) Site B1 b 60 60 50 50 >- (,) c: 40 40 Cl) ::1 0" 30 30 ...Cl) u. 20 20 10 10 0 0 C'\1 0 0 0 0 0 0 C'\1 0 0 0 0 0 0 .,... ("') .,... ("') 1.1) V C'\1 o:t 1.1) 1.1) V C'\1 o:t 1.0 A A C'\1 0 0 0 0 C'\1 0 0 0 0 .,... C'\1 ("') o:t .,... (\1 ("') o:t DBH size classes (cm) DBH size classes (cm) c) Site B2a d) Site B2b 60 60 50 50 >- (,) c: 40 40 Cl) ::1 0' 30 30 ...Cl) u. 20 20 10 10 0 0 C'\1 0 0 0 0 0 0 (\1 0 0 0 0 0 0 .,... (\1 ("') .,... ("') 1.1) V o:t 1.1) 1.1) V C'\1 o:t 1.0 A A (\1 0 0 0 0 (\1 0 0 0 0 .,... C'\1 ("') o:t .,... (\1 ("') o:t DBH size classes (cm) DBH size classes (cm) Figure 9.1: Size class distribution of Eucalyptus rudis and Melaleuca rhaphiophylla at Sites Bl & B2 on the Blackwood River. 206 a) Site B1a- Eucalyptus rudis 50 ...... - 40-l··· ...... E ~ 30 :I: 20 -!...... m c 10 Melaleuca rhaphiophylla 50 ...... b) Site B1 b ~ Eucalyptus rudis 50 ...... - 40 ...... E ~ 30 :I:m 20 c 10 0 5o ...... !Y!.~l~.~~.'!.£.1!..t!!!:!P.IJ.!.9.P.!)Y!.!.I!...... -E 40 ~ 30 :I:m 20 c 10 Figure 9.2: Diameter at breast height (cm) of individual trees measured along the transects at Site Bl on the Blackwood River. 207 a)Site B2a - Eucalyptus rudis 50-r•••••••••••••••••••••••••••••••••••••••••••••u•••••••••u••••••••••"'''""""'"""""""'"""""""""""'"""""""""""""""'"""""""'"'""""""""'"'""'"""""""""""""""' ~ 40~ ...... ,...... ,,...... ,,...... ,...... ,,, ...... ,...... E .£, 30 ...... ,...... ,...... ,...... ,. .... ,_ :I: 20 -!... .,...... m c 101...... ,...... , ...... ,,...... =:=;:~~m 50 ...... M.~!?.!f#.Y.P..If!..tfl!!!P.IJ.!.QP.!JY!.!?...... ~ 40 E .. .£, 30 oooooooooonoooooooooooooooooooooooooooooooooooooooooooooonoooooooooooonooooooooooooooooooooooooooooonoooooooounoooooooooooooooooooooooo ...... :I:m 20 c 10 50 .~.~~.~ .. ~.~~ ..~ ... ?.':!..t?..I!!:!.Y..P..~':!..~.t~~~J~...... ~ 40 E .£, 30 J:m 20 c 10 5o ...... !Y!f!!.i!lf!~P..~.t!:':.~P..IJ.!.~P!JY.!!.~ ...... , ~ 40 E .£, 30 •, ., mJ: 20 c 10 0 Figure 9.3: Diameter at breast height (cm) of individual trees measured along the transects at Site B2 on the Blackwood River. 208 Eucalyptus camaldulensis kJ Melaleuca leucadendra a) Site 01 b) Site 02 60 ;;:... 50 (.) c: Cl) 40 ::I 0" 30 ...Cl.l I.L 20 10 0 0 N 0 0 0 0 0 0 N 0 0 0 0 0 0 N M ~ LO LO ?" N M ~ LO LO V .... V I 1\ 1\ N 0 0 0 0 N 0 0 0 0 N M ~ .... N M ~ .... DBH size classes (cm) DBH size classes (cm) , c) Site 03 d) Site 04 84 85 60 ;;:... 50 (.) c: 40 Cl.l ::I 0" 30 ...(!) I.L 20 10 0 N 0 0 0 0 0 0 N 0 0 0 0 0 0 .,.. N C? ~ LO LO .... N M ~ LO LO V V 1\ A N 0 0 0 0 N 0 0 0 0 C\1 C? ~ .... N M ~ .... DBH size classes (cm) DBH size classes (cm) e) Site 05 60 ;;:... 50 (.) c: 40 Cl.l ::I 0" 30 ...Cl) u. 20 10 0 N 0 0 0 0 0 0 ?" N M ~ LO LO V I A N 0 0 0 0 ?" N M ~ DBH size classes (cm) Figure 9.4: Size class distribution of Eucalyptus camaldulensis and Melaleuca leucadendra at Sites 01 to 05 on the Ord River. 209 At all sites, large numbers of seedlings and juvenile trees were recorded with few older, large diameter trees. At all sites except Site 05 M. leucadendra was more abundant than E. camaldulensis, indicating that it is the dominant tree_ species at these sites although E. camaldulensis grows to a larger size (most abundant in >50 an size classes) and can be locally dominant. When individual trees were sorted by diameter, size (and therefore age) cohorts could be discerned for E. camaldulensis at all sites and for M. leucadendra at all sites except 01 (Figures 9.5, 9.6 & 9.7). At Site 01 stem diameters of M. leucadendra were continuous, with m obvious gap in size between groups of stems that would indicate an age difference. Distinguishable cohorts of size classes were particularly obvious for E. camaldulensis at Site 04 where diameters were in four size classes: <2 an, which were all approximately 1 year old seedlings of around 30cm in height; 2.1 to 4.8 cm diameter; 12.8 to 14.8 cm; and large trees >50 an which were probably of several ages and ranged in diameter from 51.8 to 85 cm. Size class distribution A profile diagram on the Blackwood River at Sites Bla & B2a of E. rudis shows seedlings to he most abundant at the lower elevation near the river edge but large diameter trees are also abundant at these lower elevations. Size classes were generally evenly spread along the profile (Figures 9.8 & 9.9). M. rhaphiophylla tends to be confined to areas close to the water which are regularly inundated. In this area there is a wide range of size classes. Similar patterns of establishment along the profile was evident for Sites B1b & B2b, except for the lack of the smaller size classes. Along the profile at Site B2a & b, trees have also become established away from the river edge in a slight depression. For both sites for both E. rudis and M. rhaphiophylla there was no significant relationship between size class and elevation (Figure 9.10a & b). Distribution of trees along the river reach at the Blackwood River sites show a greater density of trees closer to the summer water levels, with trees becoming much more sparse further from the river (Figure 9.11). Both sites are on long straight river pools and at Site Bl there is a steep bank on the western side with a much narrower riparian zone. 210 50 Site 9J..::... !?.YP~lY.Pt'!.~.£a".;m.;.~.a~;.~l.d,.~.u~,,.;l.e;;.;n.;.s;;.;i.s;;...... 40 E -0 30 -::r:: m 20 c 10 0 50 40 ...... ~ -E 0 30 -::r:: m 20 c 10 0 50 .~J..§.!~~... 9.?. ...7 ... FY£~!YP.!!:!.!j .. 9..~rr!.~l!!.Y.!.~IJ.~!.~ ...... 40-i ...... 30 ::r:: 20 -l ...... q.-.1- ...... m c 10 0~~~~~~~~~~~~ Melaleuca leucadendra 50 ...... - 40 ...... E ~ 30 :I:m 20 ""fo3'"''"''"'''""''"'"'"'"''"''"''""'"'"''"''"'''''"''"''"'"'"''"'"'''"'' c 10 ...... 0 Figure 9.5: Diameter at breast height (cm) of individual trees measured along the transects at a) Site 01 & b) Site 02 on the Ord River. 211 a) Ord Site03 - Eucalyptus camaldulensis 50 40 -4 ...... -E (J 30 -:I: m 20 c 10 0 50 ...... M'f!!.!.?.!.'f!!.Yei!!: .. !.?.Y.9.!4fJ!..?.l1!f.r.?...... ,,...... 40 ...... : .. -E (J 30 -:I: m 20 c 10 0 50 b) .. ~.~~~.9..~ ..~ ... ~':!€~.~Y.P.~.'!.~... €~!E.~!.~':!.!.~!!.~.~~ ...... 40 -E (J 30 ...... -:I: m 20 c 10 0 Melaleuca leucadendra 50 40 -E (J 30 -:I: m 20 c 10 0 Figure 9.6: Diameter at breast height (cm) of individual trees measured along the transects at a) Site 03 & b) Site 04 on the Ord River. 212 Site 05 - Eucalyptus camaldulensis 50-r···········.. ············ ...... ;...... 40+········ ...... -(,)E 30 - -+····...... ~ 20 -f...... , ...... c 50 ...... Melaleuca leucadendra ...... E -u 30 ...... mJ: 20 c Figure 9.7: Diameter at breast height (cm) of individual trees measured along the transects at Site 05 on the Ord River. Eucalyptus rudis Elevation (a) Site B1a 30 Meialueca rhaphiophylla 7 6 >-(J 25 c 5 c Cl) 20 0 ::l 4 ·- 0" 15 m Cl) 3 - a.. > 1.1.. 10 2~ w 5 1 0 I 0 C\1000000 C\1000000 C\1000000 C\1000000 C\1000000 C\1000000 C\1000000 C\1000000 C\1000000 C\1000000 y.-C\1(")-:i'll)ll') y.-C\IC")-:1'11')11') y..-C\IC")-:~"11')11') y..-C\1~11')11') v.-C\IM-:tll')ll') v.-C\IM-:tll')ll') y.,...C\JC")-:1'11')11') y.-NM-:1'11')11') y ..... C'o!C")-:i'li')li') y.-C\1(")-:i'll')ll') • I I I I A I I I I I A I I I I I A I I I I I J\ I I I I I I\ I I I IBA I I I I I J\ I I I I I J\ I I I I I J\ I I U I I J\ NOOOO NOOOO NOOOO NOOOO NOOOO NOOOO NOOOO NOOOO NOOOO C\10000 T""C\IC")-:1' .-NM-:1' ..-NC")-:1' .-NC")-:1' .-NM-:1' ..-NM-:1' ..-NM-:1' ..-NM-:1' ..-NM-:1' ..-NM-:1' (b) Site 81 b 30 8 >a 7 (J 25 c 6 c Cl) 20 0 ::l 5:;:: 0" 15 4 m Cl) > a.. 1.1.. 10 3~ 2w 5 1 0 0 NOOOOOO NOOOOOO C\1000000 NOOOOOO NOOOOOO C\1000000 C\1000000 NOOOOOO NOOOOOO NOOOOOO y.-NC")-:1'11')11') y,...... NM-:1'1011') y...-NC")-:1'11')11') v.,...NM-:1'11')11') y..-NM-:1'11')11') v...-NC")-:1'11')11') y..-NM-:1'11')11') y,...C\IM-:1'11')11') y.-NM-:1'11')10 y'~""NM-:1'11')11') I 8 I I I J\ S I I I I J\ I I I I I J\ I I I I I J\ • I I I I A B I I I I A 11 D li I I J\ I I D I I J\ 1 1 a 1 1 A 1 a u a 1 A NOOOC NOOOO NOOOO C\10000 NOOOO C\10000 NOOOO NOOOO NOOCO C\10000 ..-NC")-:1' .-NM-:1' ..-NM-:1' .,...NM-:1' ..-NM-:1' ..-NC")-:1' ..-C\IC")-:1' .-NM-:1' .-NM-:1' .,...NM-:1' 0-1 Om 1 0-20m 20-30m 30-40m 40-50m 50-60m 60-70m 70-80m 80-90m 90-1 OOm Figure 9.8: Size class frequency and distribution of the two main overstorey species (E. rudis and M. rhaphiophylla) along a transverse profile from the water's edge at a) Site Bla and b) Site Blb on the Blackwood River. Frequencies are from 3 combined lOm x lOm plots. Histograms show tree size classes (DBH( cm)) by species for each plot...... N w (a) Blackwood River - Site B2a Eucalyptus rudis Elevation 8 30 Melaleuca rhaphiophylla ~ 7 o 6 -E c~ ~ 5 -c ~ 0 ~ 4 - :~...~ 10 3 -ea u. 2 ~ 1 w 0 'n'f'f''f'f? I 'f''f'f I I I I I I 'f'f''i' I I I I I 'f'f'f'f I i I I I I I I I I I i 'f'f I 'f I I I I I I I I i i I I I 'f I 'f'f'f' I I I I I I I I I I I I 'f'lf''f I'" 0 NCOCCOO NCOCCCC NOCOCCC NOCCCCC NCOCCCC NOCCOCO NCCCOCC NOOCOOC NCOCOCC NCOCCOO v~N~~~~ v-N~~~~ v-N~~~~ v~N~~~~ v~N~~~~ v~N~~~ v-N~~~~ v~N~~~~ v~N~~~~ v~N~~~~ IJJ;IJ\ DtlaiJ\ DIIIIJ\ IIIIIJ\ IIIIDJ\ llUIDIJ\ II!IIJ\ IIIIIIJ\ IISIIIJ\ IIDIIJ\ NCCOO NCOCO NCOOO NOOCC NOCOC NOCOO NOOOO NCCCO NCOOC NCCOC -N~~ -N~~ -N~~ ~N~~ ~N~~ ~N~~ -N~~ ~N~~ ~N~~ ~N~~ (a) Blackwood River - Site B2b 30 ~ 7- 0 6 E c - ~ 20 5 c ~ I. 4 0 ~ . u: 10 3 -~ ~ w 0 ll~···1~1 I I I I pq'('f ' I I I I I i I I I I I I I I I I I I I I I I I I I I I !I,,,, I I I '' I I I I I I I ,,, I I I I I I I I I I I I I I I I ~ 1NCCCCOC NCCCOOC NOOCCCC NCCOOCC NCCCCCC NCOOCOO NOCOCOC NCOOCCC NCOOCCO NCOCCCC y~N~~~~ v-N~~~~ y-N~~~~ v-N~~~~ v-N~~~~ y~N~~~~ v-NM~~~ y~N~~~~ v-N~~~~ v-N~~~~ lllllj\ DIIIDJ\ 111111/\ llliiiJ\ JIIIIJ\ UBfDI.f\ IJEIDJ\ IIJIIJ\ IJJIIJ\ IIIIUJ\ NOOCO NOCCO NOOCO NOCCO NCCCO NOOOC NCOCC NCCCC NOOCC NCCCC -N~~ -N~~ -N~~ -N~~ -N~~ -N~~ -N~~ -N~~ -N~~ -N~~ 0-1 Om 1 0-20m 20-30m 30-40m 40-50m 50-:60m 60-70m 70-80m 80-90m 90-1 OOm Figure 9.9: Size class frequency and distribution of the two main overstorey species (E. rudis and M. rhaphiophylla) along a transverse profile from the water's edge at Site B2 a&b on the Blackwood River. Frequencies are from 3 combined lOm x lOm plots. Histograms show tree size classes (DBH(cm)) by species for each plot. N 1-' 11>- 215 (a) E. camaldulensis (b) M. leucodendra y = 3.911x + 0.675 r = 0.606 y = 3.829x + 0.695 r = 0.560 80 50 p<0.001 p<0.001 40 e 6o u - 30 :I:m 40 c 20 20 10 o~~~Tn~~~~nT~ 0 5 10 15 0 1 2 3 4 5 6 Elevation (m) Elevation (m) (c) E. rudis (d) M. rhaphiophylla y = -1.220x + 49.21 r = 0.060 y = -2.469x + 21.82 r = 0.255 150 60 p>O.S p>0.5 • • 50 •• -fi 100 40 • • -:I: • m • • 30 c • • 50 -• 20 10 • o~~TT~~~~~~~ 1 2 3. 4 5 6 7 0 1 2 3 4 Elevation (m) Elevation (m) Figure 9.10: Relationship between tree size (diameter at breast height) and elevation from the river for the 4 selected riparian species on the Ord River (a & b) and the Blackwood River (c & d). 216 a) Site B1a b) Site B1b summer water level .r. <2; 1000/ha .r. 1 0-20; 300/ha E.r. 10-40; 50/ha 0 20/ha 0 0 high bank E.r. 30-50; 40/ha c) Site B2a & 82b high bank 20ha Scale 100m E.r <2; 200/ha M.r. <2; 500/ha N Figure 9.11: Plan view of the riparian zone at Sites Bl & B2 on the Blackwood River showing median tree diameter size class (cm) and approximate densities per hectare(/ha) for E. rudis (E.r) and M. rhaphiophylla (M.r.). 217 On the eastern side the gradient is more gradual, allowing floodwater to flow out into this area where trees have established. At Site B1b and Sites B2a&b a second line of trees has established away from the river in a slight depression. The profile diagrams for sites on the Ord River show the abundance of new ( <1 year old) seedlings at lower elevations on the river flats and an abundance of smaller size classes throughout the profile (Figures 9.12, 9.13, 9.14 & 9.15). Frequency of larger size classes is greatest at the higher elevations and greater distance from the river, especially for E. camaldulensis. All M. leucadendra size classes tend to be at the lower elevations closer to the rivers edge. For all sites combined there was a significant correlation between diameter at breast height and elevation for both E. camaldulensis and M. leucadendra (Figure 9.10 c&d). At Site 02 on the shores of Lake Kununurra, where water levels do not fluctuate more than a metre each year, there is a large number of seedlings and smaller size classes at the water's edge. There is also a large number of larger trees of M. leucadendra at lower elevations that is now permanently immdated (Figure 9.13). At sites 03,04 & 05 there are large numbers of seedlings on sandy moist sediments on river flats where water levels are receding. On the steep banks (e.g. west bank, Site 05) larger trees occur closer to the water and there are no seedlings as there is no suitable habitat except in a slight depression upslope where there are some smaller size class stems (Figures 9.14, & 9.15). The river reach where Site 01 on the Ord River is located is on a river bend (Figure 9.16a). This shows the occurrence of new seedlings on sandbars on the insides of the bend (depositional) and the steep slope where only larger trees occur on the outsides of the bend (erosional). Establishment of younger trees has also occurred on the edge of the depositional flat where sediments have been deposited when high flows have flooded out across this area. At Site 02 the riparian zone is confined to a narrow strip of new recruitment where lake levels fluctuate and there is virtually no flow at this site (Figure 9.16b). Below this area mature M. leucadendra trees are permanently flooded, with many trees (~60%) having fallen over. Upslope, mature E. camaldulensis are now in a terrestrial environment. (a) Ord River - Site 01 Eucalyptus camaldulensis Elevation 30 ,, East bank 12 wp:;rJ Melaleuca leucadendra ;:::... 10 -E (..) ai 20 8 -c :::i 0 C" 6 ·-ea ~ -> I.L 10 4 Cl) 2 w 1 1 1 1 1 1 0 I I I I ,,., I I I I I i I I I i i i I i i I i i 1 i i i i i I i I i i I i I I i I i I i i i I i'' I i I i I 11 i i I I fi i i i i , 1 i I I 1 i I I i i i i i i 'j'''i"i"J' i I i I i i i I i i i g ljDi'l i I i I I r 0 0-1 Om 30-40m 60-70m 80-90m 100-110m 120-130m 30 (b) Ord River - Site 01 12 ;:::... 235 West bank (..) 10 c -E Cl) 20 8 :::i -c 0 0" 6 :.;= Cl) ea I.L 10 n ~ ~ 4 > - Cl) 2 w 0 l''li I I i I I i I I i I 11 I i I i 111;1i;t;; I i iiQ~ I I i I iiiQll' i I I i ii" i i I I i i i i id i i i i i I I i i Hi 8 i i i I I 1'1''ii i I I i i I I i i I I I I in I I i I I I 1,1~· I E 0 0-10m 20-30m 30-40m 40-50m 50-60m 90-1 OOm 120-130m Figure 9.12: Size class frequency and distribution of the two main overstorey species (E. camaldulensis and M. leucadendra) along a transverse profile at Site 01 on the Ord River. Frequencies are from 3 combined lOm x lOm plots. Histograms show tree size classes (DBH{ cm)) by species for each plot. ,....,.N 00 Ord River - Site 02 Eucalyptus camaldulensis elevation 30 116 6 I]] Melaleuca leucadendra 5- ~ 1: Cl) 20 :::s C" 1 ~ E Ll.. 10 nounooo•••-•nmooo••••••nouumoonoonommoouooouoomommoououmomo=u••uuoofi~l••••••••n••muouunn•uoomoommuu===-IIJ omoonnuomuuuonooouoonuumooo•o•o•f~lnu••nmmmo:J ...... 0 11 11 '" >1111"" 34 t: """ 1 > "'""" " " i i i i i I I I I I 111111. IIJI -1 W"'0 ' I n I I I I i I I I '' 11 I I I I i i i I i I 0 ~Cl) ' " " " " "'''"' """ ij: ·22 :;::- ~~ V~~ y..- 0-1 Om 40-50m 70-80m 90-1 OOm 100-11 Om Figure 9.13: Size class frequency and distribution of the two main overstorey species (E. camaldulensis and M. leucadendra) along a transverse profile at Site 02 on the Ord River. Frequencies are from 3 combined lOm x lOm plots. Histograms show tree size classes (DBH(cm)) by species for each plot. .....,N \0 (a) Ord River - Site 03 Eucalyptus camaldulensis Elevation 30 s_ ;::... EJ Melaieuca leucadendra 1 E 0 6 as 20 -c ::::s 5 0 C" 4 ~ ~ > 1.1.. 10 3 Cl) 2 w 1 1 0 i i '''''j' i ,,.,, i I I I i 0 0-1 Om 1 0-20m 20-30m 50-GOm 100-11 Om 120-130m (b) Ord River - Site 04 60 30 8 7- ~ sS as 20 ::::s 5 c C" 4 .2 Cl) 3 m 1.1.. 10 - - 2 ~ 1 w 0 's 11 i 1 ,t 0 0-1 Om 20-30m 30-40m 50-GOm 100-110m 120-130m 130-140m Figure 9.14: Size class frequency and distribution of the two main overstorey species (E. camaldulensis and M. leucadendra) along a transverse profile at a) Site 03 & b) Site 04 on the Ord River. Frequencies are from 3 combined lOm x lOm plots. Histograms show tree size classes (DBH(cm)) by species for each plot. ~ 0 Eucalyptus camaldulensis elevation Melaleuca leucadendra ~ Eucalyptus gymnotaies Ord River - Site 05 30 8 184 M I ~14 2 M I 35 Ec 7 39 Ec -E ~ 6 s 20 -c :::1 5 0 C" :;:; 4 m ~ 3 u. 10 >(1,) 2 w 1 1 0 1 i i ••! 0 0-1 Om 1 0-20m 20-30m 21 0-220m220-230m 230-240m 250-260m 280-290m Figure 9.15: Size class frequency and distribution of the main overstorey species (E. camaldulensis and M. leucadendra & E. gymnotales) along a transverse profile at Site 05 on the Ord River. Frequencies are from 3 combined lOm x lOm plots. Histograms show tree size classes (DBH(cm)) by species for each plot. N ,__.N 222 a) Site 01 dry season water level M.l. 20-40; )I depositional flat E.c. 20-40; 1200/ha M.l. 20-30; E.c. 20-30; <2; 2000/ha 20-50; 1 00/ha E.c. 20-30; 50/ha depositional M.l. 10-20; 1200/ha E.c. 40-50; 20/ha E.c. 1 0-20; 50/ha Scale 250m N b) Typha Lake Kununurra <2; 2000/ha <2; 500/ha Figure 9.16: Plan view of the riparian zone at Sites 01 & 02 on the Ord River showing the median tree diameter size class (cm) and the approximate densities per hectare(!ha) for E. camaldulensis (E.c.) and Melaleuca leucadendra (M.l.). 223 a) Site 03 20/ha high 2·1 0; 1 00/ha 2-10; 20/ha M.l. <2; E.c. <2; M.l 20-30; 1 00/ha E.c. 20-30; 25/ha Scale 250m - - tN b) Site 04 ·.. E.c. <2·1 0; 500/ha M.l. <2; 500/ha 1 0-20; 1 00/ha M.l. 20-50; 50/ha E.c. 30-50; 40/ha Figure 9.17: Plan view of the riparian zone at Sites 03 & 04 on the Ord River showing median tree diameter size classes (cm) and approximate densities per hectare(/ha) for E. camaldulensis (E.c.) and Melaleuca leucadendra (M.l.). 224 Site 05 dry season r level 2-1 0; 1 00/ha E.c. M.l. <2; 1000/ha <2; 200/ha E.c. 20-30; 20/ha E. gymnotales 20-30; 1 00/ha M .I. 2·20; 200/ha E.c. 2-10; 20/ha Scale 250 m I ...... N Figure 9.18: Plan view of the riparian zone at Site 05 on the Ord River showing median tree diameter size classes (cm) and approximate densities per hectare(/ha) for E. camaldulensis (E.c.) and Melaleuca leucadendra (M.l.). 225 At Sites 03 and 04 river water levels dry up to a small stream and eventually to isolated pools around which most seedlings are located (Figure 9.17a&b). Juvenile trees in smaller size classes occur in bands on the flood-out depositional area on the inside of the bend, with mature trees occurring on the steeper banks of the erosional area of the outside of the river bend. A similar pattern is seen at Site 05 (Figure 9.18), with large trees on the steeper western bank and juvenile trees occurring on the gradual slope of the flood-out area, and with occasional larger E. camaldulensis and E. gr;mnotales at the top of the slope. Long term hydrograph Over a 37 year period maximum discharge on the Blackwood River consistently occurred in the winter in the wet part of the year between July and August. Peak flows were regular events each year occurring with the regular winter rains. The one exception was an extremely large flood event in January 1982 caused by a cyclonic feature drifting down from the tropics into the south west land division (Figure 9.19). Maximum discharge levels are much smaller and less variable than those seen on the Ord River, with a mean maximum discharge of 161 m3s-1 ± 24.5. Even the 3 1 exceptional event in January 1982 (with a peak flow of 1187 m s- ) was much smaller than average maximum discharge on the Ord River. Average monthly discharge shows a peak in July with river levels usually receding by September. Over a 19 year period on the Ord River large river flows have been confined to the wet season with 100% of discharges greater than 500 m3s-1 occurring between November and March. Recorded flows have only occurred 3 times outside this wet season period in June 1978, May 1979 and October 1988 (Figure 9.20a). Magnitude of peak flows is highly variable, with maximum discharge of> 15000 m3s1 occurring 4 times over the 19 year period and 4 years having a discharge <1000 m3s·l, with only 375 m3s-1 in 1992. Average maximum discharge (±SE) for the period from November to March is 5565 m3s-1 ± 1475, with January having the highest mean discharge and March the lowest. For the lower Ord River below the diversion dam (near Site 01) the river flow regime prior to river regulation has been modelled from rainfall and runoff data (J. Ruprecht, unpublished data) and compared with the current flow regime (Figure 9.21). This 226 shows that variation in monthly flows has been much reduced, with a large reduction in wet season flows and greatly increased dry season base flows where, prior to regulation, the river would dry up to a series of pools. Tree age and hydrology Growth ring counts from stem sections and cores showed a significant relationship between number of rings and stem diameter for E. rudis on the Blackwood River and E. camaldulensis and M. leucadendra on the Ord River (Figure 9.22 a,b&c). Using this relationship, an estimate of the age of the trees can be made from stem diameter alone. At Site B1b on the Blackwood River a stand of trees growing in a slight depression 50 m from the water with a mean DBH of 14.3 cm in 1996 (Figure 8.8c) would be approximately 16 years old. This fairly well coincides with the very large, unseasonal flow event in January 1982 (Figure 8.19), 14 years earlier. The establishment of these trees at this time has been confirmed by the local landowner. On the Ord River at Site 04 the establishment of cohorts of E. camaldulensis at 8 and 20 years ago (Figure 9.6c) coincide with a year of high flow followed by several years of low flow (Figure 9.20b). This would create the conditions suitable for seedling establishment and allow several years without severe flooding in which to develop. Discussion Population structure of the Blackwood River species indicates that stable populations with a range of sizes occur in ungrazed areas. In areas that are grazed by livestock there is a lack of smaller size classes, indicating stagnant or senescent populations. On the Blackwood River sites trees develop to maturity throughout the river profile, from the river channel to the extent of the riparian zone and there is fairly even frequency distribution of size classes, suggesting non episodic recruitment. 800 1187 700 ~- 600 fiJ .1 500 -& 400 :I...m .c 300 0 fiJ c 200 100 1 1 I' oloiliiFii ) I } I illllhh \_ I ~ I• iiiOh~ I \ I 1 .0 ii i • il illI IU1 ..) ,~ \ I ,,r,.)i ihii, ~ plu)a\!u) r ~ 0 'nloi'?A"n?i' \ ~ ~ aa hili anuh( '' ' ) '''"'""'' ,l(rio,im\ (, \ ...... , .... ,.t\,o),r\',?1 ""< \'\<: 0 \ j ~ ilirliilii'l'il!a+j ~ ul iiliii~iAilfl) ,,, .. l,,, .. fii114il' ro m o ~ N ~ ~ ~ ~ ~ ro m o ~ N ~ ~ ~ ~ ~ ro m o ~ N ~ ~ ~ ~ ~ ro m o ~ N ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Figure 9.19: Maximum monthly river discharge over a 37 year period for the upper two thirds of the Blackwood River catchment from a recording station near Site Bla. N N 'I Discharge (m3s-1) Discharge (m3s-1) 1-1<" ...."T.I ...... (p(Jq 1\) ~ 0') CJ1 0 CJ1 1-1 ,:::: 0 0 0 0 0 0 () 1-1 0 0 0 0 0 0 Pl (0 0 0 0 0 0 0 0 0 .-t-1..0 ()• 1/1/71 1/1/71 m s~ -(/) ~ s;:: 1/1/72 (/) 1/1/72 :;::;: !""1-PJ -·.... m § ~. m o..S 1/1/73 1/1/73 0 Pl s:: 0 (J1 .-~-s ~ 1/1/74 :2:s I CJ)O :5: ...... ::~ :5: m rt:;. 1/1/75 m 1/1/75 >< oq >< -· ~~-~ -· 1/1/76 3 " ...... 1/1/76 3 c 0 ~ ~ 1/1/81 .,jl 1/1/81 s ~ ..§~ "Fr.Jro '- 1/1/89 ::l'lj :;.'"g (0 1-1 1/1/90 -i$ 1/1/90 '-< ...... 1/1/91~ 1/1/91 1-'•0 so ...... :;. 1/1/92 (0 1/1/92 ~ 0 1-1p_. 1/1/93-i-- 1/1/93 8Z:Z: 229 1800 1600 ..J 1400 -CJ -:;!: 1200 0 --o- Natural flow u: 1000 ~ ...... () ...... Regulated flow .s: -c 800 0 ::!: c 600 C\'1 Cl) ::!: 400 200 0 ..!: ·;: >- Cl) ;:...... ~ t=' u C\'1 c :; Ill Cl) Cl) (!) Cl) C\'1 C\'1 ... Q. ::l -::l .Q .Q .Q .Q ::l ::l C\'1 .;:;( ::!: -:1 0'1 0 c ... -:1 ::; E E E C\'l .Q ::!: Cl) """u Cl) (!) -:1 (!) < 0 > u u. Q. 0 ID -(!) :z c 00 Figure 9.21: Monthly river flow (GL) for the lower Ord River at Carlton Crossing below the diversion dam. Natural flows are for the river prior to dam construction in 1963 and is calculated from modelling. Regulated flows are the present flows experienced with Stage I of the Ord River irrigation district. 230 a) Eucalyptus rudis 2 40 y = 0.684x + 6.600 r = 0.687 p<0.01 0 30 C) 1:: ·;::::: 20 -0 0 z 10 0 0 5 10 15 20 25 DBH (cm) b) Melaleuca leucadendra 40 y = 0. 767x + 6.009 r2 = 0.~17 35 p 0 0 5 10 15 20 25 30 35 40 DBH (cm) c) Eucalyptus camaldulensis 40 y = 0.969x + 4.727 r2 35 p<0.01 • m 30 • C) c ·;::::: 25 20 -0 • ci 15 z 10 5 • 0 0 5 10 15 20 25 30 35 40 DBH (cm) Figure 9.22: Relationship between tree diameter at breast height (l.Sm) and tree age estimated from counting annual growth rings for (a) Eucalyptus rudis on the Blackwood River (b) Melaleuca leucadendra on the Ord River and (c)Eucalyptus camaldulensis on the Ord River. 231 On the Ord River tree populations are more dynamic, with large numbers of individuals in the <2 cm diameter size class and an uneven distribution of size classes. There are obvious cohorts of individuals which implies that environmental disturban~es are affecting recruitment which is likely to be episodic (Bradley & Smith 1986). These differences in population structures and spatial distribution in the riparian zone and the very different hydrologic regimes suggest a need for different vegetation dynamics models for the two rivers. On the Blackwood River recruitment appears to generally occur continually, with biotic processes such as competition being important. The population structure of these species can be viewed as consisting of long stable states with a short transition between states. In contrast, on the Ord River, recruitment is most likely episodic and is driven by abiotic processes such as high river flows, with the population structure described in terms of long periods of transition and short relatively stable states. Population structure On the Blackwood River E. rudis is the major overstorey species (Marchant et al. 1987) and the much greater numbers seen on the transects measured here confirm this. At the sites that are subject to stock grazing the effects on tree population structure is obvious, with no stems under 10 cm DBH recorded. This suggests that there has been no recruitment into the population for around 15 years, and with the greatest proportion of stems in the large size classes (mature trees), it would also indicate that the population is in decline Gohnson 1994). Although seed germination is taking place at these sites (see Chapter 8) seedlings are not surviving, mainly due to grazing by livestock (see Chapter 6). The understorey at these sites is also highly modified (see Chapter 4}, with extensive invasion of exotic grasses which may also contribute to poor survival of seedlings through competition (De Steven 1991). Free from livestock grazing, recruitment is taking place, with large numbers of stems in the smaller size classes indicating a more stable population Gohnson 1994). For E. rudis there are no obvious cohorts of similar sized trees which would indicate that episodic events such as flooding are not an important mechanism for recruitment and that continuous recruitment takes place for this species. Anecdotal evidence (corroborated by tree ring counts at Site 1B) does suggest however that after large unusual 232 disturbance events, such as the exceptional flood in January 1982, large recruitment events can also occur. Results here also indicate that E. rudis can establish and mature anywhere throughout the profile, including at low elevations near the water's edge, and that establishing trees are not unduly affected by floodwaters. M. rhaphiophylla at the sites surveyed on the Blackwood River is usually a mid-storey species to E. rudis but can become locally dominant (Marchant et al. 1987). It is mainly restricted to the wetter areas of the riparian zone and usually occurs on the littoral fringe of the river between summer and winter water levels. It does occur further from the river in moisture gaining depressions and this is probably related to waterlogging tolerance and/ or lack of drought tolerance. Population structure of M. rhaphiophylla on the Blackwood River shows similar patterns to E. rudis with an effect on recruitment with livestock grazing but a stable population at ungrazed sites. This species is restricted to closer to the river and does not grow as large as E. rudis but within this zone there is an even spread of size classes. The better adaptation to inundation at the seedling stage in M. rhaphiophylla was demonstrated in the flooding experiment in the previous chapter and may partly explain their distribution closer to the river channel than E. rudis. Greater tolerance, at the seedling stage, to inundation has been used to explain the differences in distribution in the riparian zone of two rainforest species in Victoria (Melick 1990). For E. camaldulensis on the Ord River the greater proportion in the smaller size classes and few larger mature trees suggest a building population (Harper 1977). The cohorts of trees identified indicate that recruitment is most likely episodic with some background recruitment. However stems of intermediate size between cohorts may also represent suppressed or exceptional individuals. There was also clearly a relationship between elevation and tree size, with m large mature trees at low elevations. This is also related to river geomorphology, with larger trees occurring at higher elevations near the river on eroding inside river bends, whereas m depositional outside bends and flood-out areas at low elevations only smaller trees occur. Therefore, while germination is more common at the lower elevations where moist open 233 sediments provide ideal sites for seedlings, growth to maturity is more likely at higher elevations which are less likely to be flooded Gones et al. 1994). Lack of stability of soils at lower elevations due to the movement of floodwaters is li~ely to be a factor in the poor survival to maturity. This lack of soil stability preventing the development of mature populations has been reported for shrubs and trees on the floodplain in a arid zone catchment (Friedel et al. 1993). The most abundant species at the Ord River sites was M. leucadendra. Population structure was similar to E. camaldulensis with an abundance of young stems and few large mature trees again indicating a building population. Like M. rhaphiophylla on the Blackwood River it is confined to the wetter areas, usually close to the river, and is therefore better adapted to inundated conditions, such as through the formati~n of adventitious roots under anoxic conditions during inundation (Boland et al. 1989). As for the Blackwood River species, the better adaptation of M. leucadendra to inundation at the seedling stage was demonstrated in Chapter 8. Tree age Estimating tree age by counting tree rings is a common technique used in determining tree population structure (Bradley & Smith 1986; Duncan 1993; Johnson 1994) or past environmental disturbances (Baker 1990; Villalba & Veblen 1997). However, there are some limitations to this approach due to technical difficulties and confounding effects (Villalba & Veblen 1997). For example, there can be large variations in the size within a cohort due to habitat heterogeneity and competition between individuals (Hutchings 1997). Also, most of this work has been done in the northern hemisphere in climates with strong seasonality which make the detection of annual rings much easier (Ogden 1978). More work is needed to develop a reliable dendrochronological technique for Australian plants but this is beyond the scope of the present study. Therefore aging of trees by ring counts in the present study must be interpreted with some caution. It does indicate however that establishment of trees is related to past hydrological events. For the Ord River this may involve the germination of seed at higher elevations after a large flow event in the wet season which would enhance seed transport through hydrochory and recharge soil moisture (Roberts 1993). Lower flows in subsequent years would prevent the 234 seedlings from being scoured out or inundated at least until they had become established. Unseasonal dry season rainfall in at least the first year after seed germination would also enhance survival. This link between tree recruitment and particular flow events has been established in many studies of riparian trees (e.g. Bradley & Smith 1986; Baker 1988; Johnson 1994). Tree vegetation dynamics Fluvial processes on the Ord River are dominated by fairly frequent, large, high energy floods that are capable of scouring out the soil, destroying the existing riparian vegetation community and preventing the establishment of stable mature stands. This is evidenced by the existence of large mature trees on these river reaches only at high elevations, in protected river bends or far from the river where only extreme floods can reach. In the years between large flood events large stands of seedlings and saplings develop but these generally do not reach maturity before they are destroyed in the next large flood. High frequency flooding can keep the vegetation in an early stage of succession (Naiman et al. 1998). Failure to reach later successional stages due to high frequency flood disturbance has been documented for New Zealand Podocarp forest (Duncan 1993). In this way, single catastrophic events change the ecosystem to an extent which is inconsistent with the theory of climax communities (George et al. 1992). These high magnitude wet season floods are, however, highly episodic and in the intervening years smaller floods allow the development of the existing trees and the recruitment of younger age classes. This regime of intermittent high frequency large disturbances prevents the establishment of stable states and in terms of the state and transition model the ecosystem is characterised by long periods of transition between short-lived stable states. This riparian ecosystem is thus driven by physical (allogenic) processes rather than by vegetation successional (autogenic) processes (Baker & Walford 1995). In contrast, lower energy seasonal flooding on the Blackwood River allows mature stands of trees to develop throughout the river profile. Recruitment is continual, although species can also respond to large flood events. This disturbance regime results in long periods of stable states with short periods of transition. Therefore the vegetation is subjected to longer periods of autogenic processes and, because of lower frequency flooding disturbance, shorter 235 periods of allogenic processes. Where flooding disturbances are chronic to frequent, riparian forests are young, as vegetation patch development is continually reset and remains in stand initiation and exclusion successional stages (Wissmar & . Swanson 1990). In environments with low magnitude, seasonally predictable floods the vegetation is likely to be more stable and more developed. The fact that no cohorts of trees were discerned at Site 01 on the regulated section of the Ord River suggests a change in dynamics of these species to a regime of continual recruitment brought about by the reduction in flooding disturbance. Large dense stands of younger aged trees now occur in a narrow band close to the river. Without large flood disturbance thinning out these stands strong competition may prevent the development of larger size trees. At Site 02 on Lake Kununurra where river flows no longer occur and the riparian zone now acts like a lentic system and water levels are permanently much higher, there are two distinct populations of trees. There are relics of larger older trees of the former riparian system and a cohort of young saplings developing on the newly created narrow littoral zone where the lake levels fluctuate. This has created a new niche of exposed moist bare sediments suitable for seed germination. Given that this area is no longer subjected to disturbance from large floods, it would be expected that population structure would become more stable at this site as saplings develop to maturity and there is less opportunity for recruitment. The hydrological regimes for the Ord and Blackwood rivers are vastly different despite similar rainfall and catchment size. This is mainly due to the different geology of the catchments (Department of Mines 1972) and rainfall patterns. The Ord River catchment is subject to high intensity rainfall events mainly generated from sub-tropical monsoons. These are confined almost entirely to the wet season between October and March and contribute high levels of runoff into the river. For the Blackwood River, although it has a strong winter (Tune-August) rainfall maximum, events are less intense and runoff is not as high (see Chapter 3). These results highlight the very different fluvial regimes that exist on these two rivers and their effect on the respective models of vegetation dynamics for the overstorey species. This is despite the species compared on the two rivers being functionally very similar. Management of 236 the riparian vegetation should therefore take into account the frequency of change in the vegetation and that the disturbed states and long periods of transition between states are part of the natural process. This would suggest that altering the. natural flow regimes, such as through river regulation, will have significant effects on riparian vegetation dynamics. This is particularly so for the effects on long term recruitment in terms of reproductive phenology and seedling survival (Bren 1992; Johnson 1997). In the case of regulation on high disturbance rivers such as the Ord River, riparian vegetation dynamics may be significantly altered to long periods of stable states with much reduced transitions. This has important implications for the long term structure and functioning of these communities. This Chapter has shown the differences in population structure and vegetation dynamics of the study species and the strong link with river hydrology, particularly for the Ord River where recruitment appears to be episodic and development of the population is continually reset by the destructiveness of large flows. Recruitment of species on the Blackwood River appears to be more subtle and not as closely linked to the river hydrology, but possibly influenced also by biotic factors. This can also be related to other elements of regeneration discussed in the preceding chapters, such as seed production and dispersal, the timing of seed release to river flows and the annual release of seed of the Ord River species. 237 CHAPTER 10 SUMMARY AND GENERAL DISCUSSION An overall aim of this thesis has been to provide information on the processes involved in the regeneration and recruitment of riparian vegetation, using examples from two contrasting river systems. There is a general paucity of basic ecological studies on riparian vegetation in Australia (Dexter 1978; Bren & Gibbs 1986; Roberts & Ludwig 1991; Roberts 1993; Walker et al. 1994). This thesis is therefore broad in, scope and provides a general picture of regeneration processes on the study rivers and establishes a starting point for more detailed work in this area. The two rivers chosen for study, despite being very different in terms of hydrology and climate, show similarities in vegetation processes which can be more generally applied. Species selected for more detailed study are functionally similar and therefore can be used to clarify ecological adaptations of each to their particular environments. Riparian landscapes are formed by the interaction of hydrologic, geomorphic and ecological processes (Malanson 1993) and consequently regeneration mechanisms are likely to be somewhat different for each river. The combination of specific knowledge of natural flow regimes and an understanding of vegetation dynamics is therefore required when planning for the effective management of riparian vegetation. The interaction of elements that have been found to influence the recruitment of riparian vegetation on the study rivers is summarised in Figure 10.1. The thesis has been structured to look firstly at the community level of riparian vegetation on both rivers. From this, particular species (Eucalyptus rudis and Melaleuca rhaphiophylla on the Blackwood River and E. camaldulensis and M. leucadendra on the Ord River) have been selected for more detailed investigation of their population structure and ecology and of the influence of environmental factors (particularly hydrology) on their recruitment. In Chapter 4 a general description of the vegetation included patterns of species richness, diversity and spatial distribution of the plant communities in the riparian zone as well as key functional groups in 238 terms of life form and species origin. From this Chapter the importance of trees in the riparian zone became evident and their key functional role is well documented (Malanson 1993; Bunn 1993; Naiman et al. 1998), so that this life form became· the obvious choice for more detailed investigation in later Chapters. It is also clear that livestock grazing has a major impact m riparian vegetation, particularly on the Blackwood River. Chapter 5 describes the relationship that exists between historical river flows for the Ord and Blackwood rivers and aspects of riparian vegetation ecology. It provides an example of how historical flow records can be used to develop a picture of the natural flow regime which can then be related to existing patterns of vegetation development and plant community patterns in the riparian zone. An assessment of the effects of livestock grazing on riparian vegetation was made and the initial response of the communities to the exclusion of livestock is examined in Chapter 6. This indicates that grazing has a major impact on the recruitment of tree species on the Blackwood River but this effect is not obvious on the Ord River. The reason for this difference is most likely the differences in the process of seedling recruitment on the two rivers but it may also be influenced by differences in stocking rates. These differences in the processes of seedling recruitment are explored in the next three chapters. From the examination of sources of seed in Chapter 7 it appears that the soil seedbank is only important for annual species and that, in particular, predation and poor survival of seed in the soil of the examined overstorey species reduces their likelihood of persisting in the soil seedbank. The fact that the reproductive phenology of the two target species on each river is timed to river flows and that seed is found in flood debris indicates that these are a more likely source of seed for recruitment. The question of whether this timing to river flows also applies to recruitment of seedlings is examined in Chapter 8. This includes determination of the optimum temperature requirements for germination of seed of these species, the site conditions likely to allow successful early establishment of seedlings and the level of tolerance over time to various degrees of flooding. Whether these findings can be related to the existing population structure and spatial distribution is explored in Chapter 9. The population structure and spatial distribution in the riparian zone of the target species suggests that different models of vegetation dynamics operate on the two rivers and these can be related to fluvial regime. From these re:5ults the general hypothesis stated in Chapter 1, that the fluvial regime 239 has a strong influence on the vegetation dynamics in the riparian zone has not been rejected. How the fluvial regime influences various aspects of the riparian vegetation is detailed in the following sections. Seed Ecology reproductive phenology dispersal types Natural flood debris (water dispersal) Fluvial soil seedbank Maintenance Regime vegetative + reproduction Seedling Survival & Growth adaptation to flooding microtopography ~Livestock t Population Structure Figure 10.1: Interaction of the elements important in recruitment of riparian vegetation. Riparian vegetation communities Describing the riparian vegetation community and the range of different plant life forms that occur is an effective way of identifying the important functional groups and the range of regeneration strategies of vegetation of the riparian zone of these rivers. Chapter 4 highlights the dominance of trees in the riparian zone on both rivers. On the Blackwood River the understorey is dominated by annual herbs and grasses, particularly at sites that are subject to livestock grazing, while perennial grasses are the dominant feature of the understorey on the Ord River. Greatest diversity on the Ord River was in the number of overstorey species, with the 240 understorey tending to be dominated by only a few species. Species diversity on the Blackwood River was most influenced by the understorey species which, particularly on the ungrazed sites, was composed of a high diversity of shrubs and perennial berbs. The overstorey was composed of almost exclusively Eucalyptus rudis. For the study rivers there was a high proportion of disturbance tolerators (ruderals), such as annual and perennial grasses, and low numbers of stress tolerators which are commonly woody perennial species (Grime 1979). This would indicate that disturbance has had a strong influence on the composition of both vegetation communities, although stress is also an important contributing factor in the form of waterlogging and drought. For the two rivers species richness tended to increase with distance from the river and therefore with lower frequency of flooding. In particular, a sharp decline in species richness is seen in plots near the water's edge in the littoral zone which is subject to the greatest flooding disturbance and greatest stress. This is generally the zone of helophytes or emergent macrophytes such as from the Cyperaceae family. Vegetation patterns, and therefore species richness and diversity, in the high disturbance riparian zone are strongly influenced by species colonising ability (White 1979). The high frequency disturbance regime and the harsh (high stress) environment precludes many species as plants need to be able to tolerate extreme conditions such as flooding in the wet season and prolonged periods of drought in the dry season. The intermediate disturbance hypothesis (Fox 1979; Huston 1979; Ward & Stanford 1983) states that greatest species richness occurs in areas where frequency of disturbance is high enough to create gaps for species establishment but not so high as to exclude species (such as through the scouring of the river banks). Clearly, the heterogeneous environment of the riparian zone allows many species to persist under these conditions and the impact along the gradient of disturbance and stress can be modified by small spatial scale variations (e.g. microtopography) (Pollock et al. 1998). Further work is required to measure species diversity differences between the riparian zone, the surrounding floodplain and the terrestrial environment. This would examine, for Western Australian rivers, whether the riparian zone is an area of high species richness and follows the generally accepted idea of the riparian zone being a transition zone of higher species richness as the river collects propagules of species throughout the watershed (Nilsson et al. 1994). 241 The fluvial regime and the influence on riparian vegetation Characterising the river fluvial regime of the study . rivers is an essential first step in determining the relationship with the riparian vegetation. Many aspects of the ecology of riparian vegetation are similar to those of vegetation in other environments such as wetlands, floodplains and even terrestrial environments. Flowing water and its effect on riparian vegetation is what makes this community type unique (Malanson 1993). This study has highlighted the importance of the fluvial regime in shaping vegetation community structure and floristics, as well as in determining the distribution and abundance of species. The fluvial regime impacts upon the dispersal of propagules, the reproductive phenology of particular species, germination of seed and the establishment of seedlings and plant community population dynamics. River flows shape the structure of the riparian habitat and contribute to the redistribution of soils and soil nutrients. Replenishment of soil water and grmmdwater across a gradient are also influenced by the duration and frequency of flows into the riparian zone. Other ecological factors obviously also influence vegetation dynamics in the riparian zone. For example, climate (including seasonality and variation in rainfall and temperature) has a strong influence on the floristics of a region (Groves 1994). This is evident in the very different floras of the two rivers in this study and from the ordination of vegetation plots in Chapter 4, which clearly shows the influence of amount of rainfall on longitudinal changes in the vegetation community along each river. Fire, as a form of disturbance, can also affect vegetation community development (Gill 1981), and specifically in the riparian zone (Douglas 1999; Busch 1995). Another disturbance which impacts on many Australian plant communities is livestock grazing (Wilson 1990), which is implicated in the degradation of the riparian zone on many rivers (CEPA 1992). These latter factors operating separately, or more likely, through interaction, create patchiness (mosaics) in the vegetation of the riparian zone across gradients (zonation), which is driven by the fluvial regime. Identifying the flow regime, in terms of variability, predictability, constancy and contingency (Colwell 1974), from the historical hydrological 242 records, can therefore be useful in describing the vegetation dynamics of a particular river (Wissmar & Swanson 1990; Naiman et al. 1998). An example of this was provided in Chapters. Livestock grazing and fire Grazing by domestic livestock is a human-induced disturbance that clearly has the potential to impact on recruitment in the riparian zone. Responses of the riparian vegetation to livestock grazing were very different for the two rivers in this study. Large sections of the riparian zone of the Blackwood River occur within a matrix of agriculturalland cleared for improved pasture and cereal cropping. Native understorey species have been effectively cleared by high intensity livestock grazing and replaced with introduced pasture species and pasture weeds. Increased fire frequency can also contribute to this process (Leigh et al. 1991; Hobbs & Huenneke 1992). Invasion of exotic species has also probably been aided by the transport of propagules in floodwaters. The combination of removal of native species, increased nutrient levels (Cale & Hobbs 1991; Scougall et al. 1993) and a high disturbance riparian environment (flooding) has provided ideal conditions for ruderals which are generally exotic annual species (Grime 1979). Grazing has also prevented the recruitment of overstorey species which constitute the 'skeleton' of the riparian vegetation, but because of the long life span of these species the effects will take a long time to became apparent (Pettit et al. 1998). Livestock grazing effects on the Ord River, in contrast, are not as easily detected. Low stocking rates, the fact that the riparian zone occurs in a matrix of predominantly natural vegetation that has not been cleared and the grassy woodland nature of the vegetation contribute to the apparent reduced impact. All areas of the Ord River have been exposed to cattle grazing, so it was not possible to compare current vegetation community composition with riparian sites that have never been grazed. In fact, only a few very isolated rivers in the Kimberley region have not been subjected to some degree of livestock grazing (Water & Rivers Commission 1997). Size class analysis of overstorey species has suggested that some recruitment of these species is taking place. Long term monitoring of exclosure plots will give some indication as to whether species richness, vegetation cover and number of shrubs is lower due to livestock grazing. Further work is required to clearly establish the vegetation 243 composition prior to grazing as is an expansion of the grazing exclusion trials with a long term monitoring programme. The response of a plant community to grazing varies with community type and evolutionary history of grazing (Milchunas & Lauenroth 1993). In environments that have had a long history of grazing, herbivore grazing is considered essential for maintaining species diversity and in many communities species diversity has decreased when grazing has ceased (Carr & Turner 1959; McNaughton 1985; Noy-Meir et al. 1989; Smith & Rushton 1994). Such communities are generally ·dominated by grasses and forbs, whereas in vegetation communities dominated by woody perennial species, species diversity decreases with livestock grazing (Williams & Ashton 1987). Severe over-grazing will clearly h~we a substantial impact on riparian vegetation, with widespread loss of plant cover and subsequent erosion problems such as have occurred in areas of the Ord River catchment (Fitzgerald 1967; de Salis 1993). The nature of the different drivers of vegetation dynamics operating in the two rivers may also contribute to this difference in response to livestock grazing, with allogenic or physical processes (such as the flooding regime) being important on the Ord River and autogenic or vegetation successional processes being dominant on the Blackwood River. Exclosure experiments on the Blackwood River show little improvement after 3 years, with only minor increases in the occurrence and cover of native species. Exclosure experiments in riparian zones in the United States have shown an increase in cover and number of species of woody perennial plants with relief from grazing (Shultz & Leininger 1990) and an increase in competition with rapid vegetative growth of strong competitors (Kauffman et al. 1983). The exclosure plots in the present study are highly degraded and perhaps do not contain sufficient storage of vegetative and sexual reproductive material of native species, such as underground storage organs or a soil seedbank, to support natural regeneration. Such sites may require intervention in the form of weed removal and reseeding or replanting to re-establish native species. Establishment in these species may be restricted by the current abundance of exotic grasses and annual herbs which have increased in cover in the absence of grazing. The effects of 244 competition from exotic weeds on the germination and establishment of riparian overstorey species is an area that requires further study. For exclosure experiments, a much longer period of observation is required to fully understand the vegetation dynamics and successional processes, particularly for long-lived perennial species such as the overstorey species. Pioneer species such as annuals may eventually be replaced by another suite of species depending on future disturbance and management (Onans & Parsons 1980; Wahren et al. 1994). For example, some episodic disturbances such as a large flood and/or particular climatic conditions may be required for a successful recruitment event to take place. As for other vegetation communities, frequent fire can have a significant influence on riparian vegetation communities. In south-west Australian rivers such as the Blackwood River. fire, as a disturbance, together with livestock grazing, can lead to greater degradation of the vegetation (Leigh et al. 1991). This is chiefly through the increased invasion of exotic ruderal species which are better adapted to this combined disturbance regime. Burning may enhance the germination of seed of native overstorey species and improve short term survival by creating areas of bare ground suitable for seed germination (Wellington & Noble 1985b) but once the biomass of annual species recover these may quickly reduce survival of seedlings. In tropical savanna in northern Australia season of fire appears very important in shaping community structure in the riparian zone (Douglas 1999). Annual late dry season burning results in significantly less diverse and abundant riparian communities and recovery is slower than after burning early in the dry season. Frequent (annual) fires in the riparian zone may also lead to the loss of rainforest type species such as lianas. Sources of seed for regeneration Strategies for regeneration of riparian vegetation include aspects of the reproductive phenology, dispersal of propagules, and the storage of mature seed ready for germination when conditions are suitable. This study indicates that the soil seedbank may be important for annual species of herbs and grasses but of only minor significance for perennial species. In lentic wetlands, the seedbank can be important in determining the vegetation community (Schneider & Sharitz 1986; 245 Thompson 1987; Brock et al. 1994) and determination of the composition of the soil seedbank is used to predict the outcome of succession after a disturbance (van der Valk 1981). Little work has been done on the role of the soil seedbank in lotic systems, however the highly disturbed nature of the riparian zone and the unstable nature of soil sediments indicate that the role would be minimal (Malanson 1993). A comprehensive survey of the soil seedbank in the riparian zone, including looking at the seedbank in different habitats (such as pool and riffle areas and erosion and depositional bends and floodplain wetlands), would provide a more complete picture of the role of the seedbank in the riparian zone. This could also include sampling at different times of the year to discern the long term from the transient seedbank. Storage of seed on the plant (serotiny) can provide another effective seedbank (Lamont et al. 1991) and in Australia many species in the Proteaceae and Myrtaceae families use this mechanism (Enright & Lamont 1989; Beardsell et al. 1993). Eucalyptus spp in general show a varying degree of serotiny. However, eucalypt seed is rarely held in the canopy for more than 3 or 4 years (House 1997). Of the overstorey species examined in this study only Melaleuca rhaphiophylla appears to use this device. Reproductive phenology of the four species studied reflects their adaptation to their environment and in particular to the fluvial regime. For E. camaldulensis, M. leucadendra and E. rudis the period of greatest seedfall is timed to coincide with receding floodwaters, when moist sediments favourable for seed germination and survival are exposed. This adaptation has been reported for riparian species of Populus and Salix in the U.S. (Bradley & Smith 1986; Johnson 1994). The Ord River species release the majority of their seed over a short period at the end of the wet season, this being the only opportunity for recruitment for the year. Other northern Australian Eucalyptus spp have been reported to display a similar reproductive phenology (Bowman et al. 1991; Fensham 1992; Setterfield & Williams 1996). For many temperate Eucalyptus spp, including E. rudis on the Blackwood River, seeds are retained in the capsules with a continuous low level of seed release occurring which is accelerated by dry conditions (Cunningham 1957; Wellington & Noble 1985a; Davies & Myerscough 1991). For E. rudis, with a ready supply of seed available in the canopy, there is also potential for explosive recruitment 246 after a major flood event coinciding with other favourable environmental conditions. Many species show a high degree of variability in reproductive phenology and trees need to be monitored over a much longer time to confirm the results reported here. Although wind dispersal was found to be the most common primary means of propagule dispersal, enhanced or secondary dispersal by floating downstream or laterally within the riparian zone or with flood debris is clearly a consequential recruitment mechanism (Barrat Segretain 1996). In rivers in Sweden, floodwaters carrying seed greatly extended the dispersal range of wind-dispersed species, depending on the ability of the seed to float and the length of time they could float (Skoglund 1990). Floating time can also be important for dispersal distance, with larger, heavier seeds establishing closer to the point of release (Nilsson et al. 1993). Lateral transport of propagules on floodwaters can also be important in depositing seed at higher elevations thus affording some protection for developing seedlings from subsequent flooding. The presence of seed of many common riparian species in debris samples in this study indicates the importance of secondary dispersal of seed by water. This has been reported for other riparian habitats (Schneider & Sharitz 1988; Johansson et al. 1996). Seed and fruit of many terrestrial species were also found in debris samples, indicating that overland flow tends to concentrate species in the riparian zone (Nilsson et al. 1994). Mechanisms of seedling recruitment Temperature and moisture play a critical role in providing the stimulus for seed germination. This is largely determined by climatic conditions, with many plants having evolved to germinate when conditions are most favourable for survival (Mott & Groves 1981). Optimum temperature for germination was similar for Ord River and Blackwood River species despite very different climatic and hydrological conditions. This may be related to the common features of riparian environments where moisture is readily available (Bell & Williams 1997). Position of seedlings within the riparian landscape is also strongly related to environmental conditions that would provide adequate moisture but protect seedlings from flooding. Microtopographical rises, or areas of slightly higher elevation, are important for seedling establishment in creating 247 safe sites to prevent prolonged submersion and waterlogging (Bren & Gibbs 1986; Skoglund & Verwijst 1989; Titus 1990). Ideally, these rises should also be slightly concave to allow collection of vital moisture from dry season rainfall events. Of the study species, the Melaleuca spp. showed a greater degree of flooding tolerance than the Eucalyptus spp., which reflects the position in the riparian landscape in which they occur. The formation of adventitious roots and stomatal closure were considered important factors in the flooding tolerance of Melaleuca quinquenervia seedlings, a species of northern Australia (Senna Gomes & Kozlowski 1980). Waterlogging has been demonstrated experimentally not to affect the growth or leaf water potential of E. camaldulensis seedlings (Pereira & Kozlowski 1977; Akilan et al. 1997). Vegetative resprouting is particularly co1nmon in communities which receive frequent periodic disturbance (James 1984), such as in many wetlands and rivers (van der Valk 1992). The importance of vegetative propagation for species is therefore mainly to allow them to persist at a site after a disturbance. The rapid recovery of established resprouters gives them a competitive advantage over seedlings so that resprouting species may dominate the post disturbance habitat (Malanson & O'Leary 1982). In highly disturbed riparian habitats, the ability to resprout as well as to produce large quantities of seed would be highly advantageous. Many species that occur in riparian habitats in North America, such as Salix spp. and Populus spp., possess the ability to resprout and to produce large quantities of seed (Johnson 1994). This is also seen in some of the riparian trees in Australia, including Eucalyptus camaldulensis and Melaleuca leucadendra on the Ord River and E. rudis and M. rhaphiophylla on the Blackwood River. Seedlings that establish on new substrates must be able to withstand high water flows in the following or subsequent seasons which could damage and defoliate the plants. An ability to resprout also enables mature plants to survive physical damage caused by flooding. However, in large floods where trees are destroyed by the scouring of sediments, recovery can only be from seed. 248 Vegetation dynamics and population stn1cture in the riparian zone A clear insight into vegetation dynamics is essential in understanding the processes of regeneration in plant communities such as the riparian zone. Changing views of vegetation dynamics suggest that changes seen after a disturbance may not be unidirectional and that successional models may be too simplistic in describing vegetation patterns (Hobbs 1994). The present state of an ecosystem is the result of a particular history of disturbance, climatic factors and/or management. The resultant mosaic of disturbance patches and successional stages in a landscape has been described in terms of multiple stable states (Forman & Godron 1981). The regeneration process and longevity of the vegetation can be critical in determining the stage and rate of transition between stable states (Hobbs 1994). Therefore, non-equilibrium vegetation dynamics described as discrete states and definable transitions between states may better reflect vegetation communities, particularly in highly disturbed environments. State and transition models described for rangeland management (Westoby et al. 1989; Friedel 1991; Laycock 1991; George et al. 1992) allow for different vegetation community composition (states) depending m the pattern of transition. In the riparian zone the high level of disturbance can result in a mosaic of vegetation communities (Kallida & Puhakka 1988) which directional succession theories do not adequately describe because of the complex nature of the interactions taking place (Szaro 1990; Baker & Walford 1995). The state and transition model may provide an appropriate framework for interpreting the response of riparian vegetation to disturbance. The results of this study do suggest that this model may more accurately describe the vegetation dynamics m these very different rivers. Fluvial processes on the Ord River are dominated by frequent, large, high energy floods that are capable of destroying the existing riparian vegetation community and preventing the establishment of stable mature stands. These high magnitude wet season floods are, however, highly episodic and in the intervening years smaller floods allow the development of existing trees and the recruitment of younger age classes. This regime of high frequency large disturbances restricts the establishment of stable states and the ecosystem is characterised by long periods of transition between short-lived stable states. This riparian ecosystem is thus driven by abiotic or allogenic processes rather than by biotic (autogenic) processes. In contrast, lower energy, seasonal flooding on the Blackwood River allows mature 249 stands of trees to develop throughout the river profile (Figure 10.2). Recruitment is continual, although species can also respond to large flood events. This disturbance regime results in long periods of stable states with short periods of transition. The vegetation is subjected to longer periods of autogenic processes and, because of lower frequency flooding disturbance, shorter periods of allogenic processes. a) ORD b) BLACKWOOD RIVER RIVER .._ Perslstance Recovery vegetation develops i vegetation develops .. ',,,',,, to maturity from recruitment of seedlings or Response vegetatively . Responsel vegetation develops systems reaction to systems reaction to from recruitment of the disturbance the disturbance seedlings or vegetatively . Figure 10.2: Pathways of vegetation development and recovery in the riparian zone in response to a) high frequency fluvial disturbance as experienced on the Ord River and b) infrequent fluvial disturbance as experienced on the Blackwood River. (adapted from Wissmar & Swanson 1990). Such a non-equilibrium model may provide an appropriate framework for management of degraded riparian zones and has been applied to riparian vegetation dynamics on a montane river in the U.S.A. (Baker & Walford 1995). Large scale disturbances, such as high water flows and flooding, are likely to be important factors in the dynamics of the riparian community. Recruitment, establishment and survival of riparian trees is thought to be closely tied to the hydrological regime (Barnes 1985; Bradley & Smith 1986; Wissmar & Swanson 1990; Johnson 250 1994; Naiman et al. 1998). Flood-induced disturbance may instigate the transition between stable states of plant composition and population structure. It can be hypothesised therefore, that river systems with frequent high energy flood eve1;1ts will support riparian communities with long periods of transition and rather shorter stable states. There may of course be islands of stability, in a general mosaic of differing states of transition, depending on where flooding has caused most damage. This may be along a gradient within the riparian zone, from low elevation high disturbance areas to areas at higher elevations with low frequency disturbance. A description of the population structure of particular species within the riparian zone can be used to assess population stability and successional trends or vegetation dynamics (Veblen 1986). In many instances, large numbers of seedlings germinate each year but there is a high mortality rate (100% in some years). Therefore, while germination is more common at lower elevations where moist open sediments provide ideal sites for seedlings, growth to maturity is more likely at higher elevations which are less likely to be flooded Gones et al. 1994). The distribution of plant communities in the riparian zone can be related to the gradient away from the river, with water being the most important factor. This includes the frequency of flooding, length of time that areas are inundated and groundwater and soil water replenishment. Lateral patterns of the riparian vegetation on low terraces and floodplains are mostly influenced by abiotic factors such as moisture gradients and flooding. On higher terraces and slopes the vegetation is more influenced by biotic factors such as competition and herbivory, although water stress may also become important as conditions become more xeric. Population structures on rivers with a high frequency of disturbance, such as occurs on the Ord River, reflect this fluvial regime. Results highlight the very different fluvial regimes that exist on the Ord and Blackwood Rivers and their effect on vegetation dynamics for the overstorey species. This is despite the species compared on the two rivers being functionally very similar. Management of the riparian vegetation should therefore take into account the frequency of change in the vegetation and that the disturbed states and long periods of transition between states are part of the natural process. This would suggest that altering the natural flow regimes, such as through river regulation, will have different effects on riparian vegetation dynamics on different rivers. 251 Management implications This work has relevance to several aspects of riparian zpne vegetation, including management of natural systems unaffected by man-made disturbances, for systems affected by livestock grazing and erosion and for areas requiring rehabilitation on regulated rivers. It highlights the importance of fluvial processes to riparian vegetation and indicates that understanding the natural flow regime of a target river is a critical first step in the management of riparian vegetation and planning of riparian zone rehabilitation. Where the riparian zone is highly modified, for example, by livestock grazing and/or weed invasion, natural regeneration of the riparian vegetation may be a long-term process. If intervention through direct seeding or replanting with seedlings is appropriate, care should be taken that species selected are adapted to particular site conditions such as flooding regime, landscape position and river geomorphology. Plants used for rehabilitation should also be adapted to the wet and dry cycle experienced in the riparian zone (Pautou & Decamps 1985) and have a reproductive cycle adapted to flowing water. Ideally, plants that are endemic to the riparian zone of the river that is to be rehabilitated should be used. There are many prescriptions for rehabilitation of riparian zones. These generally include a precise statement of objectives, good study design, collection of baseline data, monitoring before and after implementing the project, flexibility to change the procedure based on results of on-going monitoring and wide reporting of results (both successes and failures) (Henry & Amoros 1995; Kondolf 1995; O'Donnell 1995; Hancock et al. 1996). Natural regeneration of riparian zones may not always be successful because of changes to the environment. Invasion of early successional species, such as exotic annual grasses, may prevent establishment of seedlings of native perennial species through competition for light and moisture. Erosion of top soil and trampling of the soil by livestock may make conditions inhospitable for germination and establishment. There may also be a lack of propagules in the surviving vegetation. In these circumstances, natural regeneration may need to be enhanced by such methods as herbicide spraying to eliminate weeds, spreading seeds of local species and 252 replanting areas with seedlings of local species. If replanting is to be done, care should be taken with species selection to match the various types of plant communities that occur in the riparian zone. For example, different plant communities would occur 'adjacent to pools or areas of rapids or riffle zones, on the levee banks, in abandoned channels and on the floodplain. The high natural disturbance levels occurring in riparian zones may mean that there are particular requirements and adaptations of species necessary to achieve successful establishment. For example, protection of seedlings from strong flows and woody debris in floodwaters may be required for several years until seedlings are strong enough to withstand this disturbance (Hancock et al. 1996}. Different sections of a river, such as reaches and pools, riffle zones, erosion bends and depositional areas, flood plains and narrow, steep-sided sections may have different requirements for regeneration. This highlights the importance of individual responses of species to the environment and to chance events following disturbance (Gleason 1927). This aspect requires further research. The impact on riparian vegetation of river regulation has been widely studied (e.g. Nilsson et al. 1991a; Walker et al. 1994; Barnes 1997; Johnson 1997). The majority of the worlds rivers are regulated (Petts 1984) and in the northern third of the world 77% of total discharge of the 139 largest rivers are affected by some form of regulation (Dynesius & Nilsson 1994). Understanding the vegetation dynamics and the mechanisms of recruitment under natural (unregulated) flow conditions is critical to understanding or anticipating the effects of river flow regulation en riparian vegetation (Nilsson et al. 1997). Water resource policy in Australia requires setting a provision of water for the environment in regulated systems (Arthington & Zalucki 1998). Setting of environmental water requirements by documenting the variability in the natural flow regime of a river and matching this to ecologically critical flow requirements of in-stream biota has been used to define a schedule of flow releases for the North Dandalup and Harvey Rivers in south-western Australia (Davies et al. 1998). Based en the results from this study, this holistic approach (Arthington 1998} can be extended to include provisions for the riparian vegetation using quantitative vegetation data as shown in the example in Chapter 5. This would refine the setting of environmental water requirements on regulated rivers. 253 The common thread that appears to influence all aspects of the vegetation replacement processes examined in this thesis is the fluvial regime. The strof.lg relationship between fluvial regime and vegetation dynamics is important in developing the unique character of riparian vegetation (Wissmar & Swanson 1990; Naiman et al. 1998). River flows, and particularly the disturbance of large floods, affect vegetation community structure and spatial distribution, seed dispersal, seedling recruitment and population structure. The species which were examined in detail provide good examples of adaptations to conditions in the riparian zone, including reproductive phenology synchronised to seasonal flows, floating ability of seed to facilitate dispersal and tolerance of seedlings to flooding. Other aspects are also important in determining riparian vegetation community structure, including biotic (e.g. competition, livestock grazing) and abiotic (e.g. soils, topography and fire) factors, with the influence of each depending to a large extent m the interaction with the hydrology. This influence is obviously different for the two study rivers, with the riparian vegetation m the Ord River strongly influenced by the fluvial regime whereas on the Blackwood River the effects are more subtle and other factors have a stronger influence on vegetation dynamics. In terms of river restoration, this emphasises the importance of gaining a sound understanding of the fluvial regime for a particular river, along with other abiotic factors, such as soils and topography, and relating this to the biology of the vegetation community. 254 REFERENCES Abido M.S. (1985) The use of riparian vegetation in the Damascus basin. In 'Riparian Ecosystems and their Management: Reconciling Conflicting Uses.' Vol. 16 - 18 April 1985 pp. 467-469. North American Riparian Conference, Tucson, Arizona, USA. Abugov R. (1982) Species diversity and phasing of disturbance. Ecology 63: 289-293. Akilan K., Farrell R.C.C., Bell D.T. & Marshall J.K. (1997) Responses of clonal river red gum (Eucalyptus camaldulensis) to waterlogging by fresh and salt water. Australian Journal of Experimental Agriculture 37: 243-248. Allen R.B., Lee W.G. & Ranee B.D. (1994) Regeneration of indigenous forest after eradication of Norway rats, Breaksea Island, New Zealand. New Zealand Journal of Botany 32: 429-439. Andersen A.N. (1989) How important is seed predation to recruitment in stable populallons of long lived perennials? Oecologia 81: 310-315. Armour C., Duff D. & Elmore W. (1994) The effects of livestock grazing on western riparian and stream ecosystems. Fisheries 19: 9-12. Armstrong J.K. (1993). Restoration of function or diversity? In 'Nature Conservation 3: Reconstruction of Fragmented Ecosystems.' (Eds. D.A. Saunders, R.J. Hobbs & P.R. Ehrlich), Chp 20 pp. 209-214. Surrey Beatty and Sons, Sydney, Australia. Arthington A.H. (1998) Comparative evaluation of environmental flow assessment techniques: Review of holistic methodologies. Occasional Paper 26/98. Land and Water Resources Research and Development Corporation, Canberra, Australia. Arthington A.H., Bunn S.E. & Gray M. (1994) Stream hydrology and flow management in the Tully-Millstream hydroelectric scheme area. 80pp Report to Wet Tropics Management Agency, Queensland, Australia. Arthington A.H. & Zalucki J.M. (1998) Comparative evaluation of environmental flow assessment techniques: Review of Methods. Occasional Paper 27/98. Land and Water Resources Research and Development Corporation (LWRRDC), Canberra, Australia. Ashton D.H. (1975) Studies of flowering behaviour in Eucalyptus regnans F. Muell. Australian Journal of Botany 23: 399-411. 255 Ashton D.H. (1979) Seed harvesting by ants in forests of Eucalyptus regnans F Meull. in central Victoria. Australian Journal of Ecology 4: 265-277. Askey-Doran, M. and Pettit, N.E. (1999) Some impacts of human activity on riparian lands. In 'Riparian Land Management Technical Guidelines, Volume One: Principles of Sound Management' (Eds S. Lovett and P. Price). Chp 10 pp. 137-145. Land and Water Resources Research and Development Corporation, Canberra Australia. Bacon P.E., Stone C., Binns D.L., Leslie D.J. & Edwards D.W. (1993) Relationships between water availability and Eucalyptus camaldulensis growth in a riparian forest. Journal of Hydrology 150: 541-561. Baird A.M. (1977) Regeneration after fire in Kings Park, Perth, Western Australia. Journal of the Royal Society of Western Australia 60: 1-22. Baker W.L. (1988) Size class structure of contiguous riparian woodlands along a Rocky Mountain river. Physical Geography 9: 1-14. Baker W.L. (1990) Climatic and hydrologic effects on the regeneration of Populus angustifolia James along the Animas River, Colorado. Journal of Biogeography 17: 59-73. Baker W.L. & Walford G.M. (1995) Multiple stable states and models of riparian vegetation sucession on the Animas River, Colorado. Annals of the Association of American Geographers 85: 320-338. Barlow B.A. (1994). Phytogeography of the Australian region. In 'Australian Vegetation.' (Ed. R.H. Groves), pp. 3-35. Cambridge University Press, Cambridge. Bames W.J. (1985) Population dynamics of woody plants on a river island. Canadian Journal of Botany 63: 647-655. Bames W.J. (1997) Vegetation dynamics on the floodplain of the lower Chippewa River in Wisconsin. Journal of the Torrey Botanical Society 124: 189-197. Barrat-Segretain M.H. (1996) Strategies of reproduction, dispersion and competition in river plants: A review. Vegetatio 123: 13-37. Battaglia M. & Reid J.B. (1993) The effect of microsite variation on seed germination and seedling survival of Eucalyptus delegatensis. Australian Journal of Botany 41: 169-181. 256 Bayly I.A.E. & Williams W.D. (1981) Inland waters and their ecology. Longman Cheshire, Melbourne, Australia. Beard J. (1979). The vegetation of the Kimberley area, Western Australia: 1: 1,000,000 series. In Perth, W.A.: University of Western Australia Press, Perth, Australia. Beard J. (1981). The vegetation of the Swan area, Western Australia: 1: 1,000,000 series: sheet 7. In Perth, W.A.: University of Western Australia Press, Perth, Australia. Beardsell D.V., OBrien S.P., Williams E.G., Knox R.B. & Calder D.M. {1993) Reproductive biology of Australian Myrtaceae. Australian Journal of Botany 41: 511-526. Bell D.T. (1994) Interaction of fire, temperature and light in the germination response of 16 species from the Eucalyptus marginata forest of south-western Western Australia. Australian Journal of Botany 42: 501-509. Bell D.T. & Bellairs S.M. {1992) Effects of temperature on the germination of selected Australian native species used in rehabilitation of bauxite mining disturbance in Western Australia. Seed Science and Technology 20: 47-55. Bell D.T. & del Moral R. (1977) Vegetation gradients in the streamside forest of Hickory Creek; Will County, Illinois. Bulletin of the Torrey Botanical Club 104: 127-135. Bell D.T. & Heddle E.M. (1989). Floristic, morphologic and vegetational diversity In 'The Jarrah Forest.' (Eds. B. Dell, J.J. Havel & N. Malajczuk), Chp 5 pp. 53-66. Kluwer Academic Publishers, Dordrecht, Netherlands. Bell D.T., Hopkins A.J.M. & Pate J.S. (1982). Fire in the Kwongan In 'Kwongan: Plant life of the sandplain.' (Eds. J.S. Pate & J.S. Beard), Chp 9 pp. 178-204. University of Western Australia Press, Perth, Australia. Bell D.T. & Koch J.M. (1980) Post-fire succession in the jarrah forest of Western Australia. Australian Journal of Ecology 5: 9-14. Bell D.T., Plummer J.A. & Taylor S.K. (1993) Seed germination ecology in south western Western Australia. The Botanical Review 59: 24-73. Bell D.T., Vlahos S. & Bellairs S.M. (1990) Seed ecology in relation to reclamation: lessons from mined lands in Western Australia. Proceedings of the Ecological Society of Australia 16: 531- 535. 257 Bell D.T. & Williams J.E. (1997). Eucalypt ecophysiology In 'Eucalyptus Ecology.' (Eds. J.E. Williams & J.C.Z. Woinarski), pp. 168-196. Cambridge University Press, Cambridge UK. Bellairs S.M. & Bell D.T. (1990) Canopy borne seed store in three Western Australian plant communities. Australian Journal of Ecology 15: 299-305. Berg R.Y. (1975) Myrmecochorous plants in Australia and their dispersal by ants. Australian Journal of Botany 23: 475-508. Bettenay E., Russell W.G.R., Hudson D.R., Gilkes R.J. & Edmiston R.J. (1980) A description of the experimental catchments in the Collie area, Western Australia. Land Resources Management Technical Paper 7. Commonwealth Scientific and Industrial Research Organization, Australia. Bidwell R.G.S. (1974) Plant Physiology. · Macmillan Publishing Co., Inc., New York, USA. Blom C.W.P.M., Bogemann G.M., Laan P., Vandersman A.J.M., Vandesteeg H.M. & Voesenek L.A.C.J. (1990) Adaptations to flooding in plants from river areas. Aquatic Botany 38: 29-47. Blom C.W.P.M. & Voesnek L.A.C.J. (1996) Flooding: the survival strategies of plants. Trends in Ecology and Evolution 11: 290-295. Boland D.J., Brooker M.I.H., Chippendale G.M., Hall N., Hyland B.P.M., Johnston R.D., Kleinig D.A. & Turner J.D. (1989) Forest Trees of Australia. 4th Ed. Thomas Nelson Australia & CSIRO, Melbourne, Australia. Boland D.J., Brooker M.LH. & Turnball J.W. (1980) Eucalyptus Seed. CSIRO Australia, Melbourne, Australia. Boutin C. & Keddy P.A. (1993) A functional classification of wetland plants. Journal of Vegetation Science 4: 591-600. Bowen B.J. & Pate J.S. (1993) The significance of root starch in post fire shoot recovery of the resprouter Stirlingia latifolia R. Br. (Proteaceae). Annals of Botany 72: 7-16. Bowman D.M.J.S., Wilson B.A. & Hooper R.J. (1988) Response of Eucalyptus forest and woodland to four fire regimes at Munmarlary, Northern Territory, Australia. Journal of Ecology 76: 215-232. Bowman D.M.J.S., Wilson B.A. & Woinarski J.C.Z. (1991) Floristic and phenological variation in a northern Australian rocky savanna. Proceedings of the Royal Society of Queensland 101: 79- 90. 258 Bradley C.E. & Smith D.G. (1986) Plains cottonwood recruitment and survival on a prairie meandering river floodplain, Milk River, southern Alberta and northern Montana. Canadian Journal of Botany 64: 1433-1442. Bren L.J. (1992) Tree invasion of an intermittent wetland in relation to changes in flooding frequency of the River Murray, Australia. Australian Journal of Ecology 17: 395-408. Bren L.J. & Gibbs N.L. (1986) Relationships between flood frequency, vegetation and topography in a river red gum forest. Australian Forestry Research 16: 357-370. Bretschko G. (1995) River/land ecotones: Scales and patterns. Hydrobiologia 303: 83-91. Britton D:L. & Brock M.A. (1994) Seasonal germination from wetland seed banks. Australian Journal of Marine and Freshwater Research 45: 1445-1457. Brock M.A., Theodore K. & O'Donnell L. (1994) Seed bank methods for Australian wetlands. Australian Journal of Marine and Freshwater Research 45: 483-493. Brooker M.I.H. & Kleinig D.A. (1994) Field Guide to the Eucalypts: Northern Australia. Inkata Press, Sydney, Australia. Brown N.A.C. (1993) Promotion of germination of fynbos seeds by plant-derived smoke. New Phytologist 123: 575-583. BrownS. & Lugo A.E. (1994) Rehabilitation of tropical lands: a key to sustaining development. Restoration Ecology 2: 97-111. Brown V.K. (1992) Plant succession and life history strategy. Trends in Ecology and Evolution 7: 143-144. Bryant W.G. (1971) Grazing, burning and regeneration of tree seedlings of Eucalyptus pauciflora woodlands. Journal of Soil Conservation Service of New South Wales. 29: 29-44. Bunn S.E. (1988) Processing of leaf litter in a northern jarrah forest stream, Western Australia: I. Seasonal differences. Hydrobiologia 162: 201-210. Bunn S.E. (1993) Riparian-stream linkages: research needs for the protection of in-stream values. Australian Biology 6: 46-51. Bunn S.E., Davies P.M., Negus, P. and Treadwell, S. (1999) Aquatic food webs. In 'Riparian Land Management Technical Guidelines, Volume One: Principles of Sound Management.' (Eds. S. 259 Lovett and P. Price). Chp 4 pp. 25-36. Land and Water Resources Research and Development Corporation, Canberra, Australia. Burrows D.M. & Burrows W.H. (1992) Seed productioJ.! and litter fall in some eucalypt communities in central Queensland. Australian Journal of Botany 40: 389-403. Busch D.E. (1995) Effects of fire on southwestern riparian plant community structure. The Southwestern Naturalist 3: 259-267. Busch D.E. & Smith S.D. (1993) Effects of fire on water and salinity relations of riparian woody taxa. Oecologia 94: 186-194. Cale P. & Hobbs R.J. (1991). Condition of roadside vegetation in relation to nutrient status. In: 'Nature Conservation 2: The Role of Corridors.' (Eds. D.A. Saunders & R.J. Hobbs), Chp 32 pp. 353-362. Surrey Beatty and Sons, Sydney, Australia. Campbell I.C. (1993) Riparian stream linkages: an Australian perspective on instream issues. In 'Ecology and management of riparian zones in Australia.' pp. 21-30. Land and Water Resources Research and Development Corporation, Canberra, Australia. Canham C.D. & Marks P.L. (1985). The response of woody plants to disturbance: Patterns of establishment and growth. In: 'The ecology of natural disturbance and patch dynamics.' (Eds. P.S. White & S.T.A. Pickett), pp 198-216. Academic Press, San Diego, USA. Carpenter F.L. & Recher H.F. (1979) Pollination, reproduction and fire. American Naturalist 113: 871-879. Carr S.G.M. & Turner J.S. (1959) The ecology of the Bogong High Plains. I. The environmental factors and the grassland communities. Australian Journal of Botany 7: 12-33. CEP A (1992) Towards healthier rivers. December 1992. Commonwealth Environmental Protection Authority, Canberra, Australia. Chauvet E. & Decamps H. (1989) Lateral interactions in a fluvial landscape: The river Garonne, France. Journal of North American Benthological Society 8: 9-17. Cheal D.C. (1993) Effects of stock grazing on the plants of semi-arid woodlands and grasslands. Proceedings of the Royal Society of Victoria 105: 57-65. Chesterfield E.A. (1986) Changes in the vegetation of the river red gum forest at Barmah, Victoria. Australian Forestry 49: 4-15. 260 Clements F.E. (1916) Plant Succession. An Analysis of the Development of Vegetation. Carnegie Institute, Publication 242, Washington DC, USA. Colwell R.K. (1974) Predictability, constancy and contingency of periodic phenomena. Ecology 55: 1148-1153. Connell J.H. & Slatyer R.O. (1977) Mechanisms of succession in natural communities and their role in community stability and organization. The American Naturalist 111: 1119-1144. Cowling R.M., Mustart P.J., Laurie H. & Richards M.B. (1994) Species diversity; Functional diversity and functional redundancy in fynbos communities. South African Journal of Science 90: 333-337. Crawley M.J. (1992). Seed predators and plant population dynamics In 'Seeds: The ecology of regeneration in plant communities.' (Ed. M. Fermer), pp. 157-192. CAB International, W allingford, UK. Crome F., Isaacs J. & Moore L. (1994) The utility to birds and mammals of remnant riparian vegetation and associated windbreaks in tropical Queensland uplands. Pacific Conservation Biology 1: 328-343. Cunningham T.M. (1957) Seed production and seed fall of Eucalyptus regnans (F. Muell.). Australian Forestry 21: 30-39. Davidson N.J. & Reid J.B. (1989) Response of eucalypt species to drought. Australian Journal of Ecology 14: 139-156. Davies P.M. (1994). Ecosystem process: a direct assessment of river health In 'Classification of rivers.' (Ed. B. Hart), Chp 10 pp. 119-128. Davies P.M., Bunn S.E. & Arthington A.H. (1998) Environmental water requirements for lowland river systems on the swan coastal plain. (Draft) September 1998. Report for the Water and Rivers Commission of Western Australia. Department of Zoology, The University of Western Australia. Davies S.J. & Myerscough P.J. (1991) Post-fire demography of the wet mallee Eucalyptus luehmannia F. Muell. Australian Journal of Botany 39: 459-466. de Salis J. (1993) Ord River Regeneration Reserve. Miscellaneous Publication 14/93. December 1993. Western Australian Department of Agriculture, Perth, Australia. 261 De Steven D. (1991) Experiments on mechanisms of tree establishment in old-field succession: seedling emergence. Ecologtj 72: 1066-1075. Decamps H. (1993) River margins and environmental change. Ecological Applications 3: 441-445. Decamps H., Joachim J. & Lauga J. (1987) The importance for birds of the riparian woodlands within the alluvial corridor of the river Garonne, S.W. France. Regulated Rivers: Research and Management 1: 301-316. Decamps H., Planty-Tabacchi A.M. & Tabacchi E. (1995) Changes in the hydrological regime and invasions by plant species along riparian systems of the Adour river, France. Regulated Rivers: Research and Management 11: 23 - 33. Denton M. & Ganf G.G. (1994) Response of juvenile Melaleuca halmaturorum to flooding: Management implications for a seasonal wetland, Bool Lagoon, South Australia. Australian Journal of Marine and Freshwater Research 45: 1395-1408. Department of Mines. (1972) Geological Survey of Western Australia : Annual Report for the year 1971. Extract from the report of the Department of Mines, Perth, Australia. Dexter B.D. (1978) Silviculture of the river red gum forests of the central Murray flood plain. Royal Society of Victoria Proceedings 90: 175-191. Dexter B.D., Rose H.J. & Davies N. (1986) River regulation and associated forest management problems in the River Murray Red Gum forests. Australian Forestn; 49: 16-27. Dixon K.W., Roche S. & Pate J.S. (1995) The promotive effect of smoke derived from burnt native vegetation on seed germination of Western Australian plants. Oecologia 101: 185-192. Douglas M. M. (1999) Tropical savanna streams: Their seasonal dynamics and response to disturbance. Unpublished PhD Thesis, Department of Biological Sciences, Monash University, Melbourne, Australia. Drake W.E. (1981) Ant-Seed interaction in dry sclerophyll forest on North Stradbroke Island, Queensland. Australian Journal of Botany 29: 293-309. Drury W.H. & Nisbet I.C.T. (1973) Succession. Journal of Arnold Arboretum 54: 331-368. Duncan R.P. (1993) Flood disturbance and the coexistence of species in a lowland Podocarp forest, South Westland, New-Zealand. Journal of Ecology 81: 403-416. 262 Dynesius M. & Nilsson C. (1994) Fragmentation and flow regulation of river systems in the northern third of the world. Science 266: 753-762. Egler F.E. (1954) Vegetation science concepts. I. Initial floristic composition, a factor in old field development. Vegetatio 4: 412-417. Eldridge K., Davidson J., Harwood C. & van Wyk G. (1993) Eucalypt Domestication and Breeding. Oxford University Press, Oxford, UK. Ellison A.M., Denslow J.S., Loiselle B.A. & Brenes M.D. (1993) Seed and seedling ecology of neotropical Melastomataceae. Ecology 74: 1733-1749. Enright N.J. & Lamont B.B. (1989) Seed banks, fire season, safe sites and seedling recruitment in five eo-occurring Banksia species. Journal of Ecologtj 77: 1111-1122. Ernst W.H.O. (1990) Ecophysiology of plants in waterlogged and flooded environments. Aquatic Botany 38: 73-90. Estabrook G.F., Winsor J.A., Stephenson A.G. & Howe H.F. (1982) When are two phenological patterns different? Botanical Gazette 143: 374-378. Fensham R.J. (1992) The establishment of eucalyptus seedlings in tropical savanna forest. Northern Territory Naturalist 13: 30-36. Fischer S.G. (1990) Recovery processes in lotic systems: Limits of successional theory. Environmental Management 14: 725-736. Fitzgerald K. (1967) The Ord River catchment regeneration project. Bulletin 3599. West Australian Department of Agriculture, Perth, Australia. Fleischner T.L. (1994) Ecological costs of livestock grazing in western North America. Conservation Biology 8: 629-644. Fonda R.W. (1974) Forest succession in relation to river terrace development in Olympic National Park. Ecology 55: 927-942. Forman R.T.T. & Godron M. (1981) Patches and structural components for a landscape ecology. Bioscience 31: 733-739. Fowler N.L. (1988) What is a safe site?: Neighbor, litter, germination date and patch effects. Ecology 69: 947-961. Fox J.F. (1979) Intermediate disturbance hypothesis. Nature 204: 1344-1345. 263 Friedel M.H. (1991) Range condition assessment and the concept of thresholds: A viewpoint. Journal of Range Management 44: 422-426. Friedel M.H., Pickup G. & Nelson D.J. (1993) The interpretation of vegetation change in a spatially and temporally diverse arid Australian landscape. Journal of Arid Environments 24: 241-260. Friedman J.M., Scott M.L. & Lewis W.M. (1995) Restoration of riparian forest using irrigation, artificial disturbance, and natural seedfall. Environmental Management 19: 547-557. Froend R.H., Farrell R.C.C., Wilkins C.F. & McComb A.J. (1993). The effect of altered water regimes on wetland plants In 'Wetlands of the Swan Coastal Plain.' (Ed. S.A. Balla), 144 pp. Water Authority of Western Australia, Perth Western Australia. Froend R.H. & McComb A.J. (1994) Distribution, productivity and reproductive phenology of emergent macrophytes in relation to water regimes at wetlands of south-western Australia. Australian Journal of Marine and Freshwater Research 45: 1491-1508. Gentilli J. (1972) Australian Climate Patterns. Thomas Nelson Ltd, Melbourne, Australia. George M.R., Brown J.R. & Clawson W.J. (1992) Application of non-equilibrium ecology to management of Mediterranean grassland. Journal of Range Management 45: 436-440. Gibson A. & Bachelard E.P. (1987) Provenance variation in germination response to water stress of seeds of some eucalypt species. Australian Forestry Research 17: 49-58. Gibson N. & Kirkpatrick J.B. (1989) Effects of cessation of grazing on grasslands and grassy woodlands of the central plateau, Tasmania. Australian Journal of Botany 37: 55-63. Gill A.M. (1981). Coping with fire. In: 'The Biology of Australian Plants.' (Eds. J.S. Pate & A.J. McComb), Chp 3 pp. 65-87. University of Western Australia Press, Perth, Australia. Gilliam J.W. (1994) Riparian wetlands and water quality. Journal of Environmental Quality 23: 896-900. Gleason H.A. (1927) Further views of the succession concept. Ecology 8: 299-326. Glenn-Lewin D.C. & van der Maarel E. (1992). Patterns and processes of vegetation dynamics. In: 'Plant Succession: Theory and Prediction.' (Eds. D.C. Glenn-Lewin, R.K. Peet & T.T. Veblen), Chp 1 pp. 11-58. Cambridge University Press, Cambridge UK. 264 Gordon N.D., McMahon T.A. & Finlayson B.L. (1992) Stream hydrology: An Introduction for Ecologists. John Wiley and Sons, Chichester, England. Green J.W. (1985) Census of Vascular Plants of Western Australia. 2nd Ed. Western Australian Herbarium, Department of Agriculture., Perth, Australia. Greenway M. (1994) Litter accession and accumulation in a Melaleuca quinquernervia wetland in south-eastern Queensland. Australian Journal of Marine and Freshwater Research 45: 1509- 1519. Gregmy S.V., Swanson F.J., McKee W.A. & Cummins K.W. (1991) An ecosystem perspective of riparian zones. Bioscience 41: 540-551. Grein S.B. (1995) Remnant vegetation and natural resources of the Blackwood River catchment: An atlas. Miscellaneous Publication 9/95'. November 1995. Agriculture Western Australia, Perth, Australia. Grime J.P. (1974) Vegetation classification by reference to strategies. Nature 250: 26-31. Grime J.P. (1979) Plant Strategies and Vegetation Processes. John Wiley & Sons, New York, USA Grime J.P. (1980). An ecological approach to management. In: 'Amenity Grassland: An Ecological Perspective.' (Eds. I.H. Rorison & R. Hunt) pp. 13-35. John Wiley and Sons, New York, USA. Grime J.P. & Hillier S.H. (1992). The contribution of seedling regeneration to the structure and dynamics of plant communities and larger units of landscape. In 'Seeds: The Ecology of Regeneration in Plant Communities.' (Ed. M. Fermer) pp. 349-364. CAB International, W allingford, UK. Grime J.P., Hodgson J.G. & Hunt R. (1986) Comparative Plant Ecology : A Functional Approach to Common British Species. Unwin Hyman, London, UK. Gross K.L. (1990) A comparison of methods for estimating seed numbers in the soil. Journal of Ecology 78: 1079-1093. Groves R.H. (Ed.). (1994) Australian Vegetation (2nd ed.). Cambridge University Press, Cambridge, UK. Grubb P.J. (1977) The maintenance of species-richness in plant communities: The importance of the regeneration niche. Biological Review 52: 107-145. 265 Hacker R.B. (1987) Species responses to grazing and environmental factors in an arid halophytic shrubland community. Australia Journal of Botany 35: 135-150. Hancock C.N., Ladd P.G. & Froend R.H. (1996) Biodiversity and management of riparian vegetation in Western Australia. Forest Ecology and Management 85: 239-250. Hansen A., Pate J.S. & Hansen A.P. (1991) Growth and reproductive performance of a seeder and a resprouter species of Bossiaea as a function of plant age after fire. Australian Journal of Botany 67: 497-509. Harper J.L. (1977) Population Biology of Plants. Academic Press, London, UK. Hawke G.M. & Zobel D. B. (1974) Forest succession on alluvial landforms of the McKenzie River valley, Oregon. Northwest Science 48: 245-265. Henry C.P. & Amoros C. (1995) Restoration ecology of riverine wetlands: I. A scientific base. Environmental Management 19: 891-902. Herrera J. (1991) The reproductive biology of a riparian Mediterranean shrub, Nerium oleander L. (Apocynaceae). Botanical Journal of the Linnean Society 106: 147-172. Hill M.O. & Gauch H.G. (1980) Detrended correspondence analysis, an improved ordination technique. Vegetatio 42: 47-58. Hills J.M., Murphy K.J., Pulford I.D. & Flowers T.H. (1994) A method of classifying European riverine wetland ecosystems using functional vegetation groups. Functional Ecologtj 8: 242-252. Hobbs R.J. (1983) Markov models in the study of post-fire succession in heathland communities. Vegetatio 56: 17-30. Hobbs R.J. (1988) Effects of annual plants on the establishmment of shrub seedlings on a nature reserve in the West Australian wheatbelt. Annual Revegetation Newsletter 1: 33-34. Hobbs R.J. (1992). Is biodiversity important for ecosystem functioning? Implications for research and management. In 'Biodiversity of Mediterranean Ecosystems in Australia.' (Ed. R.J. Hobbs), Chp 11 pp. 211-229. Surrey Beatty and Sons., Sydney, Australia. Hobbs R.J. (1994) Dynamics of vegetation mosaics: Can we predict responses to global change? Ecoscience 1: 346-356. Hobbs R.J. & Huenneke L.F. (1992) Disturbance, diversity, and invasion: Implications for Conservation. Conservation Biology 6: 324-336. 266 Hobbs R.J. & Norton D.A. (1996) Towards a conceptual framework for restoration ecology. Restoration Ecology 4: 93-110. Hodgkinson K.C. & Harrington G.N. (1985} The case for prescribed burning to control shrubs in eastern semi-arid woodlands. Australian Rangeland Journal7: 64-74. House S.M. (1997). Reproductive biology of eucalypts. In: 'Eucalyptus Ecology.' (Eds. J.E. Williams & J.C.Z. Woinarski}, pp. 30-55. Cambridge University Press, Cambridge, UK. Huston M. (1979) A general hypothesis of species diversity. Americam Naturalist 113: 81-101. Hutchings M.J. (1997). The structure of plant populations. In: 'Plant Ecology.' (Ed. M.J. Crawley), Ch. 6 pp 325-358. Blackwell Science Limited, Oxford, UK. James S. (1984) Lignotubers and hurls - their structure, function and ecological significance in Mediterranean ecosystems. Botanical Review 50: 225-266. Janzen D.H. (1971) Seed predation by animals. Annual Review of Ecology and Systematics 2: 465- 492. Jean M. & Bouchard A. (1993) Riverine wetland vegetation - importance of small-scale and large scale environmental variation. Journal of Vegetation Science 4: 609-620. Jeffries M. & Mills D. (1990} Freshwater Ecology: principles and applications. Bellhaven Press, London, UK. Johansson M. E., Nilsson C. & Nilsson E. (1996) Do rivers function as corridors for plant dispersal? Journal of Vegetation Science 7: 593-598. Johnson W.C. (1994} Woodland expansion in the Platte River, Nebraska: Patterns and causes. Ecological Monographs 64: 45-84. Johnson W.C. (1997} Equilibrium response of riparian vegetation to flow regulation in the Platte River, Nebraska. Regulated Rivers: Research and Management 13: 403-415. Jones R.H. & Sharitz R.R. (1989} Potential advantages and disadvantages of germinating early for trees in floodplain forests. Oecologia 81: 443-449. Jones R.H., Sharitz R.R., Dixon P.M., Segal D.S. & Scheinder R.L. (1994} Woody plant regeneration in four floodplain forests. Ecological Monographs 64: 345-367. 267 Junk W.J., Bayley P.B. & Sparks R.E. (1989) The flood pulse concept in riverplain systems. In 'International Large River Symposium.' Vol. 106 (Eds D.P. Dodge.) pp. 110-127. Fisheries and Aquatic Sciences Special Publication, Canada. Kallida R. & Puhakka M. (1988) River dynamics and vegetation mosaicism: A case study of the river Kamagohka, northernmost Finland. Journal of Biogeography 15: 703-719. Kauffman J.B., Krueger W.C. & Vavra M. (1983) Effects of late season cattle grazing on riparian plant communities. Journal of Range Management 36: 685-691. Keeley J.E. (1986). Resilience of mediterranean shrub communities to fires. In: 'Resilience in Mediterranean-type Ecosystems.' (Eds. B. Dell, A.J.M. Hopkins & B.B. Lamont), Ch. 7 pp. 95- 112. Dr J. W. Junk, Dordrecht, Netherlands. Kellman M. & Tackaberry R. (1993) Disturbance and tree species coexistence in tropical riparian forest fragments. Global Ecology and Biogeography Letters 3: 1-9. Kent M. & Coker P. (1992) Vegetation Description and Analysis: A Practical Approach. Belhaven Press, London, UK. Klute A.E. (1986) Methods of Soil Analysis: Part 1 - Physical and Mineralogical Methods. 2 Edn. American Society of Agronomy Inc. & Soil Science Society of America Inc., Madison, Wisconsin, USA. Knopf F.L. & Samson F.B. (1994) Scale perspectives on avian diversity in western riparian ecosystems. Conservation BiolOgJJ 8: 699-676. Kondolf G.M. (1995) Five elements for effective evaluation of stream restoration. Restoration Ecology 3: 133-136. Kozlowski T.T. (1984) Plant responses to flooding of soil. Bioscience 34: 162-167. Kramer P.J. & Kozlowski T.T. (1979) Physiology of Woody Plants. Academic Press, New York, USA. Lamont B.B., Lemaitre D.C., Cowling R.M. & Enright N.J. (1991) Canopy seed storage in woody plants. Botanical Review 57: 277-317. Lavorel S., Lepart J., Debussche M., Lebreton J. & Beffy J. (1994) Small scale disturbances and the maintenance of species diversity in Mediterranean old fields. Oikos 70: 455-473. 268 Laycock W.A. (1991) Stable states and thresholds of range condition on North American rangelands: A veiwpoint. Journal of Range Management 44: 427-433. Leach R.H. & Edge W.D. (1994) Summer home range and habitat selection by white tailed deer in the Swan valley, Montana. Northwest Scientist 68: 31-36. Leigh J.H., Wood D.H., Slee A.V. & Holgate M.D. (1991) The effect of burning and simulated grazing on productivity, forage quality, mortality and flowering of eight subalpine herbs in Kosciusko National Park. Australian Journal of Botany 39: 97-118. Levin S.A. (1976) Population dynamic models in heterogeneous environments. Annual Review of Ecology and Systematics 7: 287-310. Lowrance R., Todd R., Fail J., Hendrickson 0., Leonard R. & Asmussen L. (1984) Riparian forests as nutrient filters in agricultural wat~rsheds. Bioscience 34: 374-377. Lovett S. & Price P. (Eds.) (1999) Riparian Land Management Technical Guidelines, Volume One: Principles of Sound Management. Land and Water Resources Research and Development Corporation, Canberra, Australia. Loyn R.H. (1987). Effects of patch area and habitat on bird abundances, species numbers and tree health in fragmented Victorian forests. In 'Nature Conservation: The Role of Remnants of Native Vegetation.' (Eds. D.A. Saunders, G.W. Arnold, A.A. Burbidge & A.J.M. Hopkins), Ch. 6 pp. 65-77. Surrey Beatty and Sons, Sydney, Australia. Luken J.O. (1990) Directing Ecological Succession. Chapman and Hall, New York, USA. Lush W.M., Kaye P.E. & Groves R.H. (1984) Germination of Clematis microphylla seeds following weathering and other treatments. Australian Journal of Botany 32: 121-129. Malanson G.P. (1993) Riparian Landscapes. Cambridge University Press, Cambridge, UK. Malanson G.P. & O'Leary J.F. (1982) Post-fire regeneration strategies of Californian coastal sage shrubs. Oecologia 53: 355-358. Malcolm C.V. (1983) Wheatbelt Salinity: A review of the salt land problem in south western Australia. Technical Bulletin 52. Western Australian Department of Agriculture, Perth, Australia. 269 Marchant N.G., Wheeler J.R., Rye B.L., Bennett E.M., Lander N.S. & Macfarlane T.D. (1987) Flora of the Perth Region Parts I and Il. Western Australian Department of Agriculture, Perth, Australia. McArthur W.M. (1991) Reference Soils of South-Western Australia. Western Australian Department of Agriculture, Perth, Australia. McDonald R.C. & Isbell R.F. (1984). Soil Profile. In: 'Australian Soil and Land Survey Field Handbook.' (Eds. R.C. Mcdonald, R.F. Isbell, J.G. Speight, J. Walker & M.S. Hopkins), pp. 83- 126. Inkata Press, Melbourne, Australia. Mclntyre S., Ladiges P.Y. & Adams G. (1988) Plant species-richness and invasion by exotics in relation to disturbance of wetland communities on the riverine plain, NSW. Australian Journal of Ecology 13: 361-373. Mclntyre S. & Lavorel S. (1994) How environmental and disturbance factors influence species composition in temperate Australian grasslands. Journal of Vegetation Science 5: 373-384. Mcintyre S., Lavorel S. & Tremont R.M. (1995) Plant life history attributes: Their relationship to disturbance response in herbaceous vegetation. Journal of Ecology 83: 31-44. McNaughton S.J. (1985) Ecology of a grazing ecosystem: The Serengeti. Ecological Monographs 55: 259-294. Melick D.R. (1990) Flood resistance of Tritaniopsis laurina and Acmena smithii from riparian warm temperate rainforest in Victoria. Australian Journal of Botany 38: 371-381. Milberg P. & Lamont B.B. (1995) Fire enhances weed invasion of roadside vegetation in southwestern Australia. Biological Conservation 73: 45-49. Milchunas D.G. & Lauenroth W.K. (1993) Quantitative effects of grazing on vegetation and soils over a global range of environments. Ecological Monographs 63: 327-366. Milchunas D.G. & Laurenroth W.K. (1995) Inertia in plant community structure: state changes after cessation of nutrient enrichment stress. Ecological Applications 5: 452-458. Miles J. (1979) Vegetation Dynamics. Chapman and Hall, London, UK. Montgomery D.R. & Buffington J.M. (1998). Channel processes, classification and response. In: 'River Ecology and Management: Lessons from the Pacific Coastal Ecoregion.' (Eds. R.J. Naiman & R.E. Bilby) pp. 13-42. Springer, New York, USA. 270 Morris J.D. (1984) Establishment of trees and shrubs on a saline site using drip irrigation. Australian Forestry 47: 210-17. Morrissy N. (1974) Reversed longihtdinal salinity profile of a major river in the southwest of Western Australia. Australian Journal of Marine and Freshwater Research 25: 327-335. Mott J.J. & Groves R.H. (1981). Germination strategies. In: 'The Biology of Australian Plants.' (Eds. J.S. Pate & A.J. McComb) pp. 307-341. University of Western Australia Press, Perth, Australia. Mucha S.B. (1979) Estimation of tree ages from growth rings of eucalypts in northern Australia. Australian Forestry 42: 13-16. Mueller-Dombois D. & Ellenberg H. (1974) Aims and Methods of Vegetation Ecology. John Wiley, New York, USA. Mulcahy M.J. & Bettenay E. (1972) Soil and landscape studies in Western Australia. (1) The major drainage divisions. Journal of the Geological Society of Australia 18: 349-357. Murdoch A.J. & Ellis R.H. (1992). Longevity, viability and dormancy. In: 'Seeds: The Ecology of Regeneration in Plant Communities.' (Ed. M. Fermer) Ch. 7 pp. 193-229. CAB International, W allingford, UK. Naiman R.J. & Bilby R.E. (Eds.). (1998) River Ecology and Management: Lessons from the Pacific Coastal Ecoregion. Springer, New York, USA. Naiman R.L Decamps H. & Pollock M. (1993) The role of riparian corridors in maintaining regional biodiversity. Ecolological Applications 3: 209-212. Naiman R.J., Fetherston K.L., McKay S.J. & Chen J. (1998). Riparian Forests. In: 'River Ecology and Management: Lessons from the Pacific Coastal Ecoregion.' (Eds. R.J. Naiman & R.E. Bilby) Ch. 12 pp. 289-323. Springer, New York, USA. Nanson G. C. & Beach H.F. (1977) Forest succession and sedimentation on a meandering flood plain, northeast British Columbia, Canada. Journal of Biogeography 4: 229-251. Newbury R.W. & Gaboury M.N. (1993) Stream Analysis and Fish Habitat Design. Newbury Hydraulics Ltd., Gibsons, British Columbia, Canada. Nilsson C., Ekblad A., Dynesius M., Backe S., Gardfjell M., Carlberg B., Hellqvist S. & Jansson R. (1994) A comparison of species richness and traits of riparian plants between a main river channel and its tributaries. Journal of Ecology 82: 281-295. 271 Nilsson C., Ekblad A., Gardfjell M. & Carlberg B. (1991a) Long term effects of river regulation en river margin vegetation. Journal of Applied Ecology 28: 963-987. Nilsson C., Gardfjell M. & Grelsson G. (1991b) Importance of hydrochory in structuring plant communities along rivers. Canadian Journal of Botany 69: 2631-2633. Nilsson C. & Grelsson G. (1990) The effects of litter displacement on riverbank vegetation. Canadian Journal of Botany 68: 735-741. Nilsson C. & Jansson R. (1995) Floristic differences between riparian corridors of regulated and free-flowing boreal rivers. Regulated Rivers: Research and Management 11: 55 - 66. Nilsson C., Jansson R. & Zinko U. (1997) Long-term responses of river-margin vegetation to water level regulation. Science 276: 798-800. Nilsson C., Nilsson E., Johansson M.E., Dynesius M., Grelsson G., Xiong S., Jansson R. & Danvind M. (1993) Processes structuring riparian vegetation. Current Topics in Botanical Research 1: 419- 431. Noble I.R. (1984) Mortality of lignotuberous seedlings of Eucalyptus species after an intense fire in a montane forest. Australian Journal of Ecology 9: 47-50. Noble I.R. & Slatyer R.O. (1980) The use of vital attributes to predict sucessional changes in plant communities. Vegetatio 43: 5-21. Northcote K.H., Bettenay E., Churchwood H.M. & McArthur W.N. (1967) Atlas of Australian Soils: Explanatory data for sheet 5 - Perth - Albany - Esperance area. Melbourne University Press, Melbourne, Australia. Noy-Meir I., Gutrnan M. & Kaplan Y. (1989) Responses of mediterranean grassland plants to grazing and protection. Journal of Ecology 77: 290-310. Noy-Meir I. & van der Maarel E. (1988) Relations between community theory and community analysis in vegetation science: some historical perspectives. Vegetatio 69: 5-15. O'Donnell S. (1995) Preliminary investigation into the management of riparian rainforests in south-east Queensland. 7 November 1995. Queensland Department of Primary Industries Brisbane, Australia. 272 O'Dowd D. & Gill A.M. (1984} Predator satiation and site alteration following fire: Mass reproduction of alpine ash (Eucalyptus delegatensis) in south eastern Australia. Ecology 65: 1052-1066. Ogden J. (1978) On the dendrochronological potential of Australian trees. Australian Journal of Ecology 3: 339-356. Olsen G. & Skitmore E. (1991} State of the rivers of the south west drainage division. 2/91. May 1991. Western Australian Water Resources Council, Perth, Australia. Onans J. & Parsons R.F. (1980) Regeneration of native plants on abandoned mallee farmland in south eastern Australia. Australian Journal of Botany 28: 479-493. Page A.L.E., Miller R.H. & Keeney D.R. (1986} Methods of Soil Analysis: Part 2 - Chemical and Microbiological Properties. 2 Edn. American Society of Agronomy Inc. & Soil Science Society of America Inc., Madison, Wisconsin, USA. Palmer M.W. (1993) Putting things in even better order: The advantages of Canonical Correspondence Analysis. Ecology 74: 2215-2230. Pate J.S. & Dixon K.W. (1981). Plants with fleshy tmderground storage organs - A Western Australian survey. In: 'The Biology of Australian Plants.' (Eds. J.S. Pate & A.J. McComb) Ch. 7 pp 181-215. University of Western Australia Press, Perth, Australia. Pate J.S., Froend R.H., Bowen B.J., Hansen A. & Kuo J. (1990) Seedling growth and storage characteristics of seeder and resprouter species of mediterranean-type ecosystems of S. W. Australia. Annals of Botany 65: 585-601. Pautou G. & Decamps H. (1985} Ecological interactions between the alluvial forests and hydrology of the upper Rhone. Arch. Hydrobiologia. 104: 13-37. Pearce-Pinto G.V.N., van der Moezel P.G. & Bell D.T. (1990) Seed germination under salinity stress in Western Australian species of Eucalyptus. Seed Science and Technology 18: 1-6. Peck A.J. (1978) Salinisation of non-irrigated soils and associated streams : A Review. Australian Journal of Soil Research 16: 157-168. Peet R.K., Knox, R.G., Case, J.S. and Alien, R.B. (1988) Putting things in order: The advantages of detrended correspondence analysis. The American Naturalist 131(6): 924-934. Pen L.J. (1999) Managing Our Rivers. Water and Rivers Commission, Perth, Australia.. 273 Pereira J.S. & Kozlowski T.T. (1977) Variations among woody angiosperms in response to flooding. Plant Physiology 41: 184-192. Perry R.A. (1970). Vegetation of the Ord-Victoria area.. In: 'Lands of the Ord-Victoria Area, Western Australia and the Northern Territory.' (Ed. G.A. Stewart), Land Research Series 28 pp. 104-119. Commonwealth Scientific and Industrial Research Organization, Canberra, Australia. Peterjohn W.T. & Correll D.L. (1984) Nutrient dynamics in an agricultural watershed: observations on the role of the riparian forest. EcolOglj 65: 1466-1475. Petheram R.J. & Kok B. (1991) Plants of the Kimberley Region of Western Australia. University of Western Australia Press, Perth, Australia. Pettit N.E. & Froend R.H. (1992) Research into reforestation techniques for saline groundwater control. Report No. WS97. November 1992. Water Authority of Western Australia, Perth, Australia. Pettit N.E., Froend R.H. & Ladd P.G. (1995) Grazing in remnant woodland vegetation: Changes in species composition and life form groups. Journal of Vegetation Science 6: 121-130. Pettit N.E., Ladd P.G. & Froend R.H. (1998) Passive clearing of native vegetation: Livestock damage to remnant jarrah (Eucalyptus marginata) woodlands in Western Australia. Journal of the Royal Society of Western Australia 81: 95-106. Petts G.E. (1983) Rivers. Butterworths, London, UK. Petts G.E. (1984) Impounded Rivers. John Wiley and Sons, London, UK. Pezeshki S.R. & Anderson P.H. (1997) Responses of three bottomland species with different flood tolerance capabilities to various flooding regimes. Wetland Ecology and Management 4: 245- 256. Pickett S.T.A., Kolasa J., Armesto J.J. & Collins S.L. (1989) The ecological concept of disturbance and its expression at various hierarchical levels. Oikos 54: 129-136. Pickett S.T.A. & Thompson J.N. (1978) Patch dynamics and the design of nature reserves. Biological Conservation 13: 27-37. Pidgeon R.W. & Cairns S.C. (1981) Decomposition and colonisation by invertebrates of native and exotic leaf material in a small stream in New England (Australia). Hydrobiologia 7.7: 113-127. 274 Piegay H. & Bravard J.P. (1997) Response of a Mediterranean riparian forest to a 1 in 400 year flood, Ouveze River, Drome-Vaucluse, France. Earth Surface Processes and Landforms 22: 31-43. Pinay G. & Decamps H. (1988) The role of woods in regulati.J::lg mtrogen fluxes between the alluvial aquifer and surface water: A conceptual model. Regulated rivers: Research and Management 2: 507-516. Pinay G., Roques L. & Fabre A. (1993) Spatial and temporal patterns of denitrification in riparian forests. Journal of Applied Ecology 30: 581-591. Pollock M.M., Naiman R.J. & Hanley T.A. (1998) Plant species richness in riparian wetlands -A test of biodiversity theory. Ecology 79: 94-105. Puckridge J.T. (1999). The role of hydrology in the biology of dryland rivers. In: 'A Free-flowing River: The Ecology of the Paroo River.' (Ed. RT. Kingsford), pp. 97-112. National Parks and Wildlife Service, NSW., Sydney, Australia. Puckridge J.T., Sheldon F., Walker K.F. & Boulton A.J. (1998) Flow variability and the ecology of large rivers. Marine and Freshwater Research 49: 55-72. Purdie R.W. & Slatyer R.O. (1976) Vegetation succession after fire in sclerophyll woodland communities in south eastern Australia. Australian Journal of Ecology 1: 223-236. Pusey B.J., Arthington A.H. & Read M.G. (1995) Species richness and spatial variation in fish assemblage structure in two rivers of the wet tropics of north Queensland. Environmental Biology of Fishes 42: 181-199. Raunkaier C. (1934) The Life Forms of Plants and Statistical Plant Geography. Clarendon Press, Oxford, UK. Richter B.D. (1997) How much water does a river need? Freshwater Biology 37: 231-249. Richter B.D., Baumgartner J.V., Powell J. & Braun D.P. (1996) A method for assessing hydrologic alteration within ecosystems. Conservation Biology 10: 1163-1174. Risser P.G. (1993) Ecotones at local to regional scales from around the world. Ecological Applications 3: 367-368. Roberts H.A. (1986) Seed persistence in the soil and seasonal emergence in plant species from different habitats. Journal of Applied Ecology 23: 639-656. 275 Roberts J. (1993) Regeneration and growth of Coolibah, Eucalyptus coolabah subsp. arida, a riparian tree, in the Cooper Creek region of South Australia. Australian Journal of Ecology 18: 345-350. Roberts J. & Ludwig J.A. (1991) Riparian vegetation along current-exposure gradients in floodplain wetlands of the river Murray, Australia. Journal of Ecology 79: 117-127. Rodolfo G. & Sala O.E. (1993) Plant functional types and ecological strategies in Patagonian forbs. Journal of Vegetation Science 4: 839-846. Rutherfurd, I., Abernethy, B. & Prosser, I. (1999) Stream erosion. In: 'Riparian Land Management Technical Guidelines, Volume One: Principles of Sound Management' (Eds. Lovett S. & Price P.) Ch. 6 pp. 61-78. Land and Water Resources Research and Development Corporation, Canberra, Australia. Sauer J.D. (1988) Plant Migration. University of California Press, Berkeley, USA. Saunders D.A., Hobbs R.J. & Margules C.R. (1991) Biological consequences of ecosystem fragmentation: A Review. Conservation Biology 5: 18-32. Schneider R.L. & Sharitz R.R. (1988) Hydrochory and regeneration in a Bald Cypress-Water Tupelo swamp forest. Ecology 69: 1055-1063. Schneider R.L. & Sharitz R.R. (1986) Seedbank dynamics in a south eastern riverine swamp. American Journal of Botany 73: 1022-1030. Schofield N.J., Ruprecht J.K. & Loh LC. (1988) The impact of agricultural development on the salinity of surface water resources of south-west Western Australia. Report No. WS 27. Water Authority of Western Australia, Perth, Australia. Scholander P.F., Hammel H.T., Hemmings E.A. & Bradstreet E.D. (1964) Hydrostatic pressure and osmotic potential in leaves of mangroves and some other plants. Proceedings of the National Academy of Science, USA 52: 119-125. Schweingruber F.H. (1992) Annual growth rings and growth zones in woody plants in southern Australia. IAWA Bulletin 13: 359-379. Scougall S.A., Majer J.D. & Hobbs R.J. (1993). Edge effects in grazed and ungrazed Western Australian wheatbelt remnants in relation to ecosystem reconstruction. In: 'Nature Conservation 276 3: Reconstruction of Fragmented Ecosystems.' (Eds. D.A. Saunders, R.J. Hobbs & P.R. Ehrlich), pp. 163-178. Surrey Beatty and Sons., Sydney, Australia. Senna-Gomes A.R. & Kozlowski T.T. (1980) Responses ~?f Melaleuca quinquenervia seedlings to flooding. Physiologia Plantarum 49: 373-377. Setterfield S.A. & Williams R.J. (1996) Patterns of flowering and seed production in Eucalyptus miniata and E. tetrodonta in a tropical savanna woodland, northern Australia. Australian Journal of Botany 44: 107-122. Shah B.H. & Thames J.L. (1985) Role of riparian vegetation in Pakistan. In: 'Riparian Ecosystems and their Management: Reconciling Conflicting Uses.' Vol. 16-18, April1985, pp 482-484. North American Riparian Conference, Tucson, Arizona, USA. Shea S.R., McCormick J. & Portlock C. C. (1979) The effects of fires on regeneration of leguminous species in the northern jarrah (Eucalyptus marginata Sm.) forest of Western Australia. Australian Journal of Ecology 4: 195-205. Shultz T.T. & Leininger W.C. (1990) Differences in riparian vegetation structure between grazed areas and exdosures. Journal of Range Management 43: 295-299. Skoglund J. & Verwijst T. (1989) Age structure of woody species populations in relation to seed rain, germination and establishment along the river Dalalven, Sweden. Vegetatio 82: 25-34. Skoglund S.J. (1990) Seed dispering agents in two regularly flooded rivers. Canadian Journal of Botany 68: 754-760. Slatyer R.O. (1970). Climate of the Ord-Victoria area. In: 'Lands of the Ord-Victoria Area, Western Australia and the Northern Territory.' (Ed. G.A. Stewart), Land Research Series 28 pp. 62-70. Commonwealth Scientific and Industrial Research Organisation, Canberra, Australia. Smith R.S. & Rushton S.P. (1994) The effects of grazing management on the vegetation of mesotrophic (meadow) grassland in Northern England. Journal of Applied Ecology 31: 13-24. Sorensen A.E. (1986) Seed dispersal by adhesion. Annual Review of Ecological Systematics 17: 443- 463. Sousa W.P. (1984) The responses of a community to disturbance: The importance of successional age and speices life histories. Oecologia 45: 72-81. 277 Spink A. & Rogers S. (1996) The effects of a record flood on the aquatic vegetation of the upper Mississipi River system: Some preliminary findings. Hydrobiologia 340: 51-57. Stewart G.A., Perry R.A., Slatyer R.O., Traves D.M.,. Dimn P.R., Jones P.J. & Paterson S.J. (1970) Lands of the Ord-Victoria Area, Western Australia and the Northern Territory. Report No. 28. Land Research Series Commonwealth Scientific and Industrial Research Organisation, Canberra, Australia. Stoneman G.L. & Dell B. (1994) Emergence of Eucaluptus marginata (jarrah) from seed in Mediterranean-climate forest in response to over:Storey, site, seedbed and ant harvesting. Australian Journal of Ecology 19: 96-102. Stoneman G.L., Dell B. & Turner N.C. (1994) Mortality of Eucalyptus marginata (jarrah) seedlings in Mediterranean-climate forest in response to overstorey, site, seedbed, fertilizer application and grazing. Australian Journal of Ecology 19: 103-109. Streng D.R., Glitzenstein J.F. & Harcombe P.A. (1989) Woody seedling dynamics in an East Texas floodplain forest. Ecological Monographs 59: 177-204. Stromberg J.C. (1993) Instream flow models for mixed deciduous riparian vegetation within a semiarid region. Regulated Rivers: Research and Management 8: 225-235. Szaro R.C. (1990) Southwestern riparian plant communities: Site characteristics, tree species distributions and size-class structures. Forest Ecology and Management 33/34: 315-334. Tabacchi E. (1995) Structural variability and invasions of pioneer plant communities in riparian habitats of the middle Adour River (SW France). Canadian Journal of Botany 73: 33-44. ter Braak C.J.F. (1987) CANOCO- A FORTRAN Program for Canonical Community Ordination by (Partial, Detrended, Cannonical) Correspondence Analysis, Principal Components Analysis and Redundancy Analysis. Agricultural Mathematics Group, Wageningen, Netherlands. Thebaud C. & Debussche M. (1991) Rapid invasion of Fraxinus ornus L. along the Herault River system in southern France: The importance of seed dispersal by water. Journal of Biogeography 18: 7-12. Thompson K. (1987) Seeds and seed banks. New Phytologist 106 (Suppl): 23-34. Thompson K. (1992). The functional ecology of seed banks. In: 'Seeds: The Ecology of Regeneration in Plant Communities.' (Ed. M. Fermer), Ch. 8 pp. 231-258. CAB International, London, UK. 278 Thorne C.R. (1990). Effects of vegetation on riverbank erosion and stability In 'Vegetation and Erosion.' (Ed. J.B. Thornes), John Wiley and Sons Ltd, New York, USA. Tille P.J. & Lantzke N.C. (1990) Busselton-Margaret River~Augusta land capability study. Land Resources Series 5. Agriculture Western Australia, Perth, Australia. Tilman D. (1986). Resources, competition and dynamics of plant communities In 'Plant Ecology.' (Ed. M. Crawley), pp. 51-75. Blackwell Scientific Publications, London, UK. Titus J.H. (1990) Microtopography and woody plant regeneration in a hardwood floodplain swamp in Florida. Bulletin of the Torrey Botanical Club 117: 429~437. van der Maarel E. (1988) Vegetation dynamics: Patterns in time and space. Vegetatio 77: 7-19. van der Moezel P.G. & Bell D.T. (1987) The effect of salinity on the germination of some Western Australian Eucalyptus and Melaleuca species. Seed Science and Technology 15: 239-246. van der Valk A.G. (1981) Succession in wetlands: A Gleasonian approach. Ecology 62: 688-696. van der Valk A.G. (1992). Establishment, colonisation and persistence. In; 'Plant Succession; Theory and Prediction.' (Eds. D.C. Glenn-Lewin, R.K. Peet & T.T. Veblen) Population and Community Biology pp. 60-102. Chapman and Hall, London, UK. van Groenewoud H. (1992) The robustness of Correspondence, Detrended Correspondence and TWINSPAN Analysis. Journal of Vegetation Science 3: 239-246. Vannote R.L., Minshall G.W., Cummins K.W., Sedell J.R. & Cushing C.E. (1980) The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences 37: 130-137. Veblen T.T. (1986) Age and size structure of subalpine forests in the Colorado front range. Bulletin of the Torrey Botanical Society 113: 225-240. Villalba R. & Veblen T.T. (1997) Regional patterns of tree population age structures in northern Patagonia: Climatic and disturbance influences. Journal of Ecology 85: 113-124. Vlahos S. & Bell D.T. (1986) Soil seed-bank components of the northern jarrah forest of Western Australia. Australian Journal of Ecology 11: 171-179. Wahren C.H.A., Papst W.A. & Williams R.J. (1994) Long-term vegetation change in relation to cattle grazing in subalpine grassland and heathland on the Bogong High Plains: An analysis of vegetation records from 1945 to 1994. Australian Journal of Botany 42: 607-639. 279 Walker K.F., Boulton A.J., Thorns M.C. & Sheldon F. (1994) Effects of water-level changes induced by weirs on the distribution of littoral plants along the River Murray, South Australia. Australian Journal of Marine and Freshwater Research.45: 1421-1438. Walker K.F., Sheldon F. & Puckridge J.T. (1995) A perspective on dryland river ecosystems. Regulated Rivers: Research amd Management 11: 85 - 104. Ward J.V. & Stanford J.A. (1983). The intermediate disturbance hypothesis: An explanation for biotic diversity patterns in lotic systems. In: 'Dynamics of Lotic Systems.' (Eds. T.D. Fontaine & S.M. Bartell) pp 347-355. Ann Arbor Science, Ann Arbor, Michigan, USA. Ward rv. & Stanford J.A. (1995) Ecological connectivity in alluvial river ecosystems and its disruption by flow regulation. Regulated Rivers: Research and Management 11: 105 - 119. Water and Rivers Commission (1997) The state of the northern rivers. WRAP 10. December 1997. Water and Rivers Commission, Perth, Australia. Watson J.R. (1991). The identification of river foreshore corridors for nature conservation in the South Coast Region of Western Australia. In: 'Nature Conservation 2: The Role of Corridors.' (Eds. D.A. Saunders & R.J. Hobbs) Ch. 7 pp. 63-68. Surrey Beatty and Sons, Sydney, Australia. Weaver D.M. & Reed A.E.G. (1998) Patterns of nutrient status and fertiliser practice on soils of the south coast of Western Australia. Agriculture, Ecosystems and Environment 67: 37-53 . Weaver J.E., Hanson H. C. & Aikman J.M. (1925) Transect method of studying woodland vegetation along streams. Botanical Gazette 80: 168-187. Webb L.L Tracey J.G. & Williams W.T. (1972) Regeneration and pattern in subtropical rain forest. Journal of Ecology 60: 675-695. Wellington A.B. & Noble I.R. (1985a) Seed dynamics and factors limiting recruitment of the mallee Eucalyptus incrassata in semi arid, southeastern Australia. Journal of Ecology 73: 657- 666. Wellington A.B. & Noble I.R. (1985b) Post-fire recruitment and mortality in a population of the mallee Eucalyptus incrassata in semi-arid, south-eastern Australia. Journal of Ecology 73: 645- 656. Werner P.A. (1990) Principles of restoration ecology relevant to degraded rangelands. Australian Rangeland Journal12: 34-39. 280 Westoby M., Walker B. & Noy-Meir I. (1989) Opportunistic management for rangelands not at equilibrium. Journal of Range Management 42: 266-274. Wheeler J.R., Rye B.L., Koch B.L. & Wilson A.J.G. (Ed.) .. (1992) Flora of the Kimberley Region. Department of Conservation and Land Management, Perth, Australia. Whelan R.J. (1989). The influence of fauna on plant species composition. In: 'Animals in Primary Succession: The Role of Fauna in Reclaimed Lands.' (Ed. J.D. Majer), Chp 5 pp 107-142. Cambridge University Press, Cambridge, UK. White P.S. (1979) Pattern, process, and natural disturbance in vegetation. Botanical Review 45: 229-299. White P.S. & Pickett S.T.A. (1985). Natural disturbance and patch dynamics: An introduction. In: 'The Ecology of Natural Disturbance and Patch Dynamics.' (Eds. P.S. White & S.T.A. Pickett) pp. 2-13. Academic Press Inc., New York, USA. Williams O.B. (1969) Studies in the ecology of the riverine plain. V. Plant density response of species in a Danthonia caespitosa grassland to 16 years grazing by merino sheep. Australian Journal of Botany 17: 255-268. Williams P.J., Pen L.J., Stein J.A., Stein J.L. & Prince B. (1996) The identification, conservation and management of wild rivers in the Kimberley region of Western Australia. In: 'River Management in Australia.' Melbourne, Australia. Williams R.J. (1990) Cattle grazing within subalpine heathland and grassland communities on the Bogong High Plains: Disturbance, regeneration and the shrub-grass balance. Proceedings of the Ecological Society of Australia 16: 255-265. Williams R.J. & Ashton D.H. (1987) Effects of disturbance and grazing by cattle on the dynamics of heathland and grassland communities on the Bogong high plains, Victoria. Australian Journal of Botany 35: 413-31. Williams W.D. (1983) Life in Inland Waters. Blackwell Scientific Publications, Melbourne, Australia. Willson M.F. (1992). The ecology of seed dispersal. In: 'Seeds: The Ecology of Regeneration in Plant Communities.' (M. Fenner Ed.), Ch. 3 pp. 61-86. CAB International, London, UK. 281 Willson M.F., Rice B.L. & Westoby M. (1990) Seed dispersal spectra: A comparison of temperate plant communities. Journal of Vegetation Science 1: 547-562. Wilson A.D. (1990) The effect of grazing on Australian ecosystems. Proceedings of the Ecological Society of Australia 16: 235-244. Wimbush D.J. & Costin A.B. (1979) Trends in Vegetation at Kosciusko. I. Grazing trials in the Subalpine zone, 1957-1971. Australian Journal of Botany 27: 741- 87. Wissmar R.C. & Swanson F.J. (1990). Landscape disturbances and lotic ecotones. In: 'The Ecology and Management of Aquatic -Terrestrial Ecotones.' (Eds.R.J. Naiman & H. Decamps), Man and the Biosphere 4 pp. 65-89. UNESCO, Paris, France. Withers J.R. (1978) Studies on the status of unburnt Eucalyptus woodland at Ocean Grove, Victoria. II. The differential seedling establishment of Eucalyptus ovata Labill. and Casuarina littoralis Salisb. Australian Journal of Botany 26: 465-483. Xiong S. & Nilsson C. (1997) Dynamics of litter accumulation and its effects on riparian vegetation: A review. The Botanical Review 63: 240-264. Xiong S. & Nilsson C. (1998). Dynamics of plant litter in riparian meadows: Setting a reference for management. In: 'European WetGrasslands: Biodiversity, Management and Restoration.' (Eds. C.B. Joyce & P.M. Wade), Ch 13 pp. 193-200. John Wiley & Sons Ltd., New York, USA. Yates C.J., Hobbs R.J. & Bell R.W. (1994) Landscape scale disturbances and regeneration in semi arid woodlands of south-western Australia. Pacific Conservation Biology 1: 214-221. Yates C.J., Taplin R., Hobbs R.J. & Bell R.W. (1995) Factors limiting the recruitment of Eucalyptus salmonophloia in remnant woodlands II. Post-dispersal seed predation and soil seed reserves. Australian Journal of Botany 43: 145-155. Young W.J. (1999). Hydrologic descriptions of semi-arid rivers: An ecological perspective. In: 'A Free flowing River: The Ecology of the Paroo River.' (Ed. R.T. Kingsford), pp. 77-96. National Parks and Wildlife Service NSW., Sydney, Australia. Zar J.H. (1984) Biostatistical Analysis. 2 Edn. Prentice-Hall, Inc, New Jersey, USA. 282 APPENDIXl List of species found at the study sites on the Blackwood River. A total of 226 species, including 44 exotic species, were recorded for the Blackwood River sites during the study. These species came from 54 families and and 163 genera. The most common families were Poaceae (29 species from 22 genera), Asteraceae (22 species from 20 genera), Cyperaceae (20 species from 10 genera) and Myrtaceae (17 species from 8 genera). Common species that were found at all five sites included the exotic annual herbs Hypochaeris glabra and Romulea rosea, the exotic annual grass Lolium rigidum and the native riparian tree Eucalyptus rudis. Melaleuca rhaphiophylla also occured at all sites other than the saline site (B3). Location of sites in the Blackwood River catchment are shown on the map in Figure 3.2 and include: B1a - ungrazed site near the town of Nannup at the downstream end of the middle catchment. B1b- grazed site near the town of Nannup at the downstream end of the middle catchment. B2a - ungrazed site near the town of Boyup Brook in the centre middle catchment. B2b - grazed site near the town of Boyup Brook in the centre middle catchment. B3a- ungrazed site on the Beaufort River in the upper catchment. B3b - ungrazed site on the Beaufort River in the upper catchment. B4 - grazed site in the middle catchment. BS - grazed site in the middle catchment. * Exotic species FAMILY SPECIES NAME SITES Aizoaceae Mesembryanthemum crystallinum L. B3b Amaranthaceae Alternathera nodiflora R. Br. B1a, B1b Ptilotus manglesii (Lindley) F. Muell. B3a, B3b Anthericaceae Arthropodium capillipes Endl. B2a, BS Arthropodium preissii Endl. B1a, B1b Chamaescilla corymbosa (R. Br.) F. Muell ex Benth. B2a,B2b,B3b,B4,B5 Corynotheca micrantha (Lindley) J.F. MacBride BS Hodgsoniola junciformis F. Muell. B1b, B2a, B2b, B4 Laxmannia grandiflora Lindley B3b Thysanotus multiflorus R. Br. B1a 283 Anthericaceae Tricornye elatior R. Br. Bla, B2a Apiaceae Dauctts glochidiatus (Labill.)Fischer, C. & Ave-Lall. B3b Eryngium rostratum Cav. B3b Asparagaceae Asparagus asparagoides (L.) Wight B2a, B2b Asteraceae Angianthus priessianus (Steetz) Benth. B4 Arctotheca calendula (L.) Levyns.* Blb, B2a, B2b Aster subulatus Michaux* Blb, B2a, B2b Calocephalus angianthoides (Steetz) Beth. B3a, B3b Carduus pycnocephalus L. ~- Blb Circium vulgare (Savi) Ten.* B4 Cotula turbinata L.* B3b, B4, BS Dittrichia graveolens (L.) Greuter* Bla, Blb Gnaphalium sphaericium Will d.* Blb Hyalosperma cotula (Benth.) P.G. Wilson B3a, B3b Hypochaeris glabra L.* Blb, B2a, B2b, B3a, B3b, B4, BS Podolepis lessonii (Cass.) Benth. B3a, B3b Psuedognaphalium luteoalbum (L.) B2a, B2b Hilliard&B.L.Burtt* Quinetia urvillei Cass. B3b Siloxerus humifusus Labill. B3b Sonchus asper Hill* B4 Sonchus oleraceus L. * Blb, B2a, B2b, B3b, B4 Rhodanthe citrina (Benth.) P.G. Wilson B3a, B3b Rhodanthe manglesii Lindl. B2a Trichocline spathulata (Cunn. ex DC.) J.H. Wills B3b Ursinia anthemoidies (L.) Poiret* B2a, B2b, B3b, B4 Campanulaceae Wahlenbergia capensis (L.) A. DC.* B4 Caryophyllaceae Cerastium glomeratum Thuill. * Blb, B2a, B4 Petrorhagia velutina (Guss.) P. Ball & Heyw.* B4 Sagina apetala Ard. * B4 Spergularia salina J.S. Presl & C. Presl* B3b, B4 Casuarinaceae Casuarina obesa Miq. B3a, B3b Centrolepidaceae Aphelia brizula F. Muell. Blb,B3b Centrolepis aristata (R.Br.) Roemer & Schultes B3b Chenopodaceae Atriplex prostrata M. Boucher ex DC* Blb, B2a, B2b Chenopodium album L.* B2a, B2b Chenopodaceae Halosarcia lepidosperma Paul G. Wilson B3a, B3b 284 Crassulaceae Crassula alata (Viv.) Berger* B3a, B3b, BS Crassula glomerata P. Bergius* BS Cyperaceae Carex appressa R. Br. B2a Chorizandra enodis Nees B2a, B3a, B3b, B4, BS Cyathochaeta avenacea Benth. BS Cyperus sp. L. Bla, Blb Gahnia sp Forst. & Forst. B2a, B4 Gahnia trifida Labill. B3a, B3b Isolepis cernua (M. Vahl) Roemer & Schultes B2a, B2b, BS Isolepis nodosa (Rottb.) R. Br. BS Lepidosperma angustatum R. Br. Bla, B2a Lepidosperma drummondii Benth. B2a Lepidosperma effusum Benth. Bla Lepidosperma leptostacyum Benth. B2a, B4 Lepidosperma longitudinale Labill. Bla, B4 Lepidosperma squamatum Labill. BS Lepidosperma tenue Benth. B2a, BS Mesomelaena tetragona (R. Ik) Benth. BS Schoenus elegans S.T.Blake B2a, B3b Schoenus nanus (Nees) Benth. B3b Schoenus sp.L. B3a, B3b Tetraria octandra (Nees) Kuek. Bla, Blb, B2a, BS Dasypogonaceae Lomandra micrantha (Endl.} Ewart B3b Lomandra sericea (Endl.) Ewart Bla Dilleniaceae Hibbertia montana Steudel B2a Droseraceae Drosera gigantea Lindley Bla Drosera glanduligera Lehm. BS Drosera macrantha Endl. Bla,Blb, B2a Drosera pallida Lindley B4, BS Epacridaceae Astroloma drummondii Sonder Bla Astroloma pallidum R. Br. Bla, B2a Leucopogon propinquis R. Br. Bla Euphorbiaceae Euphorbia australis Boiss. * B3b Phyllanthus calycinus Labill. Bla Poranthera microphylla Brongn. B4 Gentianaceae Centaurium erthyraea Rafn. * Blb, B2a Geraniaceae Erodium botryus (Cav.) Bertol.* Blb, BS Geranium retrorsum L'Her. ex DC. BS 285 Haemodoraceae Conostylis aculeata R. Br. Bla, B2a Conostylis setosa Lindley Bla Haemodorum laxum R. Br. B2a, B2b Haemodorum spicatum R. Br. Bla Hypoxidaceae Hypoxis glabella R. Br. Bla, B4 Irdiaceae Hesperantha jalcata (L. f.) Ker Gawler B2a Hexaglottis lewisiae Goldblatt* B2a Homeria jlaccida Sweet* B4, B5 Patersonia occidentalis R. Br. B2a Patersonia umbrosa Endl. Bla Romulea rosea (L.) Ecklon* Bla, Blb, B2a, B2b, B3b, B4, B5 Juncaceae Juncus bufonius L.* B2a, B2b, B4 Juncus capitatus Weigel * Bla, B2b, B3a Juncus pallidus R. Br. B4, B5 Juncus pauciflorus R.Br. B4 Juncus subsecundus Wakef. Bla, Blb Lamiaceae Mentha pulegium L.* Blb Lauraceae Cassytha racemosa R. Br. B2a, B4, B5 Lobeliaceae Lobelia alata Labill. Bla, B2a, B2b, B3a, B3b Monopsis deblis (L.f.) C. Presl.* B4 Menyanthaceae Villarsia capitata Nees Bla Mimoaceae Acacia pulchella R. Br. Bla, B2a Acacia acuminata Benth. B4 Acacia denifera Benth. Bla Acacia saligna (Labill.) H. L. Wendl Bla, Blb, B2a, B2b, B4 Myoporaceae Myoporum caprarioides Benth. Bla Myrtaceae Astartea jasicularis (Labill.) DC. Bla, Blb, B2a, B2b, B5 Corymbia calophylla(Lindl.)K.D.Hill&L.A.S.Johnson Bla, Blb, B2a, B2b, B4, B5 Eucalyptus marginata Donn ex Smith Bla, Blb, B2a, B2b, B4, B5 Myrtaceae Eucalyptus rudis Endl. Bla, Blb, B2a, B2b, B3a, B4, B5 Eucalyptus wandoo Blakely B2a, B3b Hypocalymma angustifolium Endl. Bla, B2a Kunzea ericifolia (Smith) Heynh. B2a 286 Llwtskya acutifolia Lindl. B3 Melaleuca acuminata F. Muell. B2a Melaleuca hamulosa Turcz. B3a Melaleuca incana R. Br. B2a Melaleuca lateriflora Benth. B3a Melaleuca rhaphiophylla Schauer Bla, Blb, B2a, B2b, B4, BS Melaleuca uncinata R. Br. B3a Melaleuca urceolaris F. Muell. ex Benth. B3a Melaleuca viminea Lindley Bla Verticordia densiflora Lindley B3a Onagraceae Epilobium billardierianum Ser. B2a Orchidaceae Elythranthera brunonis (Endl.) A.S. George B3b Diuris longifolia R. Br. B2a Microtis unifolia (G. Forster) H.G. Reichb. Bla Monadenia bracteata (SW.) T. Durand & Schinz* Blb,B2b,B3b,B4 Pterostylis nana R. Br. B3b Thelymitra antennifera (Lindl.) J.D. Hook. B2a Oxalidaceae Oxalis corniculata L.* Bla, B2a, B2b, B3b, B4 Oxalis hirta L. * B2a Oxalis pes-caprae L. * B2a Papilionaceae Brachysema celsianum Lemaire B2a Bossiaea linophylla R. Br. Bla Davesia horrida Priess ex Meissner Bla Gastrolobium bilobum R. Br. Bla Gompholobium marginatum R.Br. Bla Hovea trisperma Benth. Bla Jacksonia furcellata (Bonpl.) DC B2a, B4 Jacksonia sternbergiana Huegel B2a Lotus suaveolens Pers. * B4, BS Lupinus angustfolius L. B4 Papilionaceae Trifolium angustifolium L.* B4 Trifolium campestre Schreber * Bla, Blb, B2a, B2b, B4, BS Trifolium repens L. * B4 Trifolium subterraneum L. * Blb, B4 Viminaria juncea (Schrader & Wendl.) Hoffsgg. Bla Philydraceae Philydrella pygmeae (R. Br.) Caruel B3b 287 Phorminaceae Dianella divaricata R. Br. Bla1 Blb1 B2a1 B3a, B4 Stypandra glauca R. Br. B2a1 B3a Pittosporaceae Billardiera floribunda (Putterl. )F .Muell. Bla Sollya heterophylla Lindley Bla1 B2a Plantaginaceae Plantago lanceolata L.* Bla1 B3b1 B4 Poaceae Aira caryophyllea L.* Bla1 Blb, B2a1 B2b, B3a1 B3b 1 B4 1 B5 Amphipogon turbinatus R.Br. B 1 a Astrostipa compressa (R.Br.) S.W.L.Jacobs & J.Everett B2a, B5 Austrostipa elegantissima Labill. S.W.L. Jacobs & B3a, B3b J.Everett Astrostipa semibarbata R. Br. B2a1 B2b Avena Jatua L.* Blb1 B2a1 B2b Briza maxima L.* Blb1 B2a1 B2b, B41 B5 Briza minor L. * B2a, B2b 1 B3a1 B3b, B4 1 B5 Bromus diandrus Roth* B2a1 B2b 1 B4 1 B5 Bromus hordeaceus L. * Blb, B2a, B2b, B4, B5 Bromus madritensis L.* B4 Cynodon dactylon (L.) Pers. Bla1 Blb1 B2a 1 B2b, B4 1 B5 Danthonia caespitosa Gaudich. B2a Ehrharta calycina Smith* B2a1 B2b 1 B4 1 B5 Ehrharta longiflora Smith* B2a1 B2b1 B5 Holcus lanatus L.* Bla Hordeum leporinum Link* B3b 1 B4 1 B5 Lolium perenne L. * BS Lolium rigidum Gaudin* Blb1 B2a1 B2b 1 B3b, B4, BS Microlaena stipoides (Labill.) R. Br. Bla, B2a1 B4 Neurachne alopecuroidea R. Br. Bla1 Blb1 B2a, B2b, B3b1 B4 Parapholis incurva (L.) C.E. Hubb.* B3b Pennisetum cladestinum Hochst. ex Chiov. * Blb Pentaschistis airiodes (Nees) Stapf * B3b Phalaris minor Retz. * B4 Polypogon monspeliensis (L.) Desf.* B3a1 B3b 288 Puccinellia ciliata Bor * B3b Sporobolus virginicus (L.) Kunth* Blb Vulpia myuros (L.) Gmelin* Blb, B2b, B4, BS Polygalaceae Comesperma confertum Labill. Bla, B2a Comesperma virgatum Labill B3b Polygonaceae Emex australis Steinh.* Blb Muelenbeckia adpressa (Labill.) Meissner Bla Persicaria decipiens (R.Br.) K.L. Wilson Bla, Blb, B4 Rumex acetosella L.* B4, BS Rumex crispus L. * Bla, B2a, B2b, B4, BS Primulaceae Anagallis arvensis L.* Blb, B2a, B2b Proteaceae Banksia seminuda (A.S.George) B.L.Rye Bla, Blb, B2a Grevillea diversifolia Meissner Bla Grevillea manglesii (,Graham) Plauch. Bla Grevillea pilulifera (Lindley) Druce Bla Grevillea vestita (Endl) Meissner B2a Hakea lissocarpha R. Br. B2a Hakea prostrata R. Br. B2a Ranunculaceae Ranunculus muricatus L.* Bla Restionaceae Anarthria gracilis R.Br. B2a Desmocladus fiexuosa (R. Br.) Benth. Bla, Blb, B2a Hypolaena exsulca R. Br. B3b Lepyrodia glauca (Nees) F. Muell. Blb Meeboldina coangustata (Nees) B.G.Briggs & L.A.S. Blb Johnson Restionaceae Meeboldina royceii L.A.S. Johnson & B.G. Briggs BS Rosaceae Rubus discolor Weihe & Nees * Bla, Blb Rubiaceae Galium diaricatum Pourret ex Lam.* Blb Galium murale (L.) All.* B3b,B4 Santalaceae Exocarpos sparteus R. Br. B2a Sapindaceae Dodonaea viscosa Jacq. B2a Scrophulariaceae Bellardia trixago (L.) All.* B3b Parentucellia latifolia (L.) Caruel* Bla, Blb, B2b, B3b Parentucellia viscosa (L.) Caruel* B2a Stylidiaceae Levenhookia pusilla R. Br. B2a Stylidium amoenum R. Br. Bla Stylidium piliferum R. Br. B2a Stylidium schoenoides DC. Bla 289 Xanthorrhoeaceae Xanthorrhoea preissii Endl. Bla, B2a Zaminaceae Macrozamia reidii (Fischer ex Gaudich.) C. Gardner Bla, B2a 290 APPENDIX2 List of species found at the study sites on the Ord River A total of 151 species, including 23 exotic species, were recorded for the Ord River sites during the study. These species came from 46 families and 104 genera. The most common families were Poaceae (24 species from 18 genera), Cyperaceae (19 species from 5 genera), Papilionaceae (12 species from 7 genera) and Myrtaceae (10 species from 4 genera). Common species that were found at all five sites included the trees Eucalyptus camaldulensis, Melaleuca leucadendra, Lysiphyllum cunninghamii and Pandanus spiralis and the exotic creeper Passiflora foetida. Location of sites in the Ord River catchment are shown on the map in Figure 3.1 and include: 01 - Lower Ord River upstream of Carlton crossing. 02 - Lake Kununurra near the settlement of Crossing Falls 03 -Upper Ord on the Behn River near the Duncan Rd crossing. 04 -Upper Ord on the Negri River near the Duncan Rd Crossing 05- Upper Ord at the Water and Rivers Commission gauging site at Old Ord River Station. * Exotic species FAMILY SPECIES NAME SITES Amaranthaceae Achyranthes aspera L. 01, 03, 04, 05 Aerva javanica (Burm.f.) Juss.ex Schultes* 04,05 Alternanthera nana R.Br. 01 Alternanthera nodiflora R.Br. 01, 03, 04, 05 Gomphrena affinis F. Muell. ex Benth 04,05 Ptilotus exaltatus N ees 05 Asclepiadaceae Calotropis procera (Aiton) W.T. Aiton* 05 Marsdenia viridiflora R. Br. 01 Sarcostemma esculentum (L.f.) Holm 01 Tylophora flexuosa R.Br. 05 Asteraceae Bidens pilosa L.* 01 Blumea axillaris (Lam) DC. 05 Blumea diffusa R. Br ex Benth. 01 Flaveria australasica Hook. 05 Vernonia cinerea (L.) Less. 02 Asteraceae Xanthium occidentale Bertol. * 01 291 Bixaceae Cochlospermum fraseri Planchon B3 Bombacaceae Adansonia gregorii F. Muell. 01 Boraginaceae Heliotropium indicum L.* 01 Heliotropium sp L. 01 Caesalpiniaceae Lysiphyllum cunninghamii (Benth.) de Wit 01, 02, 03, 04, 05 Parkinsonia aculeata L.* 01, 02,05 Capparaceae Cleome viscosa L.* 03,04 Centrolepidaceae Centrolepis exserta(R.Br.) Roemer & Schultes 05 Chenopodiaceae Dysphania plantaginella F. Meull. 05 Salsola kali L.* 04,05 Combretaceae Terminalia canescens (DC.)Radlk.ex T.Durand 01,04,05 Terminalia hadleyana W. Fitzg. 03 Terminalia oblongata F. Muell. 03,04 Terminalia platyphylla F.Muell. 03,04,05 Commelinaceae Commelina ciliata T.D. Stanley 01,05 Commelina ensifolia R.Br. 01 Convolvulaceae Ipomoea muelleri Benth. 05 Jacquemontia browniana Ooststr. 03 Cucurbitaceae Luffa graveolens Roxb. 01 Cyperaceae Bulbostylis barbata (Rottb.) C.B.Clarke 01 Cyperus breviculmus R.Br. 05 Cyperus castaneus Willd. 01 Cyperus conicus (R.Br.) Boeckler 02 Cyperus cunninghamii (C.B.Clarke) C.Gardner 01,02 Cyperus cuspidatus Kunth 01 Cyperus difformis L. 01 Cyperus lwloschoenus R. Br. 01,05 Cyperus sp L. 03 Cyperus vaginatus R.Br. 02 Eleocharis atropurpuea Kunth 02 Eleocharis sp A 01 Fimbristylis caespitosa R. Br. 01 Fimbristylis littoralis Gaudich. 05 Fimbristylis phaeoleuca S. T. Blake 01 Cyperaceae Fimbristylis punctata R.Br. 01 Fimbristylis sp V ahl 05 Schoenoplectus dissachanthus (S.T.Blake) Raynal 02 292 Schoenoplectus littoralis (Schrader) Palla 03,04 Euphorbiaceae Euphobia hirta L.* 01 Euphorbia australis Boiss. 01,05 Excoecaria parvifolia Muell. Arg. 05 Flueggea virosa (Roxb. ex Willd.) Voigt 01, 02, 03, 04 Petalostigma pubescens Domin 01 Phyllantus maderaspatensis L. 01 Hernandiaceae Gyrocarpus americanus Jacq. 01,03 Lamiaceae Hyptis suaveolens (L.) Poit.* 03,04 Lecythidaceae Barringtonia acutangula (L.) Gaertner 01,03,04 Malvaceae Abelmoschus ficulneus (L.) Wight & Arn. ex Wight 01 Malvaceae Abutilon indicum (L.) Sweet 01 Abutilon otocarpum F.Muell. 01 Abutilon oxycarpum (F. Muell.) F.Muell.exBenth. 04,05 Hibiscus sabdariffa L. * 03 Malvastrum americanum (L.) Torrey* 03,04,05 Sida acuta Burrm. f.* 01 Sida fibulifera Lindley 05 Sida spinosa L. 01,04 Marsileaceae Marsilea mutica Mett. 02 Melastomataceae Melastoma affine D.Don 01 Meliaceae Melia azedarach L. 01,04 Menispermaceae Tinospora smilacina Benth. 01 Mimosaceae Acacia farnesiana (L.)Willd.* 01, 03, 04, 05 Acacia holosericea Cunn. ex Don 01 Calthormion umbellatum (Vahl)Kosterm. 01,02 Dichrostachys spicata (F. Muell.)C. Gardner 01 Neptunia monosperma F. Muell. 03,05 Moraceae Ficus coronulata Miq. 01,03,04,05 Fictts opposita Miq. 01,03 Ficus racemosa L. 01,04. 05 Myrtaceae Eucalyptus camaldulensis Dehnh. 01, 02, 03, 04, 05 Myrtaceae Eucalyptus gymnoteles L.A.S. Johnson & K.D.Hill 04,05 Eucalyptus houseana W. Fitzg. ex Maiden 01,03, 04 Eucalyptus microtheca F. Muell. 01, Lophostemon grandiflorus (Benth.) P.G.Wilson & J.T.Waterhouse03, 04,05 Melaleuca argentea W. Fitzg. 01,02 293 Melaleuca leucadendra (L.) L. 01, 02, 03, 04, 05 Melaleuca minutifolia F.Muell. 01 Melaleuca viridiflora Sol. ex Gaertner 01,03 Syzygium eucalyptoidies (F.Muell.) B. Hyland 01,03 Nymphaeaceae Nymphaea violacea Lehm. 02 Oleaceae Jasminum didymum G. Forster 05 Onagraceae Ludwigia perennis L. 05 Opiliaceae Opilia amentacea Roxb. 01 Pandanaceae Pandanus spiralis R.Br. 01, 02, 03, 04, 05 Papilionaceae Abrus precatorius L. 01,02 Aeschynomene indica L. 01 Clitoria ternatea L. * 01 Crotalaria medicaginea Lam. 05 Crotalaria novae-hollandiae DC. 04 Crotalaria retusa L. 04,05 Erythrina vespertilio Benth. 01 Indigofera linifolia (L.f.)Retz. 01,05 Indigofera parviflora Heyne ex Wight & Arn. 01 Sesbania cannabina (Retz.) Poiret 01 Sesbania formosa (F.Muell.) N.Burb. 01 Sesbania simpliciuscula F.Muell. ex Benth. 04 Passifloraceae Passiflora foetida L. * 01, 02, 03, 04, 05 Poaceae Aristida hygrometrica R. Br. 01,03 Bothriochloa bladhii (Retz.) S.T.Blake 03 Brachyachne convergens (F. Muell.) Stapf 01 Cenchrus ciliaris L. * 04,05 Cenchrus echinatus L. * 04,05 Cenchrus setiger V ahl * 04,05 Poaceae Chionachne hubbardiana Henrard 01,05 Chloris barbata Sw. 01 Chrysopogon fallax S.T. Blake 04 Cynodon dactylon (L.) Pers. 01,04,05 Dichanthium sericeum (R.Br.) A. Camus 04,05 Echinochloa colona (L.) Link* 01 Eragrostis cummgii Steudel 01,02 294 Eragrostis falcata (Gaudich.) Gaudich. ex Steudel 04,05 Eragrostis tenellula (Kunth) Stuedel 05 Eulalia aurea (Bory)Kunth 04,05 Heteropogon contortus (L.)P.Beauv.ex Roemer & Schultes 03,04,05 Iseilema vaginiflorum Domin 01 Panicum decompositum R.Br. 01 Panicum mindanaense M err. 01 Panicum seminudum Domin 01 Paspalidium jubiflorum (Trin.)Hughes 01 Phragmites karka (Retz.) Trin. 01,05 Whiteochloa airoides (R. Br.) Lazarides 01 Polygonaceae Persicaria attenuata (R. Br.) Sojak 01,04,05 Portulacaceae Calandrinia quadrivalvis F .Muell. 01 Portulaca oleracea L. 04,05 Rhamnaceae Ziziphus quadrilocularis F.Muell. 04 Rubiaceae Nauclea orientalis L. 01,03, 04, 05 Santalaceae Santalum lanceolatum R.Br. 05 Sapindaceae Cardiospermum halicacabum L. * 01 Atalaya hemiglauca (F.Muell.) F.Muell. ex Benth. 01 Solanaceae Datura inoxia Miller* 05 Physalis minima L. * 04,05 Solanum lucani F.Muell. 01 Tiliaceae Grewia breviflora Benth. 01 Typhaceae Typha domingensis Pers. 01,02,05 Violaceae Hybanthus enneaspermus (L.) F.Muell. 01 Vitaceae Cissus adnata Roxb. 01 295 Appendix3 List of submitted publications and conference presentations related to the thesis topic. Pettit N.E. and Froend RH. (in press) Variability in flood disturbance and the impact en riparian tree recruitment in two contrasting river systems. Weland Ecology and Management x pp __ . Pettit N.E., FroendRF. and Davies P.M. (in press) Identifying the natural flow regime and the relationship with riparian vegetationfor two contrasting Western Australian rivers. Regulated Rivers: Research and Management__ . Pettit N.E. (in press) Riparian vegetation of a permanent waterhole on Cooper Creek, south western Queensland. Proceedings of the Royal Society of Queensland __ . Askey-Doran, M., Pettit, N.E., Robins, L. and McDonald, T. (1999) The role of vegetation in riparian management. In 'Riparian land management technical guidelines, volume one: Principles of sound management' (Eds. S. Lovett and P. Price). Chp 8 pp 97-120. Land and Water Resources Research and Development Corporation, Canberra. Askey-Doran, M. and Pettit, N.E. (1999) Some impacts of human activity on riparian lands. In 'Riparian land management technical guidelines, volume one: Principles of sound management' (Eds. S. Lovett and P. Price). Chp 10 pp 137-145. Land and Water Resources Research and Development Corporation, Canberra. Pettit, N.E. and Froend, RH. (1999) Riparian vegetation on the Blackwood and Ord Rivers in Western Australia and some factors influencing regeneration. Centre for Ecosystem Management ReportNo.1999-1 prepared for the Water and Rivers CommissionNovember1999102pp. Pettit, N.E., Froend, RH. and Davies, P.M. (1999) The influence of river hydrology on tree establishment in the riparian zone of two contrasting river systems. In : Ecological Society of Australia Open Forum and Symposium Conference, Fremantle 26th September - 1st October 1999. Pettit, N.E. and Froend, RH. (1998) Tree recruitment in the riparian zone of two contrasting river systems. In: INTECOL98 VII International Congress of Ecology, Florence, Italy, July 1998. 296 Froend, R.H., Pettit, N.E. and Payne, A. (1998) Riparian vegetation on the Fitzroy River: Water requirements. In: Limnology of the Fitzroy River, Western Australia: A Technical Workshop, Perth Western Australia, February 1998. Pettit, N.E., Froend, R.H., Davies, P. and Horwitz, P. (1996) Riparian vegetation: What factors influence recruitment and regeneration. Poster paper: INTECOL96 International Wetlands Conference, Perth, Western Australia October 1996. 297 Appendix 4 Copy of Manuscript accepted for publication in Wetland Ecology and Management. (due for publication in September 2000) Wetlands Ecology and Management 00: 1-13,2000. 1 © 2000 Kluwer Academic Publishers. Printed in the Netherlands. Variability in flood disturbance and the impact on riparian tree recruitment in two contrasting river systems N.E. Pettit & R.H. Froend Centre for Ecosystem Management, Edith Cowan University, Joondalup, Western Australia 6027 Received 15 March 1999; accepted in revised form 20 December 1999 Key words: flood disturbance, riparian vegetation, tree recruitment, water regime Abstract The vegetation within the riparian zone performs an important ecological function for in-stream processes. In Australia, riparian zones are regarded as the most degraded natural resource zone due to disturbances such as river regulation and livestock grazing. This study looks at factors influencing vegetation dynamics of riparian tree species on two contrasting river systems in Western Australia. The Blackwood River in south-western Australia is influenced by a Mediterranean type climate with regular seasonal winter flows. The Ord River in north-western Australia is characterized by low winter base flows· and episodic, extreme flows influenced by monsoon rains in the summer. For both rivers, reproductive phenology of studied overstory species is timed to coincide with seasonal river hydrology and rainfall. An even distribution of size classes of trees on the Blackwood River indicated recruitment into the population is continual and related to the regular predictable seasonal river flows and rainfall. In contrast, on the Ord River tree size class distribution was clustered, indicating episodic recruitment. On both rivers tree establishment is also influenced by elevation above the river, microtopography, moisture status and soil type. In terms of vegetation dynamics riparian vegetation on the Ord River consists of long periods of transition with short lived stable states in contrast to the Blackwood river where tree population structure is characterized by long periods of stable states with short transitions. Introduction and forest (Bell and Koch, 1980) as well as lentic wet lands (van der Valk, 1981; Boutin and Keddy, 1993; In most countries, changes to land use in river catch Hills et al., 1994). However, little work has been ments have had a major impact on riparian zones. done specifically on riparian vegetation (Barnes, 1985; Impacts and .activities include industrial and urban Baker, 1988; Baker, 1990; Nilsson et al., 1994) and pollution (Jeffries and Mills, 1990), agriculture and especially not in Australia (Dexter, 1978; Bren and forestry (Abido, 1985; Shah and Thames, 1985; Fleis Gibbs, 1986; Bren, 1992; Roberts, 1993). In many chner, 1994). In Australia, riparian zones are regarded areas of Australia there is a high level of community as the most degraded natural resource zone (CEPA, interest in restoration of rivers. However there is a lack 1992). These ecosystems are important in terms of of ecologically- based knowledge of the processes of overall landscape ecology and for the protection and regeneration within riparian vegetation. maintenance of the instream river ecosystem (Bunn, Traditional 'Clemensian' successional models 1993). In order to manage these systems knowledge have been found to be inappropriate for many veget of the ecological processes that are important for ation communities (Glenn-Lewin and van der Maarel, regeneration of the riparian vegetation is required. 1992). Rather than unidirectional succession, dy Ecological studies on the regeneration of plant namics within vegetation communities create a series communities have been done for many types of plant of patches made up of transitions between more or communities including grasslands (Grime and Hillier, less stable states mediated by disturbance, climate or 1992; Mcintrye et al., 1995), heath (Bell et al., 1982), management (Hobbs, 1994). The present state of an GSB/Prepr; WETL SM-02/DISK Pipsnr. 261102; Ordernr.: 59-JM.WEB2C (wetlkap:bio2fam) v.1.1 wetlsrn02.tex; 29/02/2000; 20:52; p.l HOR iT PROOF· SPARE SET -- 2 ecosystem is the result of a particular history of dis cies that is known to hybridize with E. camaldulensis turbance, climatic factors and/or management. The (Boland et al., 1989). Melaleuca leucadendra on the regeneration process and longevity of vegetation can Ord River and M. rhaphiophylla on the Blackwood also be critical in determining the stage and rate of River are also similar species with a high tolerance transition states. of waterlogged conditions. By studying functionally State and transition models have been sugges equivalent riparian tree species in two contrasting river ted as a tool for rangeland management (Westoby systems and climatic zones, the influence of the dif et al., 1989; George et al., 1992), as an alternative ferent physical environments on vegetation persistence to linear succession, to allow for different vegeta and recruitment in the riparian zone can be elucidated. tion community composition (states) depending on This paper compares recruitment patterns of func the transitions pattern. Non-equilibrium vegetation tionally similar overstory species that occur on two dynamics may be described by discrete states and contrasting river systems. These patterns will be dis definable transitions between states. Transitions are cussed in terms of a state and transition model for pop usually instigated by natural events and may occur ulations of these riparian species and the implications quickly or over an extended period and change to a for riparian vegetation management. persistent state depending on events during the trans ition (George et al., 1992). This model may provide an appropriate framework for management of degraded Methods riparian zones and assist in interpreting the response of riparian vegetation to disturbance in terms of possible Study sites states and transitions of riparian vegetation. Large scale disturbances, such as high water flows The study sites are located in the south-west (Black and flooding, are likely to be important factors in the wood River) and in the Kimberley region of the north dynamics of the riparian community. Recruitment, es west of Western Australia (Ord River) (Figure 1). The tablishment and survival of riparian trees has been Blackwood river was chosen as being representative of thought to be closely tied to the hydrological regime temperate river systems in the south-west of Western (Barnes, 1985; Bradley and Smith, 1986; Johnson, Australia while in contrast the Ord is a sub-tropical 1994). Flood induced disturbance may instigate the northern Australian river with flows greatly affected transition from stable states of plant composition and by monsoon rains resulting in large seasonal flows. population structure. It can be hypothesized therefore, The riparian zone of each of these rivers has, at least in that river systems with frequent high energy flood some sections, become degraded by livestock grazing events will support riparian communities with long and as a consequence of agricultural clearing. periods of transition and rather shorter stable states. The Ord River, in the tropical north-west of Aus Blackwood River tralia, is affected by seasonal monsoon rains during the The Blackwood River originates in the low rainfall wet season, when episodic high intensity events result area and flows westward to enter the sea at the south in high river flows. For the rest of the year virtually west tip of the continent. The river is approximately no rain falls and the river, especially those sections 300 km long and carries the largest flow of any river in draining the semi-arid interior, dry up completely or to the south west. Rainfall in the catchment varies from a series of permanent pools. The Blackwood River is 400 mm yr-1 in the east to 1200 mm yr-1 near the one of the largest rivers in south-western Australia and coast. The upper and middle reaches are greatly influ drains the low rainfall agricultural area before flowing enced by large areas of cleared agricultural land with through forest and entering the ~ea on the southwest many areas having been cleared for 100 years and with corner of the continent. River flows are influenced by livestock grazing taking place right up to the river's the Mediterranean climate with regular high seasonal edge (Olsen and Skitmore, 1991). Clearing in the nat winter flows and low summer base flows. Two dom urally highly saline soils of the upper catchment has inant overstory species from each river system have resulted in rising salinity levels of the river with the been chosen for this study. Eucalyptus camaldulensis waters of the upper Blackwood tributaries (Arthur and occurs as a dominant riverine species throughout con Beaufort Rivers) being saline to brackish (Schofield tinental Australia except for the south west where it et al., 1988). Water in the middle to lower sections is replaced by E. rudis which is a closely related spe- which are drained by mainly forested areas is marginal wetlsm02.tex; 29/02~2000; 20:52; p.2 3 -- 20°S -- tN Scale 0 200 400km Figure 1. Location of the Ord and Blackwood river catchments in Western Australia. to fresh before becoming influenced by the salt wedge dammed in two places which results in three discreet as the river nears the coast. sections. Upstream from the Ord River Dam and Lake Argyle the river flow is essentially natural with flows OrdRiver influenced by the climate, consisting of a distinct wet The Ord River stretches from it's source north-west of season (December to March) with monsoon rainfall Halls Creek to empty into Cambridge Gulf north-west initiating high river flows and a dry season when the of Kununurra. Vegetation in the catchment is predom river dries up to a series of pools. In the middle sec inately open grassy woodland (Beard, 1979) which is tion of the river, between the Ord River Dam and the grazed by cattle. On the Ord River plains in the up diversion dam at Kununurra, the river has become per per catchment degradation of the vegetation through manent and water levels alter little throughout the year. a long history of overgrazing has resulted in severe Downstream from the diversion dam, water flows are erosion problems and extremely high sediment loads artificially regulated to provide water for irrigation. in the river. On large areas, such as Ord River Sta tion, rehabilitation through fencing and de-stocking has taken place (Fitzgerald, 1967). The river has been wetlsrn02.tex; 29/02/2000; 20:52; p.3 4 Tree species parameters were recorded including surface soil type, microtopography, surface soil moisture status, eleva Eucalyptus camaldulensis and E. rudis (in the in tion above river base flow level, bare ground, litter, formal subgenus Symphyomyrtus (Boland et al., distance to water and distance to nearest seed tree. 1989)) are taxonomically and functionally very similar Frequency of seedling plots for each environmental species. Both are riverine trees which grow to 20 m tall parameter was. collated. For quantitative parameters and fill the same ecological niche in the riparian zone, (elevation, litter, distance to water and distance to occupying the river bank and in some instances the nearest seed tree) correlation analysis against seedling adjacent floodplain. E. camaldulensis is widespread density was performed. For each categorical parameter throughout mainland Australia having essentially a (soil type, microtopography, surface soil moisture linear geographic distribution along watercourses. E. status) a Krusall-Wallis test (non-parametric ANOVA rudis has a much more restricted distribution to south analog) was conducted for seedling density to test if western Australia where it replaces E. camaldulensis there were a significant difference between categories as the main riparian species. Both species are toler for each parameter. As there was unequal numbers of ant of periodic inundation and waterlogged conditions. quadrats in the categories for each parameter the five Melaleuca leucadendra on the Ord River can be eo seedling density scores around the median score were dominant with E. camaldulensis but tends to occupy used for the analysis. the wetter littoral edges of the riparian zone probably due to greater waterlogging tolerance. Similarly M. rhaphiophylla on the Blackwood River occurs with Results E. rudis and is a small tree which tends to persist on the fringe of the littoral zone and in other areas where prolonged inundation can exclude other species. Long term hydrograph and climate On the Ord River over the, 1975 to, 1993 period Survey methods large river flows have been confined to the wet sea son with 100% of dischargc;s greater than 500 m3/s For each river long term climate and river flow occurring between November and March. Recorded data were collected from nearby climate and stream flows have only occurred 3 times outside this period gauging stations operated by the Waters and Rivers in June, 1978, May, 1979 and October, 1988 (Fig Commission of Western Australia. Sites were selec ure 2a). Magnitude of peak flows are highly variable ted at two areas on reaches of both rivers. The river with maximum discharges of > 15000 m3 Is occurring reach scale was considered the appropriate scale for 4 times over the, 19 year period with 4 years having a this study as it includes a range of disturbance patches discharge <1000 m3/s with only 375 m3/s in, 1992. of different ages but in close proximity to reduce other Average maximum discharge (± SE) for the period environmental sources of variation such as changes in from November to March is 5565 m3/s ± 1475 with geology. At the study sites on the upper Ord river in the January having the highest mean discharge and March umegulated section and on the middle section of the the lowest. Rainfall is also confined to the wet season Blackwood River (Figure 1) six 5 m x 200 m belt tran with 87% falling in this period with dry season falls sects where set up (in, 1996) perpendicular to the river. of >10mm having occurred only 6 times in the, 19 Along each transect the diameter at breast height (1.3 year period. Rainfall magnitude is also very variable m) of all trees of the study species in the transect were and there is a significant correlation between discharge measured as well as their elevation (using an auto-level and rainfall (r = 0.632, p = 0.01). A very close cor and staff) relative to the river dry season base flow relation would not be expected between discharge and level and distance from river base flow. In addition, re rainfall as isolated intense rainfall events in other parts productive status of each tree was recorded. Sites were of the catchment may affect discharge at the site of the revisited regularly to record reproductive phenology. stream gauging station where data were collected. One m square seedling quadrats were randomly In contrast, on the Blackwood River over a 37 year laid out adjacent to the transects and on two addi period maximum discharge consistently occurs in the tional transects where seedlings occurred. Within each winter in the wet part of the year between July and quadrat, heights of all 1 year old seedlings were re August. Peak flows are regular events each year occur corded. For each quadrat a number of environmental ring with the regular winter rains. The one exception wetlsm02.tex; 29/02/2000; 20:52; p.4 5 (a) Ord River 15000 24455 400 300 200 100 0~~~~~"-~~~~~~~~~~~~~~~ (b) Blackwood River 250 0 Year Figure 2. Monthly river discharge and monthly rainfhll data over a 18 year period for the (a) Ord River and the (b) Blackwood River. was an extremely large flood event in January, 1982 charge on the Ord River. Average monthly discharge caused by a cyclonic feature drifting down from the shows a peak in July with river levels usually receding tropics in to the south-west land division (Figure 2a). by September. Rainfall patterns in the lower Black Maximum discharge levels are much smaller and less wood show a regular and distinct winter (June-August) variable than those seen on the Ord River with a mean maximum but with summer rainfall events of> 10 mm maximum discharge of 161 m3/s ± 24.5. Even the quite common (Figure 2b). There was a strong cor exceptional event in January, 1982 with a peak flow of relation between maximum discharge and rainfall for 1187 m3/s is much smaller than average maximum dis- the Blackwood river (r = 0.741, p = 0.01) if the out- wetlsm02.tex; 29/02/2000; 20:52; p.S 6 (a) Ord River Melalauca leucodendra Seed!all Flower development Fl . Fruit development ~--~<(· .. ··)> ~--~---O~!JIJ.II..~~-~ Eucalyptus camaldulensis Fruit development Fruit development 5000 ~ ~ Flower daveloJll:llent Flo . <111.. 11------'S...,e;we.wdlllaJal~~..l --1..-~ ...... ~..... ~ -<4 wenng ... +---4 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Oec (b) Blackwood River Melaleuca rhaphiophylla Fruit development Flowar development ~ ~.-.: .-.: ~ :.~ ••••••••••••• •~E!ey Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Fig11re 3. Relationship between phenology of selected riparian species with montly river flow levels for the (a) Ord River and the (b) Blackwood River. lier event in January is removed. This is because the and compared with mean monthly river discharge greatest amount of rainfall fell in the upper part of the (Figures 3 a and b). For both Eucalyptus species seed catchment. fall appears to be timed to coincide with that period of the year when river water levels are receding. That is, Tree reproductive phenology March, April for E. camaldulensis and September to November for E. rudis. For E. camaldulensis flower Seasonal reproductive phenology for Eucalyptus ca ing and seed fall occur in a single annual cycle with maldulensis andMelaleuca leucadendra on the Ord fruit maturing and releasing seed with little seed ap River and E. rudis and M. rhaphiophylla on the Black pearing to remain in the canopy. In contrast E. rudis wood river has been summarized for a 12 month period flowers from August to November with fruit and seed wetlsm02.tex; 29/02/2000; 20:52; p.6 7 developing over summer. Greatest seed release is not higher in loam, loamy sand and sandy soils than in from these matured capsules until the following early the clay soil and no plots occurred in gravel soil (H = summer period. Seed is therefore retained on the tree 10.99, p = 0.011). Frequency of plots for microto so that some seed may be released throughout the year pography showed more than 50% of plots occurring with a peak in the early summer period. on the river flat area but analysis of seedling density A similar contrast is seen for the Melaleuca speeies showed river flat, depression and rise had signific~ with M. leucadendra on the Ord River releasing all its antly greater seedlings/m2 than plots on a slope (H = seed from February to April, the period after flower 11.66, p =0.0086). Moist sites had the greatest fre ing and fruit development from August to January the quency of seedling plots with wet and moist sites previous year. M. rhaphiophylla on the Blackwood having a significantly greater density of seedlings than River flowers in September to November. Fruit de did dry sites (H = 10.47, p =0.0053). There was a velopment occurs over summer and autumn and the significant correlation between seedling density and woody fruit and seed are retained on the plant for sev distance to nearest seed tree (r =0.410, p = 0.05) eral years. Seed release appears to be opportunistic with most plots in close proximity to a seed tree and after the desiccation of the fruit. no plots > 30 m from a seed tree. There was also no significant correlation between seedling density and% Seedling establishment litter or% bare ground. Like the Ord River plots, fre quency of plots with 0-20% litter and 0-20% bare In the Ord River seedling plots the majority had loam ground was much greater than the other cover classes to sandy soils (Figure 4a). When seedling density (Figure 4b). There was a strong negative relationship was compared for the different soil types, numbers of between seedling density and elevation (r = 0.758, seedlings/m2 was significantly higher in loam, loamy p =0.001) with most seedlings occurring at lower sand and sandy soils than in the clay and gravel soil elevations (Figure Sb). There was also a significant types (H = 15.87, p = 0.003). Frequency of plots was relationship between elevation and distance from the fairly similar for the different microtopography sites river (r =0.815, p =0.001). with slightly more plots occurring on the river flat area but analysis of seedling density showed that plots in depressions, slopes and rises had significantly more Size class distribution seedlings/m2 (H = 13.28, p =0.004). Moist and dry sites had the greatest frequency of seedling plots but A profile diagram of an Ord river transect shows waterlogged sites had significantly greater density of abundance of new ( < 1 year old) seedlings at the lower seedlings than the other moisture-related parameters elevations on the river flats and an abundance of smal (H =9.74, p =0.02). There was no significant correl ler size classes throughout the profile (Figure 6a). ation between seedling density and distance to nearest Frequency of larger size classes is greatest at the seed tree with a large group of plots in close prox higher elevations and greater distance from the river, imity and another group at some distance (>200 m) especially for E. camaldulensis. All M. leucadendra from a seed tree. There was also no significant cor size classes tend to be at the lower elevations closer relation between seedling density and litter or bare to the rivers edge. There was a significant correlation ground. Frequency of plots with 0-20% litter and 0- between diameter at breast height and elevation for 20% bare ground was much greater than the other both E. camaldulensis and M. leucadendra (Figure 7 a cover classes (Figure 4a). There was a strong negative and b). For E. rudis on the Blackwood River the profile relationship between seedling density and elevation diagram shows seedlings most abundant at the lower (r = 0.875, p =0.001) with most seedlings occurring at elevation near the river edge but large diameter trees lower elevations (Figure Sa). There was, however, not are also most abundant at these lower elevations. Size a strong relationship between elevation and distance classes were generally more evenly spread along the from the river (r =0.284, p =0.10). This reflects the profile (Figure 6b). M. rhaphiophylla tends to be con wide terrace profile of the river at the Ord River sites. fined to areas close to the water which are regularly The majority of Blackwood River seedling plots inundated. In this area there is a wide range of size were loamy sand and sandy soils (Figure 4b ). When classes. For both E. rudis and M. rhaphiophylla there seedling density were compared for the different was no significant relationship between size class and soil types, numbers of seedlings/m2 was significantly elevation (Figure 7 c and d). wetlsm02.tex; 29/02/2000; 20:52; p.7 8 (a) Ord River 100 90 80 70 !! 60 Cl ii: 50 ...Cl 10 0 Micro- Soil Type topography Moisture %bare %litter (b) Blackwood River 100 90 80 70 Ill 0 60 i'i: .... 50 Cl 40 20 10 0 "CC'tl ~~E <:: <:: w ... .. 0 .. .. --::..-.. .,.. .. !!! '1:1 eJ <:: ~ ~ .!:!.. Micro- Soil Type topography Moisture %bare % litter Figure 4. Percentage of I m x 1 m seedling plots with particular environmental attributes for (a) the Ord River and (b) the Blackwood River. Discussion mainly generated from sub-tropical monsoons. These are confined almost entirely to the wet season between October and March and contribute high levels of run This study highlights the vastly different hydrological off into the river. The Blackwood River, although it regimes of the Ord and Blackwood rivers despite sim has a strong winter (June-August) rainfall maximum, ilar rainfall and catchment size. This is mainly due to events are less intense and run-off not as high. the different geology of the catchments (Department Fluvial processes on the Ord River are domin of Mines, 1972) and rainfall patterns. The Ord River ated by fairly frequent large high energy floods that catchment is subject to high intensity rainfall events wetlsm02.tex; 29/02/2000; 20:52; p.B 9 (a) Ord River 30 N y = 9.574 * 10 -O.OB!x r= 0.791 E 25 [J ri .... [J (I) 20 1:11 .!: ;:; 15 ID [J [J ID (I) 10 d z 5 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Elevation (m) (b) Blackwood River 50 y = 22.357 * 10 -O.ISOx r = 0.755 1: 40 [J Ill 1:11 .!: ;:; 20 ill 5I d 10 z 0 1 2 3 4 5 6 7 11 9 10 11 12 13 14 15 Elevation (m) Figure 5. Relationship between elevation from the river and number of seedlings in I m x I m plots for(a) the OrdRiver and (b) the Biackwood River. are capable of destroying the existing riparian vegeta isting trees and the recruitment of younger age classes. tion community and preventing the establishment of This regime of intermittent high frequency large dis stable mature stands. This is evidenced by the ex turbances prevents the establishment of stable states istence of large mature trees on these river reaches and in terms of the st;te and transition model the eco only at high elevations, in protected river bends or far system is characterized by long periods of transition from the river where only extreme floods can reach. between short-lived stable states. This riparian ecosys In the years between large flood events large stands tem is thus driven by allogenic or physical processes of seedlings and saplings develop but these generally rather than vegetation successional (autogenic) pro do not reach maturity before they are destroyed in the cesses (Baker and Walford, 1995). In contrast, lower next large flood. Failure to reach later successional energy seasonal flooding on the Blackwood River al stage due to high frequency flood disturbance has lows mature stands of trees to develop throughout the been documented for New Zealand Podocarp forest river profile. Recruitment is continual although species (Duncan, 1993). In this way single catastrophic events can also respond to large flood events. This disturb change the ecosystem to an extent which is inconsist ance regime results in long periods of stable states with ent with the theory of climax communities (George et short periods of transition. Therefore the vegetation al., 1992). These high magnitude wet season floods is subjected to longer periods of autogenic processes are, however, highly episodic and in the intervening and because of lower frequency flooding disturbance, years smaller floods allow the development of the ex- shorter periods of allogenic processes. wetlsm02.tex; 29/02/2000; 20:52; p.9 10 Ill Eucalyptus camaldulensis --- elevation Ill Melalouca leucadondra a) Ord River site a-- 184MI 42MI 7-S. 39 Ec 6 1: 5 0 4":;:: 3 Ill 2 i; 1 ii:i 0 1! 11 1}\ I I 1 IIJ\ '''''/\. •••••A '''''J\ •••• '1\ ~~ ~ ('~ """""""'.....,....,. N~ 20-30m 1 0·20m 0·1 Om 0-10m 10·20m 20-30m 40-50m 7 o-aom (b) Blackwood River site Ill Eucalyptus rudls --- elevalion Ill Melaleuca C'>IOOQC::U::lO v,...NM'Q'lrun tIt I fJ\ J I I I IJ\ I I I 1 IJ\ It I l fl\ I I It I of\ It I I I}\ NOOOO NOOOO NOOOO .-(';IM'I;f ('od~~~~ rC'\IM" Figure 6. Size class frequency and disttibution of the two main overston:y species along a !rdllSVerse profile at a) ihe Ord River and b) ihe Blackwood River. Histograms show tree size classes (dbh(cm)) by species for each plot. For both rivers the environmental conditions that fact, for the Ord River successful long term establish appear to correlate with seedling establishment are ment to a stable state may require a period of no large similar, with most seedlings occurring in river sed floods. On the Blackwood River seedling density was iment of larger particle size of sandy to loam soils. closely related to distance to seed trees and no seed Seedlings are also more dense at lower elevations and lings were found further than 30 m from a potential in microtopographic sites where moisture is gaining seed tree. For E. camaldulensis on the Murray River and there is a slight rise. These rises are important in south-eastern Australia a similar close association for seedling establishment by creating a safe site to between seedlings and seed trees was found (Dexter prevent waterlogging (Bren and Gibbs, 1986; Sko et al., 1986). This suggests that seeds are generally glund and Verwijst, 1989; Titus, 1990). This once dispersed by wind with little dispersal on floodwaters. again highlights the importance of fluvial processes Gerinination tests on seeds of E. rudis indicate that in recruitment. Litter levels and areas of bare sedi the seed germinate rapidly after wetting (1 to 2 days) ment do not appear to be important for establishment so that they may not have time to be dispersed by of seedlings. These elements may however become water. Seed or fruit of E. rudis was not found in any important for the successful initial establishment of debris samples. However, all species appear capable germinating seed (Nilsson and Gunnel, 1990; Gre of responding to floods with large recruitment events enway, 1994; Johnson, 1994). Seedling plots on the but the mechanisms for doing so appear different. For Ord River can occur at large distances from mature example there was a large recruitment event of E. rudis trees and therefore from a source of seed. This may after the unseasonal extreme flood in January, 1982. indicate that seed can be dispersed by flood waters. Roberts (1993) found a good correlation between re Observations of flood debris lines indicate that seed cruitment and flood events, which replenish soil water is being transported in this way as evidenced by the for E. coolibah on a semi-arid ephemeral creek in presence of numerous open fruit and newly germin eastern Australia. ated seed. Thus seed can be transported to areas which Reproductive phenology of the four species stud afford some protection from subsequent flooding. In ied also reflects their adaptation to their environment. wetlsm02.tex; 29/02/.2000; 20:52; p.lO 11 (a) E.· camaldulensis (b) M. leucodendra 80 y = 3.911x + 0.675 r=0.606 50 y =3.829x + 0.695 r= 0.560 lit Ill lit E!l 40 'E 60 ~ 30 ::1: 11> m 40 c 20 lit 20 10 ' 5 10 15 1 2 3 4 5 6 Elevation (m) El13vation (m) (c) E. rudis (d) M. rhaphiophylla 150 y = -1.220x + 49.21 r=0.060 60 y = "2.469x + 21.82 r=0.255 • 11> • 50 .. <11 01> ...... ~100 • • 40 • :I: fill lit .. m "' 30 c 50 - "' - 20 •• 0 0 1 2 3 4 5 6 7 0 1 2 3 Elevation (m) Elevation (m) Figure 7. Relationship between tree size (diameter at breast height) and elevation from the river for the 4 selected riparian species on the Ord River (a & b) and the Blackwood River (c & d). Maturation of flowers and seed takes place within a unpublished data). Therefore post-dispersal conditions couple of months in many tropical eucalypts (Brooker may limit recruitment and if subsequent conditions are and Kleinig, 1983; Eldridge et al., 1993) as opposed not favorable the community will remain in transition. to more than two years in some southern eucalypts E. rudis on the Blackwood River retains mature seed (Ashton, 1975). For E. camaldulensis, M. leucadendra in the canopy for at least 12 months. Although there is and E. rudis the period of greatest seedfall is timed to a time of peak seed release, some seed is released all coincide with receding floodwaters when moist sed year round so that there is potential for germination of iments favorable for seed germination and survival seed whenever conditions are favorable thus providing are exposed. This adaptation has been reported for continuos recruitment into the stable community. For riparian species of Populus and Salix in the U.S. (Brad many temperate eucalypts seed maturation can take ley and Smith, 1986; Johnson, 1994). Indeed pheno several months and seeds are retained in the capsules logical events in many species are thought to be timed with a continuos low level of seed release occurring to abiotic conditions (Estabrook et al., 1982). The Ord which is accelerated by dry conditions (Cunningham, river species release the majority of their seed over a 1957; Wellington and Noble, 1985; Davies and Myer short period of time at the end of the wet season which scough, 1991). Interestingly, reproductive phenology provides their only opportunity for annual recruitment. of E. camaldulensis occurring on the Murray River in Other northern Australian Eucalyptus spp. have been south-eastern Australia is similar to that documented reported to display a similar reproductive phenology here for E. rudis on the Blackwood River (Dexter et (Bowman et al., 1991; Fensham, 1992; Setterfield and al., 1986). These rivers have similar fluvial regimes Williams, 1996). The small seed of these species ·do and climatic conditions. M. rhaphiophylla retains ma not survive long in the soil and therefore do not be ture fruit on the tree and appears to only release seed come incorporated into the soil seedbank (N. Pettit, with the desiccation of the woody fruit. This can wetlsm02.tex; 29/02/2000; 20:52; p.11 12 happen through disturbance such as flooding where Baker, W.L. 1988. Size class structure of contiguous riparian wood branches are commonly broken off or during strong lands along a Rocky Mountain river. Physical Geography 9: winds. This may result in the release of seed when 1-14. Baker, W.L. 1990. Climatic and hydrologic effects on the regen conditions are most favorable for successful germina eration of Populus angustifolia James along the Anirnas River, tion such as when flood waters are receding. Fire may Colorado. 1. Biogeography 17: 59-73. also result in the release of seed (Froend et al., 1993). Baker, W.L: and Walford, G.M. 1995. Multiple stable states and These results highlight the very different fluvial re models of riparian vegetation succession on the Anirnas River, Colorado. Annals of the Association of American Geographers gimes that exist on these two rivers and· their effect 85: 320-338. on the vegetation dynamics for the overstory species. Bames, W.J. 1985. Population dynamics of woody plants on a river This is despite the species compared on the two rivers island. Canadian J. Bot 63: 647-655. Beard, J. 1979. The vegetation of the Kimberley area, Western Aus being functionally very similar. Management of the tralia: 1: 1,000,000 series. Perth, W.A., University of Western riparian vegetation should therefore take into account Australia Press. 118 pp. the frequency of change in the vegetation and that the Bell, D.T., Hopkins, A.J.M. and Pate, J.S. 1982. Fire in the disturbed states and long periods of transition between Kwongan. ln: Pate, J.S. and Beard, J.S. (eds.), Kwongan: Plant Life of the Sandplain. pp. 178-204. University of Western states are part of the natural process. This would sug Australia Press, Perth Western Australia. gest that altering the natural flow regimes, such as Bell, D.T. and Koch, J.M. 1980. Post-fire succession in the jarrah through river regulation will have significant effects forest of Western Australia. Australian J. Ecol. 5: 9-14. on riparian vegetation dynamics. This is particularly Boland, D.J., Brooker, M.I.H., Chippendale, G.M., Hall, N., Hy land, B.P.M., Johnston, R.D., Kleinig, D.A. and Turner, J.D. so for the effects on long term recruitment in terms of 1989. Forest Trees of Australia. Thomas Nelson Australia and reproductive phenology and seedling survival (Bren, CSffiO, Melbourne. 687 pp. 1992; Johnson, 1994). In the case of regulation on Boutin, C. and Keddy, P.A. 1993. A Functional Classification of high disturbance rivers such as the Ord River, riparian Wetland Plants. J. Vegetation Science, 4: 591-600. Bowman, D.M.J.S., Wilson, B.A. and Woinarski, J.C.Z. 1991. Flor vegetation dynamics may be significantly altered to istic and Phenological variation in a northern Austalian rocky long periods of stable states with much reduced trans savanna. Proceedings of the Royal Society of Queensland 101: itions. This has important implications for the long 79-90. Bradley, C. E. and Smith, D.G. 1986. Plains cottonwood recruitment term structure and functioning of these communities. and survival on a prairie meandering river floodplain, Milk River, southern Alberta and northern Montana. Canadian J. Bot. 64: 1433-1442. Acknowledgements Bren, LJ. 1992. Tree invasion of an intermittent wetland in relation to changes in flooding frequency of the River Murray, Australia. Australian J. Ecol. 17: 395-408. The authors thank R. Doupe and M. Angelorii for their Bren, L.J. and Gibbs, N.L. 1986. Relationships between flood fre invaluable assistance in the field and knowledge of the quency, vegetation and topography in a river red gum forest. Australian Forestry Research 16: 357-370. Ord River, the staff at the Water and Rivers Commis Brooker, M.I.H. and Kleinig, D.A. 1983. Field Guide to the sion in Kununurra for the use of field equipment and a Eucalypts: Northern Australia. Inkata Press, Sydney. 383 pp. vehicle, and in Perth for the supply of historical flow Bunn, S.E. 1993. Riparian-stream linkages: research needs for the data, Dr L. Penn (WRC) for comments on a draft of protection of in-stream values. Australian Biology 6:46-51. CEPA. 1992. Towards healthier rivers. Commonwealth Environ the article, and A. Karafilis for her support and assist mental Protection Authority. Report 21. Australian Common ance on the Blackwood River. The project was funded wealth Government, Canberra. 21 pp. by the Land and Water Resources Research and De Cunningham, T.M. 1957. Seed production and seed fall of Euca velopment Corporation through the National Riparian lyptus reg11ans (F. Muell.). Australian Forestry 21:30-39. Davies, S.J. and Myerscough, P.J. 1991. Post-fire demography of Program and the Water and Rivers Commission of the wet mallee Eucalyptus luehmannia F. Muell. Australian J. Western Australia. Bot. 39: 459-466. Department of Mines. 1972. Geological Survey of Western Aus tralia: Annual Report for the year, 1971. Extract from the report of the Department of Mines. Perth, Western Australia. References Dexter, B.O. 1978. Silviculture of the river red gum forests of the central Murray flood plain. Royal Society of Victoria Proceed Abido, M.S. 1985. The use of riparian vegetation in the Damascus ings 90: 175-191. basin. Riparian Ecosystems and their Management: Reconciling Dexter, B.O., Rose, H.J. and Davies, N. 1986. River regulation and Conflicting Uses, Tucson, Arizona, North American Riparian associated forest management problems in the River Murray Red Conference. pp. 467-469. Gum forests. Australian Forestry 49: 16-27. Ashton, D.H. 1975. Studies of flowering behaviour in Eucalyptus regnans F. Muell. Australian. J. Bot. 23: 399-411. wetlsm02.tex; 29/02/2000; 20:52; p.12 13 Duncan, R.P. 1993. Rood Disturbance and the Coexistence of Mclntrye, S., Lavorel, S. and Tremont, R.M. 1995. Plant life Species in a Lowland Podocarp Forest, South Westland, New history attributes: their relationship to disturbance response in Zealand. J. Ecol. 81: 403-416. herbaceous vegetation. J. Ecol. 83: 31-44. Eldridge, K., Davidson, J., Harwood, C. and van Wyk, G. 1993. Nilsson, C., Ekblad, A., Dynesius, M., Backe, S., Gardfjell, M., Eucalypt domestication and breeding., Oxford University Press, Carlberg, B., Hellqvist, S. and Jansson, R. 1994. A comparison Oxford. 265 pp. of species richness and traits of riparian plants between a main Estabrook, G.F., Winsor, J.A., Stephenson,- A.G. and Howe, H.F. river channel and its tributaries. J. Ecol. 82: 281-295. 1982. When are two phenological patterns different? Botanical Nilsson, C. and Gmmel, G. 1990. The effects of litter displacement Gazette 143: 374-378. on riverbank vegetation. Canadian J. Bot. 68: 735-741. Fensham, R.J. 1992. The establishment of eucalyptus seedlings in Olsen, G. and Skitrnore, E. 1991. State of the rivers of the south west tropical savanna forest. Northern Territory Naturalist 13: 30-36. drainage division. Western Australian Water Resources Council. Fitzgemld, K. 1967. The Ord River catchment regeneration pro 171 pp. ject. West·Australian Department of Agriculture. Perth Western Roberts, J. 1993. Regeneration and growth ofCoolibah, Eucalyptus Australia. Bulletin 3599. 24 pp. coolabah Subsp. Arida, a riparian tree, in the Cooper Creek Fleischner, T,L. 1994. Ecological Costs of Livestock Grazing in region of South Australia. Australian J. Ecol. 18: 345-350. Western l'jorth America. Conservation Biology 8: 629-644. Schofield, N.J., Ruprecht, J.K. and Lob, I.C. 1988. The impact Froend, R.H., Farrell, R.C.C., Wilkins, C.F. and McCpmb, A.J. of agricultural development on the salinity of surface water re 1993. Wetl~ds of the Swan Coastal Plain: The Effect of Altered sources of south-west Western Austmlia. Water Authority of Water Regimes on Wetland Plants. Water Authority of Western Western Australia. 83 pp. Australia, Perth Western Australia. 178 pp. Setterfield, S.A. and Williarns, R.J. 1996. Patterns of flowering and George, M.R., Brown, J.R. and Clawson, W.J. 1992. Application seed production in Eucalyptus miniata and E. tetrodonta in a of non-eqnilibriuni ecology to management of Mediterranean tropical savanna woodland, northern Australia. Australian J. Bot. grassland. J. Range Management 45: 436--440. 44: 107-122. Glenn-Lewin, D.C. and van der Maarel, E. 1992. Patterns and pro Shah, B.H. and Thames, J.L. 1985. Role of riparian vegetation in cesses of vegetation dynamics. Plant Succession: Theory and Pakistan. Riparian Ecosystems and their Management: Reconcil Prediction. Cambridge University Press, Cambridge. pp. 11-58. ing Conflicting Uses, Tucson, Arizona, North American Riparian Greenway, M. 1994. Litter accession and accumulation in a Conference. pp. 482-484. Melaleuca quinquemervia wetland in south-eastern Queensland. Skoglund, J. and Verwijst, T. 1989. Age structure of woody spe Austmlian J. Marine and Freshwater Research 45: 1509-1519. cies populations in relation to seed rain, germination and es Grime, J.P. and Hillier, S.H. 1992. The contribution of seedling tablishment along the river Dalalven, Sweden. Vegetatio 82: regeneration to the structure and dynamics of plant communities 25-34. and larger units of landscape. Seeds: The ecology of regeneration Titus, J.H. 1990. Microtopography and woody plant regeneration in in plant communities. Wallingford UK, CAB International. pp. a hardwood floodplain swamp in Florida. Bull. Torrey Botanical 349-364. Club 117: 429--437. Hills, J.M., Murphy, K.J., Pulford, I.D. and Flowers, T.H. 1994. van der Valk, A.G. 1981. Succession in wetlands: A Gleasouian A method of classifying European riverine wetland ecosystems Approach. Ecol. 62: 688-696. using functional vegetation groups. Functional Ecol. 8: 242-252. Wellington, A.B. and Noble, I.R. 1985. Seed dynamics and factors Hobbs, R.J. 1994. Dynamics of vegetation mosaics: Can we predict limiting recruitment of the mallee Eucalyptus incrassata in semi responses to global change? Ecoscience 1: 346-356. arid, south eastern Australia. J. Ecol. 73: 657-666. Jeffries, M. and Mills, D. 1990. Freshwater Ecology: piinciples and Westoby, M., Walker, B. and Noy-Meir, I. 1989. Opportunistic man applications. Bellhaven Press, London. 285 pp. agement for rangelands not at equilibrium. J. Range Managmt. Johnson, W.C. 1994. Woodland expansion in the Platte River, Neb 42: 266-274. raska: Patterns and causes. Ecological Monographs 64: 45-84. wetlsrn02.tex; 29/02_/2000; 20:52; p.13