Mapping Terrestrial Groundwater Dependent Ecosystems:

Method Development and Example Output

P. Evan Dresel Rob Clark Xiang Cheng Mark Reid Alister Terry Jonathon Fawcett Denise Cochrane

If you would like to receive this information/publication in an accessible format (such as large print or audio) please call the Customer Service Centre on 136 186, TTY 1800 122 969, or email [email protected].

Citation:

Dresel, P. E., Clark, R. Cheng, X., Reid, M., Fawcett, J., and Cochraine, D. (2010) Mapping Terrestrial Groundwater Dependent Ecosystems: Method Development and Example Output. Victoria Department of Primary Industries, Melbourne VIC. 66 pp.

ISBN 978-1-74264-321-2 (print) ISBN 978-1-74264-322-9 (CD-ROM) ISBN 978-1-74264-323-6 (online)

Published by the Department of Primary Industries , April 2010

© The State of Victoria 2010.

This publication is copyright. No part may be reproduced by any process except in accordance with the provisions of the Copyright Act 1968.

Authorised by the Department of Primary Industries 1 Spring Street, Melbourne 3000.

Disclaimer This publication may be of assistance to you but the State of Victoria and its employees do not guarantee that the publication is without flaw of any kind or is wholly appropriate for your particular purposes and therefore disclaims all liability for any error, loss or other consequence which may arise from you relying on any information in this publication.

For more information about DPI go to www.dpi.vic.gov.au or phone the Customer Service Centre on 136 186.

Contents

Figures 4

Tables 5

Acronyms and Definitions 6

1 Introduction 9

1.1 Project Objectives 9

2 What are Groundwater Dependent Ecosystems? 11

2.1 Original Classification System from Hatton & Evans (1998) 11

2.2 Additional Classes of GDEs 11

2.3 Alternative GDE Classification System by Eamus et al. (2006) 12

2.4 ADE classification & identification by Colvin et al. (2007) 12

3 Description of GDE Classes 14

3.1 Terrestrial vegetation 14

3.2 Wetlands - fresh and saline groundwater dependency 15

3.3 Coastal estuarine and near shore marine systems 17

3.4 River base flow systems 18

3.5 Aquifer and cave ecosystems 18

3.6 Terrestrial fauna 18

4 Known Victorian areas with GDEs 20

5 Identification of GDEs 21

5.1 Vegetation analysis 21

5.2 Groundwater – surface water interaction 22

5.3 Remote sensing approaches 23

5.4 Geological mapping approaches 23

6 Development of state-wide potential GDE maps 25

6.1 Spatial data sources for GDE mapping 25

6.2 Definition of Aridity Zones 27

6.3 Remote sensing methodology 28

6.3.1 Contrasting mid-summer photosynthetic activity 28

Mapping Terrestrial Groundwater Dependent Ecosystems 3

6.3.2 Consistent annual photosynthetic activity 29

6.3.3 Unsupervised classification of Landsat spectral data 33

6.4 Integration of data sets and development of Model layers 33

6.5 Classification and Attribution of GDEs 37

6.5.1 Groundwater Interactive Map 38

6.5.2 Geomorphologic, Geologic, and Chemical Attributes 39

7 CMA region potential GDE map results 42

7.1 Mallee CMA 42

7.2 Wimmera CMA 42

7.3 Glenelg Hopkins CMA 42

7.4 North Central CMA 46

7.5 Corangamite CMA 48

7.6 Port Phillip and Westernport CMA 50

7.7 Goulburn Broken CMA 50

7.8 West Gippsland CMA 50

7.9 North East CMA 50

7.10 East Gippsland CMA 55

8 Conclusions and Recommendations 58

8.1 Limitations of the maps 58

8.2 Recommended usage of the maps in current form 59

9 References 61

Figures

Figure 1. Distrubution of wetlands where previous research has established a link with groundwater and location of stream gauges where base flow calculations have been made 20 Figure 2. Catchment Management Authority (CMA) regions for Victoria 25 Figure 3. Climate zones for the Corangamite CMA region derived from aridity indices (A) and comparing the climate zones to the MODIS EVI time series data for summer 2003 (B) 27 Figure 4. Process flow for development of potential terrestrial GDE map layer 28 Figure 5. Mid summer NDVI for the combined years of 1995 and 2002 in the North Central CMA region 30 Figure 6. Time series of MODIS EVI values showing consistent annual photosynthetic activity of redgum forest in comparison to variable annual photosynthetic activity of pasture 31 Figure 7. EVI standard deviation layer for the North Central CMA region, showing the 27 classes with the lowest standard deviation 32 Figure 8. Distribution of selected EVC groupings used in the unsupervised classification in Victoria. Black outline defines the Glenelg-Hopkins CMA region. 34 Figure 9. Application of land use mask to potential terrestrial GDEs. A. Land use classification for section of North East CMA region. B. Potential terrestrial GDEs before and after application of mask. 36 Figure 10. Indicative groundwater interactive map (GIM) for the North Central CMA region showing the location of piezometers that have recorded a watertable depth of less than 5 m. Colour coding shows five different percentile

4 Mapping Terrestrial Groundwater Dependent Ecosystems

ranges indicating percentage of piezometer record for which the water table was less than 5 m below ground surface 40 Figure 11. Potential terrestrial GDEs in the Mallee CMA. A. Overview. B. GDEs mapped in area of extensive mallee vegetation. C. Sparse association of mapped GDEs with lakes, salinas and areas of shallow water table 43 Figure 12. Potential terrestrial GDEs in the Wimmera CMA. A. Overview. B. Area of known groundwater-surface water interaction near Edenhope. C. Area of potential GDEs near the Wimmera River at Dimboola 44 Figure 13. Potential terrestrial GDEs in the Glenelg Hopkins CMA. A. Overview. B. Detail of GDEs shown around the southern Grampians. C. Potential GDEs and spring locations northwest of Portland. 45 Figure 14. Map of the Cockajemmi Lakes region showing potential GDEs and the extent of the Quaternary age basalt flow 46 Figure 15. Potential GDE map for the North Central CMA region. A. Overview of the CMA region. B. Area between Swan Hill and Kerang showing the association of potential GDEs in areas where surface water drainages or wetlands intersect groundwater interactive landscapes. C. Relation of GDEs to spring locations in the vicinity of Daylesford. 47 Figure 16. Map of potential GDEs for the Tarpaulin Creek area, showing landscapes that may contain GDEs with shallow rooted vegetation (cyan) and landscapes which might include deep rooted vegetation and fall outside the indicative groundwater interaction map (yellow) 48 Figure 17. Potential terrestrial GDEs in the Corangamite CMA. A. Overview. B. Detail showing areas of potential groundwater dependency east and south of Ballarat. C. Potential terrestrial GDEs around lakes near Colac, including Lake Corangamite 49 Figure 18. Potential terrestrial GDEs in the Port Phillip and Westernport CMA. A. Overview. B. Detail showing areas of potential groundwater dependency along streams near Bacchus Marsh. C. Potential GDEs in the Yarra River riparian zone near Healesville 51 Figure 19. Satellite image showing potential terrestrial GDEs along Yarra River 52 Figure 20. Potential terrestrial GDEs in the Goulburn Broken CMA. A. Overview. B. Detail of GDEs in the Murray Riverfloodplain (large green area is the Barmah Forest), along the lower Goulburn River (lower left) and along other streams including the Broken Creek 53 Figure 21. Potential terrestrial GDEs in the West Gippsland CMA. A. Overview. B&C. Detail showing areas of potential groundwater dependency along streams and waterways in the Latrobe Valley (B) and north of Lake Wellington (C) 54 Figure 22. Potential terrestrial GDEs in the North East CMA. A. Overview. B. Detail showing areas of potential groundwater dependency along the Murray and Ovens River floodplains 55 Figure 23. Potential terrestrial GDEs in the East Gippsland CMA. A. Overview. B. Detail showing potential groundwater dependency in areas surrounding drainage lines and water bodies and other scattered areas. The large blocks are responding to open water 57 Figure 24. Potential terrestrial GDEs in a section of the North East CMA region, classified by groundwater interactive landscape. The areas of greatest interest are expected to be in the shallow water table zones 60

Tables

Table 1 Some references of some research into groundwater-surface water interaction and groundwater use by vegetation within Victorian and adjacent landscapes 20 Table 2. Tools to identify GDE's (after Clifton et al. 2007) 22 Table 3. EVC groups and examples used to develop the indicative response layer 33 Table 4. Example weighting factors for use in model combining remote sensing GDE classes 35 Table 5. Land use classes used to exclude areas from classification as potential terrestrial GDEs 37 Table 6. Attributes assigned to each potential GDE location 41

Publication title 5

Acronyms and Definitions

AGO Australian Greenhouse Office

CGDL State of Victoria Corporate Geospatial Data Library

CMA Catchment Management Authority

EVC Ecological Vegetation Class

EVI Enhanced Vegetation Index

GDE Groundwater Dependent Ecosystem

GFS Groundwater Flow System

GIM Groundwater Interactive Map

GMU Geomorphologic Management Unit

NDVI Normalised Difference Vegetation Index

6 Mapping Terrestrial Groundwater Dependent Ecosystems

1 Introduction

Groundwater and surface water systems are commonly intimately linked and, therefore, understanding how they respond to climate change and land and water management is fundamental to protecting groundwater dependent ecosystems (GDEs). It is well understood that land use practices and extraction of groundwater resources can affect groundwater levels, flows, and quality. However, there is less understanding about the implications of such effects on the health of above- and below-ground ecosystems that have a dependence on groundwater. Trying to improve this understanding together with knowledge of GDE occurrence, significance, sensitivity and threat has only recently become a National priority.

Possible threats to GDEs include groundwater extraction, climate change, salinity, nutrients, and altered surface water management. An outcome of the 1994 Council of Australian Governments (COAG) Water Reform Framework (Council of Australian Governments 1994) was that water allocation planning is required to protect ecosystems, including GDEs that have an important function or conservation value. The need to improve knowledge of GDEs across Victoria has been made more urgent due to prolonged dry conditions since 1997 and an increased reliance upon groundwater as a resource. The last National Land and Water Audit in Australia estimated that groundwater use had increased by 88% from 1983 to 1997 (Department of the Environment, Water, Heritage and the Arts 2006). Anecdotal evidence, including increased groundwater licensing and exploitations of new or unmanaged aquifers since 1997, suggest that the general increase in groundwater use continues to the present day.

GDEs can include subterranean fauna (stygofauna), wetlands, streams, and lakes, remnant pools of ephemeral streams, and terrestrial vegetation with roots tapping the aquifer (phreatophytes). Groundwater can maintain stream flow during periods of low precipitation, provide nutrients for aquatic biota, and provide thermally stable refuges for fauna. The identification of communities that use groundwater will be used as an indication of potential dependence in this review. Determining dependence may require field experimentation and verification.

In addition to the intrinsic importance of maintaining the biodiversity of GDEs, they provide important resources for rural lifestyle. GDEs support fishing, waterfowl hunting, bird watching, and other recreation. The locations of GDEs provide valuable insight to the hydrologic system within catchments and understanding GDEs is important to water resource, environmental, and agricultural management.

This report is the major milestone and culmination of DPI CMI project No 102552, “Groundwater dependent ecosystems”. It began in 2007 and was funded initially from the National Action Plan (NAP) State Reserve and later by the State Water Fund. Project governance has been shared between the Victorian NAP Office and DSE Water Allocation and Licensing Branch. The project addresses for Victoria the first step in developing informed management action plans for protection or enhancement of GDEs, namely the prediction of GDE occurrence across the State. The issue addressed by this project is rated very highly by the National Water Initiative (Raising National Water Standards), the Victorian White Paper for Water (Securing Our Water Future), the Murray-Darling Basin Authority, the Co-operative Research Centre for eWater and the NAP. The persistence of dry conditions over a number of years further reinforces the high priority attention to GDEs.

The report includes a detailed literature review on GDEs and describes a new integrated method that has been developed by the authors and applied to produce the first ever series of regional potential GDE maps for Victoria. To the knowledge of the authors, this is the first time that mapping of GDEs has been attempted at a large scale in Australia and represents one of very few attempts internationally. The strengths and limitations of the currently produced maps are explained in Section 8 together with recommendations for further validation and refinement of the maps.

1.1 Project Objectives

The initially recorded project objectives were to: (i) identify threatened high-value GDEs in Victorian NAP regions (i.e. Goulburn-Broken CMA, North Central CMA, Wimmera CMA, Mallee CMA, Glenelg Hopkins CMA and Corangamite CMA), (ii) assess their groundwater dependency (nature and degree), and (iii) identify the threats to these GDEs, their susceptibility to these threats and likely consequences.

Because this project was venturing into a new research field, it was allowed considerable flexibility to begin with. In the early stages of the project, it was recognised, with agreement from the project funders, that the initial project objectives (above) were too ambitious in the project timeframe and needed to be revised to three equally important pre-requisite objectives, as follows:

1) complete a detailed literature review to improve knowledge of GDE types, function and identification, and to help inform development of a broad scale mapping method;

Mapping Terrestrial Groundwater Dependent Ecosystems 9

2) develop a method to predict the occurrence and distribution of GDEs at a regional scale; and

3) produce a series of potential GDE maps for all Victorian NAP regions (Note: with supplementary funding, this was later expanded to the whole State to also include maps for Port Phillip and Western Port CMA, West Gippsland CMA, East Gippsland CMA and North East CMA regions)

The main outcomes sought from the project are:-

1) Valued new knowledge to assist development and refinement of State and regional policies on GDE management and water management in areas of resource competition (including regional Sustainable Water Strategies);

2) Valued new knowledge to assist targeting and management of high-value wetlands/streams and groundwater management areas;

3) Provision of a sound basis for targeting detailed investigations of GDEs; .

4) Reduction in the impact of farming systems on GDEs at risk of disturbance through land use practices, including groundwater pumping and drainage diversion.

10 Mapping Terrestrial Groundwater Dependent Ecosystems

2 What are Groundwater Dependent Ecosystems?

There have been many definitions for GDEs given by various authors, but they are all along the same theme, being ecosystems that use groundwater either at some stage of their life cycle or by one generation which is critical to the existence of that species. Demonstration of groundwater use does not necessarily equate to groundwater dependence. By dependence it is meant that the ecosystem would be significantly altered and even irreversibly degraded if groundwater availability was altered beyond its ‘normal’ range of fluctuation (Colvin et al. 2003).

Hatton and Evans (1998), in the publication, ‘Dependence of Ecosystems on Groundwater and its Significance to Australia’, defined groundwater ‘for the purposes of identifying ecosystem dependence’ as extractable groundwater utilised to a greater or lesser extent, by plants and animals. They noted that the term groundwater should not include transient shallow (soil) water and perched systems including shallow through flow down hillslopes. Hatton and Evans (1998) used vegetation assemblages to classify broad groups of ecosystems that potentially are dependent on groundwater, with dependence being sometimes cryptic or subtle. Most of the current literature refers to GDEs based on an assumption that the groundwater is found in aquifers and able to be abstracted. Colvin et al. (2007) consider the term Aquifer Dependent Ecosystem (ADE), to be more precise; ADE, as used in South Africa, is functionally equivalent to GDE.

Smith et al. (2006) added a time dimension to the term and defined groundwater dependent ecosystems as ecosystems that rely wholly or partially on groundwater to maintain an adequate level of ecosystem function and maintenance of community composition over multiple generations of the longest lived species within the community.

GDE dependence on groundwater is highly variable, ranging from partially and infrequently to continually and wholly dependent. These ecosystems including wetlands, vegetation, mound springs, river base flows, cave ecosystems, playa lakes saline discharges, springs, mangroves, river pools, billabongs and hanging swamps represent complex and important components of biological diversity (Sinclair Knight Merz 2001).

Murray et al. (2003) defined GDEs as requiring the input of groundwater to maintain their current composition and functioning. Removal of groundwater from these ecosystems, or a change in the timing, quantity, quality or distribution of groundwater may influence ecosystems by, for example, changing the availability of water for transpiration by vegetation and the recruitment of seedlings into the adult population. This generally results in changes in associated fauna assemblages.

2.1 Original Classification System from Hatton & Evans (1998) There are a number of different GDE classifications proposed in Australia. Hatton and Evans (1998) identified four types of GDEs primarily based on geographic setting:

1. terrestrial vegetation - vegetation communities and dependent fauna that have seasonal or episodic dependence on groundwater;

2. river base flow systems - aquatic and riparian ecosystems that exist in or adjacent to streams that are fed by groundwater base flow;

3. aquifer and cave ecosystems - aquatic ecosystems that occupy caves or aquifers;

4. wetlands - aquatic communities and fringing vegetation dependent on groundwater fed lakes and wetlands.

2.2 Additional Classes of GDEs Based on the same approach as Hatton & Evans, Sinclair Knight Merz (2001) identified two additional types of GDEs:

5. terrestrial fauna - native animals that directly use groundwater rather than rely on it for habitat;

6. estuarine and near-shore marine ecosystems - coastal, estuarine and near-shore marine plant and animal communities whose ecological function has some dependence on discharge of groundwater.

Mapping Terrestrial Groundwater Dependent Ecosystems 11

2.3 Alternative GDE Classification System by Eamus et al. (2006)

Eamus et al. (2006) grouped ecosystem types for their GDE classification scheme. They proposed three simple primary classes based upon the type of groundwater reliance. This type of classification is useful for management as it allows GDEs to be separated into readily recognisable ecosystem classes that permit the use of similar techniques and approaches for the identification of GDEs and the assessment of ecological risk. The classes are: 1. aquifer and cave ecosystems, where stygofauna (groundwater-inhabiting organisms) reside within the groundwater resource;

2. all ecosystems dependent on the surface expression of groundwater;

ƒ River base flows

ƒ Wetlands, swamplands

ƒ Seagrass beds in estuaries

ƒ Floodplains

ƒ Mound springs

ƒ Riparian vegetation

ƒ Saline discharge to lakes

ƒ Low lying forests

3. all ecosystems dependent on the subsurface presence of groundwater, often accessed via the capillary fringe (non- saturated zone above the water table) when roots penetrate this zone.

ƒ River Red Gum forests

ƒ Banskia woodlands

ƒ Riparian vegetation in the wet/dry tropics

2.4 ADE classification & identification by Colvin et al. (2007)

Colvin et al. (2007) created a new approach to identify Aquifer Dependent Ecosystems (ADEs) based on a combination of aquifer and habitat types in South Africa, or type setting. These specific combinations help structure understanding of these systems and classify them according to the aquifer - ecosystem interface. The actual classes of the GDE habitats was similar to that developed by Hatton and Evans (1998), being: ƒ terrestrial

ƒ riverine aquatic

ƒ spring

ƒ riparian

ƒ wetland/ seep

ƒ estuarine/coastal

ƒ in-aquifer

The aquifer types in South Africa were categorised into 6 principle types based on lithology. These are as follows; ƒ Fractured Metasedimentary

12 Mapping Terrestrial Groundwater Dependent Ecosystems

ƒ Carbonates

ƒ Unconsolidated sediments

ƒ Dolerite dykes and sills

ƒ Basement complexes and younger granites

ƒ Igneous extrusive

Mapping Terrestrial Groundwater Dependent Ecosystems 13

3 Description of GDE Classes

The following sections contain a brief overview of each of the major classes of groundwater dependent ecosystems, following the classification scheme of Sinclair Knight Merz (2001). This classification forms the basis for the regional terrestrial GDE mapping in this project. Although the state-wide mapping process is focused on the terrestrial GDEs, the process also identifies areas in the riverine aquatic and wetland classes of the original Hatton and Evans (1998) system. Supplemental geographic information on water bodies and wetland areas are overlain with the GDE mapping results to aid analysis and interpretation because the boundaries between the classes may be gradational (e.g. in riparian flood-plains).

3.1 Terrestrial vegetation

A definition for terrestrial vegetation GDE’s was provided by Sinclair Knight Merz (2001, p. 9) as:

‘This class of groundwater dependent ecosystem includes vegetation communities that do not rely on expressions of surface water for survival, but which have seasonal or episodic dependence on groundwater’

Groundwater use by terrestrial vegetation has long been established (e.g. Thorburn et al. 1993a; Thorburn et al. 1993b), however defining a degree of dependency has proven more difficult, considering for example that a species may use groundwater once every decade to survive or once each year. Terrestrial vegetation GDEs may be found in areas with no proximal surface water expression. However, they grade into riparian zones of ephemeral streams, streams, or other water bodies.

A large number of studies have been undertaken over the past two decades to identify different terrestrial ecosystems in Australia and overseas. A mixed Eucalypt forest within the Brindabella Range, ACT was recorded to use groundwater only during a severely dry period during the early 1980s (Talsma and Gardner 1986). Eucalyptus camaldulensis (river red gum) along the Chowilla flood plain use groundwater during summer months and use surface and groundwater during winter months or when fresher sources are available (Mensforth et al. 1994), indicating this species has established an annual pattern of groundwater use. An experimental study in the Murray-Darling basin showed Eucalyptus camaldulensis moisture stress was lower between flood events in plots underlain by shallow aquifers, implying groundwater dependency (Bacon et al. 1993). Banksia species on the Gnangara Mound, Western Australia were also shown to be groundwater dependent (Canham et al. 2009).

Lamontagne et al. (2005a) used δ2H (deuterium) isotopic and soil water matric potential measurements to delineate groundwater dependency of tree species in a tropical savanna riparian zone in the Northern Territory. Melaleuca argentea and Barringtonia acutangula appeared to be obligate phreatophytes while Cathorium umbellatum and Acacia auriculiformis appeared to be facultative phreatophytes. Zencich et al. (2002) used δ2H to determine seasonal variations in groundwater use by Banksia ilicifolia and Banksia attenuate on the Swan coastal plain, Western Australia.

O’Grady et al. (2006) evaluated groundwater dependence of subtropical woodland species in northern Queensland. They found evidence of some degree of groundwater use in all communities studied. Combined measurement of evapotranspiration rate, leaf and soil water potential, and δ18O of soil water/groundwater/xylem water were used to model the depth of water extraction for different species in one stand (Cook and O’Grady 2006). The model was used to infer groundwater usage by species of Eucalyptus platyphylla, Melaleuca viridiflora, Lophostemon suaveolens, and Corymbia clarksoniana. The estimated dry season groundwater contribution ranged from < 15% for E. platyphylla and L. sauveolens, to 100% for C. clarksoniana.

Costelloe et al. (2008) extended the investigation of GDEs to saline groundwater use by arid zone riparian Eucalyptus coolabah in the Lake Eyre Basin, South Australia. They used a combination of δ18O, δ2H, soil chloride, and soil matric potential profiles. E. coolabah adapted to the dry saline conditions by depending to a greater extent on soil water including bank storage, by maintaining a low evapotranspiration rate compared to other riparian species, and by a higher salinity tolerance than other riparian eucalypts.

Gries et al. (2003) demonstrated that the growth rate of woody vegetation in the Taklamakan Desert, China, depended on the depth to groundwater for Populus species but to a lesser degree for Tamarix. The ability to use groundwater from depths up to 24 m depth resulted from upward growth of above ground shoots as the sand dunes shifted. Thus the vegetation is susceptible to rapid water table decline due to anthropogenic activities. Additional evidence for groundwater dependency of forests at the periphery of the Taklamakan Desert was presented by Bruelheide et al. (2010). Although groundwater depths

14 Mapping Terrestrial Groundwater Dependent Ecosystems

ranged from 2.3 to 17.5 m, periodic flooding and vadose zone soil moisture use was discounted as water sources for the vegetation. Forestry or water usage may permanently impact the ability of these phreatophyte species to respond to shifting dunes and other environmental disturbances for which they are adapted. Hao et al. (2010) determined that, along the Tarim River on the margin of the Taklamakan Desert, the optimum groundwater level for major plant growth is 2-4 m with a threshold for groundwater dependency of approximately 6 m.

A number of different terrestrial ecosystems dependent on groundwater have been described in South Africa (Colvin et al. 2007). Of note are the abundant Acacia erioloba stands on Kalahari sands that have root systems up to 30-60 m deep. The authors consider the areas with groundwater dependent species to be keystone ecosystems, supporting a rich variety of fauna, and the deep-rooted trees acting as nutrient pumps and providing water to shallower-rooted plants by hydraulic lift.

Terrestrial GDEs have also been identified in North America. Woody and herbaceous groundwater dependent species were identified in a study of the impacts of groundwater withdrawal on vegetation in the Great Basin and Mojave deserts (Patten et al. 2008). The authors identified phreatophitic species of Atriplex, Prosopis, Isocoma, Chrsothamnus, Distichlis, Sporobolus, Artemesia, Salix, and others in the upland zone adjacent to springs and in the wetland/upland transition zone. The expected response to declining water tables would be replacement of transitional communities by upland communities and increasing encroachment of salt tolerant plants (halophytes). Conversely in upland communities, a decline in soil salinity could occur with declining water tables, with loss of halophytes. Phreatophytic upland communities would be altered to favour deep-rooted species or replaced by non-phreatophytes.

Effects of groundwater decline on Dutch forests include loss of moist forest types, in particular those dependent on Ca-rich groundwater (Van Tol et al. 1998). Groundwater level decline was associated with decline in base-exchangeable cations, decrease in pH, and increased nitrogen availability. Wagner and Bretschko (2003) found diurnal flow variation in an Austrian stream due to groundwater use by trees in the riparian zone. The interconnection between the stream and groundwater was through a network of preferential flow paths.

3.2 Wetlands - fresh and saline groundwater dependency

Wetlands can be defined as ‘land permanently or temporarily under water or waterlogged, with temporary wetlands having surface water or waterlogging of sufficient frequency and/or duration to affect the biota’ (Paijmans et al. 1985, cited in Hatton and Evans 1998). Most, but not all of the wetlands in Australia depend on groundwater to some degree. Groundwater dependent wetlands are generally considered to be vulnerable to changes in groundwater level, because small changes in the depth to groundwater can significantly reduce the water available to vegetation.

Wetland GDE vegetation type is often strongly related to groundwater chemistry (Bedford and Godwin 2003). Examples include fens in the North America and Europe fed by calcareous, high pH groundwater, (Siegel and Glaser 1987; Almendinger and Leete 1998; Podniesinski and Leopold 1998; Kemmers et al. 2003). Conversely, organic acids produced in the wetlands neutralize the groundwater so that, depending on the groundwater composition and flux, a continuum is formed from neutral to alkaline – ‘extremely rich fens’, through ‘rich fens’, ‘poor fens’, to pH 3.6-4.2 ‘bogs’ (Siegel et al. 2006). In addition, the interaction of groundwater with organic matter in wetlands can result in redox gradients that influence soil composition and phosphorous availability, and thus the plant species (Boomer and Bedford 2008).

It may be difficult to draw a clear distinction between terrestrial vegetation GDEs and wetland GDEs. The former rarely have surface water expression while the latter have frequent or episodic inundation by surface water or groundwater discharge. Thus, strictly speaking a terrestrial GDE should only occur in a zone not classified as wetland. However, management changes, such as stream regulation and control of flooding in the Murray-Darling basin, may increase groundwater dependency and former wetlands may then be considered terrestrial systems. There can be considerable overlap in vegetation species wetlands and terrestrial GDEs. For example, Eucalyptus coolabah, is a salt-tolerant tree that grows in the riparian zone floodplains of the Lake Eyre Basin and Murray-Darling Basin. It may be mostly or completely reliant on saline groundwater use or may use a mixture of groundwater and infiltrating floodwater (Costelloe et al. 2008).

Hatton and Evans (1998) felt that groundwater dependent wetlands form the most extensive and diverse set of the ecosystems within Australia that interact with groundwater. They described a number of these groundwater dependent wetland ecosystems including:

Mapping Terrestrial Groundwater Dependent Ecosystems 15

Swamp sclerophyll forests and woodlands – These ecosystems occupy the riparian corridors of ephemeral or base flow dependent streams and exhibit at least seasonal dependency on groundwater. They include a wide range of mostly eucalypt species such as E.ovata, E.viminalis and E.leucoxylon communities in South Australia, E.camaldulensis and E.largiflorens woodlands of the Murray and Darling River floodplain and of the inland river systems of central Australia.

Swamp scrubs and heaths – This group of ecosystems normally occupy sandy or peaty soils in landscapes ranging from coastal dunes to swampy areas fed by snow melt in the southern Australian highlands. Farrington et al. (1990) found substantial use of groundwater by swamp scrub on the sumplands and damplands of the Swan Coastal Plain in Western Australia.

Swamp shrublands –Lignum shrublands (Muehlenbeckia cunninghamii) often grow in association with E. largiflorens woodlands in the Murray-Darling basin on heavy-textured grey and brown soils that are periodically inundated. Halophytic shrublands (Chenopodium auricomum, C. nitrariaceum, Atriplex nummularia, Maireana spp., Bassia spp. and Rhagodia spp.) occur around the margins of inland salt lakes in central Australia and also colonise groundwater discharge areas and the edges and beds of salt lakes that are only inundated periodically across Australia (Arthrocnemum spp.). The links between these swamp shrublands and groundwater is not known at this stage.

Sedgelands – Sedgeland communities are found in the coastal, floodplain and valley floor environments of eastern Australia. Most require at least seasonal waterlogging. Those that require permanent surface wetness are almost certainly groundwater dependent e.g. Eleocharis sphacelata sedgelands in lagoons of the Murray River and tributaries. Baumea sedgelands occur in permanent and semi-permanent wetlands from the coastal areas of south-east Queensland down to Tasmania and around to south-west Western Australia. Their need for continual surface wetness suggests that they are likely to be interacting with groundwater. Button Grass (Gymnoschoenus spaerocephalus) sedgelands occupy waterlogged, peaty, infertile soils that are periodically flooded on valley floors in Tasmania’s south-west and on the tablelands of New South Wales. Where the sites are continually wet, button-grass may be replaced by Leptocarpus tenx, Xyris operculate, Lepidosperma longtitudinale, L. filiforme and Restio tatraphyllus. Groundwater is almost certain to be influencing the vegetation community at these sites. Carex sedgelands occupy permanently waterlogged sites that are periodically flooded on the tablelands of New South Wales, Victoria and Tasmania. Groundwater clearly has a role in maintaining the permanently wet state at some sites, but at others the degree of interaction is not clear.

Swamp grasslands – Phragmites and Typha swamp grasslands are common in seasonally and permanently waterlogged locations across coastal and inland parts of south-eastern Australia. Their groundwater dependency was considered by Hatton and Evans (1998) to vary widely, but communities associated with more permanent water features are considered the most likely to exhibit some degree of dependency. Sod tussock grasslands characterised by Poa caespitosa, Themeda australis, Danthonia nudiflora and Calorophus lateriflorus occur in the subalpine and alpine areas of Victoria. Any ground water interaction is likely to be limited. Canegrass (Eragrostis australasica) grasslands grow in low-lying, flood-prone locations in North-western Victoria. The regular flooding regime at these sites suggests at most a limited dependency on groundwater.

Swamp herblands – Floating and floating leaved herblands are common in coastal rivers and dune swales and lakes throughout Australia. The characteristic wetness of the locations implies some role for groundwater and associated ecosystem dependency.

Hypersaline lakes – A number of the RAMSAR wetlands in the Kerang-Swan Hill region and Lake Tyrrell wetland complex in northern Victoria and numerous lakes across the Basalt Plains in southern Victoria are examples of unique and viable ecosystems that have successfully developed around these hostile environments. These areas are dependent on groundwater and the aquifers may have had their groundwater function altered/threatened by irrigation practices and water supply management. Natural spatial and temporal variability in salinity is an important characteristic of saline wetlands (Jolly et al. 2008). The salinity and chemistry of saline lakes in the Wimmera is dependent on the chemistry of the source groundwater and the hydrologic setting, leading to high diversity between individual lakes (Radke and Howard 2007). The effects of changing salinity on vegetation in the salt marsh fringe of Lake Austin, Western Australia was studied by van Etten and Vellekoop (2009).

16 Mapping Terrestrial Groundwater Dependent Ecosystems

3.3 Coastal estuarine and near shore marine systems

Coastal estuaries including brackish lakes and salt marshes can be zones of groundwater discharge. The groundwater sources can have important impacts on salinity and nutrient levels in the ecosystem and groundwater contamination may harm the ecological receptors. Recent research into submarine groundwater discharge (SGD) is increasing understanding of the role of the direct discharge to coastal waters in the regional water budgets and to marine vegetation and aquatic communities. However, the extent of groundwater dependency is generally poorly understood.

Gallardo and Marui (2006) review research on submarine groundwater discharge, focusing on the physical processes rather than ecological effects. Burnett et al. (2003) review the physical and geochemical effects of SGDs. Bokuniewicz et al. (2003) discuss the importance of SGDs to coastal management and present a global assessment of coastal SGD areas. Evaluation of these GDEs is complicated by tidal effects leading to variability in discharge quantity, quality, and location of the groundwater interface with marine water and coastal streams (Acworth et al. 2007).

Loaiciga and Zektser (2003) reviewed methods for estimating SGD. Smith and Nield 2003) used a numerical groundwater flow model to estimate SGD to Cockburn Sound, Western Australia, while in the same area, Tanaguchi et al. (2003) compared estimated SGD from groundwater and porewater temperatures to seepage meter measurements. Danielescu et al. (2009) used thermal infrared imaging to evaluate groundwater discharge to a coastal estuary in Prince Edward Island, Canada. Several studies have used radon and radium as tracers to investigate SGD areas and estimate flux (e.g. Moore 2003; Lambert and Burnett 2003).

Nutrient phosphate and ammonia flux to estuaries and SGDs is a particular issue of coastal groundwater discharge. Crotwell and Moore (2003) used natural radium as a tracer to determine nutrient flux to Port William Sound, South Carolina, USA and suggest that the nutrient load may be reduced by regional groundwater use.

Several examples of groundwater dependent coastal and marine ecosystems described by Sinclair Knight Merz (2001) are presented below, with additional references.

Coastal mangroves and salt marshes – mangroves are widely distributed around the Australian coast. While most common in northern Australia, they may be found as far south as Corner Inlet in Victoria. While seawater is considered to be the primary water source for most of these vegetation communities, sites have been noted where mangroves occupy relatively fresh groundwater discharge areas (Adam 1994). Semeniuk (1983) related the distribution of mangroves in Northwestern Australia to areas of groundwater discharge where conditions were more arid. Drexler and Ewel (2001) investigated a coastal area in Micronesia and showed that groundwater flowed from a freshwater swamp to a mangrove stand during normal conditions but was reversed during an El Niño-Southern Oscillation (ENSO) drought and suggested possible mechanisms of groundwater dependency for the mangroves.

Salt marshes tend to replace mangroves in coastal locations in southern Australia. The nature of any groundwater dependency is unknown. Protection of coastal mangroves and salt marshes from clearing and drainage may play an important role in maintaining groundwater discharge and preventing the activation of acid sulphate soils.

Coastal lakes – coastal lakes along the south-west coastline of Western Australia support the development of stromatolites (Palinska et al. 1999) and have quite varying aquatic communities. Groundwater is the principal source for many of the lakes. Some Victorian coastal lakes and wetlands maintain fresh to brackish species compositions due to the discharge of relatively fresh groundwater.

Sea grass beds – the distribution of sea grass beds in some coastal areas is influenced by groundwater discharge. Anthropogenic impacts to groundwater nitrogen may increase nutrient loading in shallow coastal waters, increase macroalgae growth, and decrease eelgrass distribution. Eutrophication and other effects associated with shift from sea grass to magroalgae then lead to decreased benthic animal abundance and fish kills (Valiela et al. 1990).

Marine animals – some marine and estuarine animals depend on groundwater discharge to provide a suitable habitat or an appropriate environment in which the species of plant and/or animal they eat will prosper. Groundwater discharge may be in the form of direct off-shore discharge or base flow into streams that discharge to the ocean. Examples of groundwater dependent fauna include crocodiles, turtles, fish and macro-invertebrates.

Mapping Terrestrial Groundwater Dependent Ecosystems 17

3.4 River base flow systems

This category of ecosystem was devised by Hatton and Evans (1998) to include the many ecosystems that are dependent on groundwater derived base flow in streams and rivers. Base flow is that part of stream flow derived from groundwater discharge and bank storage. Base flow may contribute year round to flows in coastal streams in south-eastern Australia and may contribute to flow in inland streams, although the extent of the contribution may be difficult to determine in some cases due to river regulation (Hatton and Evans 1998). The coastal rivers of south-eastern Australia maintain base flow throughout the year and support riparian forests, scrub, sedgelands and grasslands, as well as in-stream biota and floating and emergent herbfields.

Riparian and aquatic ecosystems in base flow dependent streams would be, to a greater or lesser extent, groundwater dependent themselves. Demarcation between groundwater dependent terrestrial vegetation, wetlands, and base flow systems may be difficult. The three types of community represent ranges on a spectrum of habitat and groundwater dependency.

O’Grady et al. (2006) point out that riparian woodland species tend to be opportunistic in the sources of water they are able to use. They provide the examples of Lamontagne et al. (2005a) and Zencich et al. (2002) of Casuarina, Acacia, and Banksia species using groundwater when it is shallow and accessible, but other sources as available.

Groundwater level in a riverine aquifer is also important in terms of maintaining a hydraulic gradient towards the stream that supports the necessary discharge flux. Hatton and Evans (1998) noted that across at least some of its range, the platypus was an example of groundwater dependent fauna. In some parts of this species’ range, groundwater discharge is required to sustain the flow or pools in which it feeds.

Contamination of riverine aquifers by nutrients, pesticides and other toxicants may adversely affect dependent ecosystems in base flow streams. Aquatic communities would be expected to be the worst affected.

3.5 Aquifer and cave ecosystems

This category comprises the aquatic ecosystems that may be found in free water within cave systems and within aquifers themselves. Australia studies of these ‘stygean’ ecosystems have traditionally related to cave, rather than aquifer systems, however there is a growing body of information on the latter (e.g. Boulton et al. 2003; Eberhard et al. 2009; Hancock and Boulton 2009; Bradford et al. 2010). Karst and other cave systems also support diverse ecosystems (e.g. Piccaninnie Ponds in South Australia - Scholz 1990; palaeovalleys within central Australia - Humphreys 2006).

Aquifers themselves support a diverse array of ecosystems. Some ecosystems (e.g. in riverine plains) exist along a continuum between fully aquatic communities and completely aquifer dwelling communities (Danielopol 1989).

The environment in which aquifer ecosystems develop is characterised by darkness, consistency and persistence of habitat and low energy and oxygen availability. The organisms that inhabit these environments are often specialised morphologically and physiologically. They are also predominantly invertebrates. These stable and confined environments result in high levels of endemism and high proportions of relict species in comparison to surface environments (Danielopol 1989).

3.6 Terrestrial fauna

Descriptions of surface groundwater dependent ecosystems in the previous sections have mainly concentrated on plant communities. These communities provide habitat for a variety of terrestrial, aquatic and marine animals, which by extension must also be groundwater dependent.

However, there is an additional group of groundwater dependent fauna whose reliance on groundwater is not based on the provision of habitat, but as a source of drinking water. Groundwater, as river base flow or discharge into a spring or pool, is an important source of water across much of the country, particularly in northern and inland Australia and other areas with semi- arid climate. Its significance is greater for larger mammals and birds, as many smaller animals can obtain most of their water requirements from respiration.

18 Mapping Terrestrial Groundwater Dependent Ecosystems

Pastoralists in inland Australia have made extensive use of groundwater to supply drinking water to grazing stock. In addition to watering stock, groundwater is also used by native fauna (e.g. kangaroos) and pest and feral animals. Groundwater dependent terrestrial and riparian vegetation and wetlands may be used by terrestrial fauna as drought refuges. Access to groundwater allows the vegetation to maintain its condition and normal phenology (e.g. nectar production, new foliage initiation, seeding). Populations of some birds and mammals retreat to these areas during drought and then recolonise drier parts of the landscape following recovery. The long term survival of such animal populations relies on maintaining the vegetation communities and ensuring their water requirements are met. The key groundwater attributes will be flux and pressure. These are dependent on the hydrology of the groundwater flow system providing the water.

Mapping Terrestrial Groundwater Dependent Ecosystems 19

4 Known Victorian areas with GDEs

An extensive literature review was conducted to identify locations within Victoria where field studies had established a groundwater connection with a landscape and highlighted areas that will contain GDEs. Known locations from the review are shown on Figure 1, along with locations where stream base flow measurements have been made. Base flow data are important for evaluating the groundwater dependency of wetland and base flow systems and are useful for understanding the hydrogeologic context for terrestrial systems. It was not the purpose of the review to critically evaluate previous research, but to determine the general landscape settings of known locations of potential GDEs within Victoria; some examples are provided in Table 1. The data set was used to assist in the model development, acting as a guide to likely locations for GDEs

Figure 1. Distrubution of wetlands where previous research has established a link with groundwater and location of stream gauges where base flow calculations have been made Table 1 Some references of some research into groundwater-surface water interaction and groundwater use by vegetation within Victorian and adjacent landscapes Landscape Some cited literature Murray River Lamontagne et al. (2002); Lamontagne et al. (2005b) Groundwater interaction streams Hatton and Evans (1998); Nathan and Weinmann (1993), Lacey (1996). Northern Loddon plains Macumber (1991) Palaeozoic Upland Springs Shugg and Brumley (2003); Shugg (2009) Basalt Plains wetlands Barton et al. (2006) Kerang Wetlands Macumber (1991) Avoca River wetlands Macumber (2004) West Wimmera wetlands Fawcett and Huggins (2005) Mallee Wetlands incl. Lake Tyrrell Macumber (1980); Macumber (1991); Macumber (1992) Terrestrial Vegetation Macumber (2004); Benyon and Doody (2004); Benyon et al. (2006); Morris and Collopy (1999), Southern Victorian wetlands Fluvial Systems (2006); Mensforth (1996); Turnbull (2006) Coastal setting Corangamite CMA. Fitzpatrick et al. (2007)

20 Mapping Terrestrial Groundwater Dependent Ecosystems

5 Identification of GDEs

The focus of this project is on the identification of areas of potential GDEs using large scale techniques but these can be supported by smaller scale field techniques.

Investigations to identify GDEs typically employ combinations of several indicators, each of which only provide partial information related to the groundwater use. First, the types of information will be described and then a number of examples from the literature will be discussed to illustrate the overall approaches.

A fundamental tenet of ecology is that ecosystems generally will use resources in proportion to their availability - whether the resource is light, water, nitrogen or anything else - and the availability of different resources will be a significant determinant to their structure, composition and dynamics (Colvin et al. 2003). Thus where groundwater is accessible, ecosystems will probably develop some degree of dependence on it and that dependence is likely to increase with increasing aridity of the associated environment (Hatton and Evans 1998).

An analysis of vegetation assemblages can show the presence of species likely having the ability to use groundwater. Hydrogeologic studies can identify where the groundwater is discharging to surface springs, streams, lakes, or coastal zones. Specifically studies of groundwater-surface water interaction can show areas, not only of aquatic groundwater dependency but where riparian vegetation is using groundwater. Other geologic and hydrogeologic features of importance include bedrock type, regolith depth, and depth to the water table.

Eamus (2009) discusses field methods to evaluate the groundwater dependency of ecosystems. He recommends determinations based on the use of vegetation and hydrologic measurements including pre-dawn leaf water potential, stable isotopes of oxygen and hydrogen in tree xylem and soils, soil sampling to establish root depth, leaf aria index measurement, depth to groundwater, diurnal changes in groundwater depth, and water balance calculations.

Colvin et al. (2003) and Clifton et al. (2007) produced tool boxes of the various techniques available to indicate groundwater use. They consider each method in terms of its technical basis, costs, constraints, suitability, time requirements, resolution of results, format of outputs, and previous use of the technique.

Clifton et al. (2007) provided a description of applications and limitations of a range of scientific approaches in GDE assessments (Table 2). With the exception of the geological mapping that already exists, the methods detailed in Table 2 are generally not sufficiently cost or time effective to be applied to broad scale identification or characterisation of GDEs.

5.1 Vegetation analysis

Where fresh groundwater discharges to the surface, the constant availability of water sustains plant photosynthetic activity longer in summer in that area (Tweed et al. 2007). Areas where shallow saline groundwater is close to, or discharges to the surface, are also evident from the limited vegetation activity. Soil subject to saline groundwater accumulates salts, thus limiting photosynthetic activity of salt sensitive vegetation throughout the entire year whilst supporting photosynthetic activity of salt-tolerant vegetation (Tweed et al. 2007).

Areas of vegetation that use groundwater typically exhibit low seasonal variability of photosynthetic activity when compared with productive agricultural areas. The inter-annual variability of the vegetation activity can also contain useful information to assist mapping of groundwater discharge areas (Tweed et al. 2007).

Vegetation that is relatively lush in comparison to surrounding areas is often associated with discharge areas. (Colvin et al. 2007). Leaf area index (LAI) is often used as an indicator to identify this type of vegetation. LAI is defined as the one sided green leaf area per unit ground area in broadleaf canopies, or as the projected needle leaf area per unit ground area in needle canopies. LAI is a strong indicator of water availability in semi-arid to arid environments (Hatton and Evans 1998; Colvin et al. 2007).

Mapping Terrestrial Groundwater Dependent Ecosystems 21

Table 2. Tools to identify GDE's (after Clifton et al. 2007) Tool Description

1. Mapping tools Mapping of geology and geological structures, water table depth or aquifer pressure and the distribution, composition and/or condition of vegetation as a means of identifying ecosystems that are likely to have access to and use groundwater.

2. Water balance Identify and quantify groundwater use by measurement or estimation of various components of techniques an ecosystem’s water cycle.

3. Pre-dawn leaf water Pre-dawn leaf water potential measurements to identify groundwater use and depth of water potential uptake.

4. Stable isotope Comparison of the fractionation of isotopes in plant xylem water with potential source waters to analysis - vegetation identify groundwater uptake.

5. System response to Long term monitoring of ecosystem composition and ecological function in response to change management intervention, climate, and soil water, surface water and groundwater conditions.

6. Groundwater - Application of hydraulic principles, statistical analyses of stream hydrographs and site surface water hydraulics measurements to derive the degree of interaction between groundwater and surface water features. 7. Physical properties of Measurement of water electrical conductivity and temperature change along the length of a river water / wetland, or over time to identify a groundwater contribution.

8. Analysis of water Chemical analysis of surface water and groundwater for isotopes, major anions and cations, chemistry and trace elements. Mixing relationships identify groundwater contribution.

9. Introduced tracers Use of introduced chemical tracers to observe mixing and dilution relationships and assess the contribution of groundwater to a stream.

10. Plant water use Mathematical representations (or models) of plant water balance to estimate plant water modelling requirements and/or groundwater uptake, and/or response to water table drawdown.

11. Groundwater Two or three-dimensional mathematical representations (or models) of water movement in the modelling saturated and unsaturated zones to assess the potential level of interaction between groundwater and surface water bodies and between groundwater and terrestrial ecosystems.

12. Conceptual Use of expert knowledge of similar ecosystems, biophysical environments and relevant data to modelling develop a conceptual model of the ecosystem and its interaction with groundwater.

13. Root depth and Assessment of the depth and morphology of plant root systems, and comparison with measured morphology or estimated water table depth to assess potential for groundwater uptake.

14. Analysis of aquatic Use of ecological survey techniques to identify aquatic species with reproductive behaviour or ecology habitat requirements that indicate groundwater dependency.

5.2 Groundwater – surface water interaction

Investigation of groundwater-surface water interactions is a broad and rapidly developing field that will only be briefly touched on here. Identification of groundwater discharge zones and quantity of discharge is not only important for determining the

22 Mapping Terrestrial Groundwater Dependent Ecosystems

dependency of the surface water system, but the groundwater-surface water interface, termed the hyphoreic zone, often forms an important ecological niche.

Brodie et al. (2007a) provide an extensive discussion of groundwater-surface water interaction from the perspective of Australian water resource management. Field methods developed to investigate the spatial-temporal variability of recharge and discharge processes are typically site specific (Kishel and Gerla 2002). Field methods include: • Water balance, seepage meters, piezometer, and stream gauging/hydrograph separation (Brodie et al. 2007a, 2007b) • Temperature monitoring (Brodie et al. 2007b). • Geophysical surveys, especially electrical conductivity in the subsurface or surface water (Brodie et al. 2007b) • Chemical and oxygen/hydrogen isotopic relationships between groundwater and surface water (e.g. Chapman et al. 2003; Gibson et al. 2000; Zench et al. 2002) • Tritium and 3He measurements to characterize exchange between groundwater and wetlands, estimating transit times in catchments, and constrain flow models (e.g. Harvey et al. 2006; Stewart et al. 2010) • Strontium isotopes in groundwater and surface water (e.g. Kirchner 1995; Kirchner 1997; De Villiers 2005), • Radon and radium isotopic measurements have been used to characterize groundwater discharge to streams and the ocean (e.g. Ellins et al. 1990; Moore 2003; Wu et al. 2004; Dulaiova et al. 2005; Oliveira et al. 2006; Schmidt et al. 2009)

5.3 Remote sensing approaches

Smith et al. (2006) devised a method of determining use of groundwater by terrestrial ecosystems by modelling the groundwater levels and overlaying vegetation maps obtained from satellite imagery. This involved matching woody vegetation within the regions of influence of groundwater pumping to predict GDEs which may be as risk from groundwater extraction.

Tweed et al. (2007) applied a combination of remote sensing and geographic information system (GIS) techniques to map groundwater recharge and discharge areas in the Western District of Victoria. By reviewing hydrological processes, a series of surface and subsurface indicators of recharge and discharge were established. Various remote sensing and GIS techniques were then used to map these surface indicators including; terrain analysis, vegetation activity or normalised difference vegetation index (NDVI) obtained from satellite imagery and mapping of infiltration capacity.

Münch and Conrad (2007) used a combination of remote sensing and GIS modelling to produce a GDE probability map in South Africa. Starting with remote sensing data, satellite images were assessed to identify probable GDEs. A GIS model predicted landscape wetness based on terrain and geomorphologic character or land wetness potential. It was modified to highlight groundwater generated landscape wetness. Other techniques were then applied to the modelling and included a groundwater elevation model, interpolated and digital elevation data and biomass indicators generated from Landsat. The information was then classified and combined with GIS layers and followed up with field verification. The result was a map of areas of potential GDE’s based on groundwater level and soil moisture availability.

Bierwirth and Welsh (2000) used a combination of airborne radiometric and satellite remote sensing measurements to investigate groundwater recharge in the Great Artesian Basin.

5.4 Geological mapping approaches

Regional information on geology, structural geology, regolith mapping and the geomorphology can be assessed to help search for areas where potential GDEs may exist. The surface and sub-surface characteristics generated by the geology and structures define groundwater flow pathways and thus the establishment and support of ecosystems within certain areas. Salama et al. (1993, 1994) used geological and geomorphological data to determine the likelihood of groundwater discharge sites. Salama et al. (1994) integrated the interpretation of geomorphic and geological features, from Landsat satellite imagery and aerial photographs, with on-ground hydrogeological studies to map recharge and discharge areas. It was found that recharge areas were linked to increases in permeability of the surface geology, whereas discharge areas were associated with major drainage lines, geological boundaries and topographic depressions.

Mapping Terrestrial Groundwater Dependent Ecosystems 23

• Groundwater discharge and shallow watertable levels correlate to areas of soil salinity, saline lake systems and fresh water springs. (Macumber 1991; Tweed et al. 2007) and areas of GDEs. (Colvin et al. 2003). Groundwater transpiration by vegetation and direct discharge to wetlands and rivers, means that shallow water tables are an essential component of sustaining ecologically significant areas. (Batelaan et al. 2003).

• Geology, geomorphology and structural geological features influence groundwater flow and groundwater discharge (and potentially the location of GDEs.) (Macumber 1991; Colvin et al. 2003) Ecosystems or species which occur in association with (potential) discharge areas, such as topographic low points or along dykes or fault lines can be used as groundwater indicators (Colvin et al. 2003)

• Potential indicators of groundwater discharge include terrain indicators such as topographic depressions and break of slope and groundwater flow direction towards surface water bodies (rivers, wetlands, and the ocean) (Tweed et al. 2007)

• The influence of stratigraphical layers and jointing patterns on spring locations within the bedrock of New Mexico was determined by Walsh (2008).

Macumber (1991) identified distinct geomorphology units where groundwater discharges to the surface in the north-west regions of Victoria. Colvin et al. (2003) used geological features as indicators of fractured aquifer potential or areas where groundwater availability was greatest. This was followed up with botanical field verification of these areas to see if the areas supported GDEs in South Africa. The reasoning behind the study was that geology and lithology control groundwater flow and impose aquifer boundaries. Fractures, faults, folds and intrusive dykes may form a preferred pathway or barriers to groundwater flow and thus availability of groundwater for terrestrial ecosystems.

Brodie et al. (2007b) developed a mapping method to predict surface and groundwater interaction using site and area specific characteristics and formulae developed for weighting these inputs. The method takes into account hydrological and hydrogeological factors. It focuses on the conductance of the geological material and watertable depths to derive an indicator for the potential for groundwater – surface water connectivity. It provides sufficient information to identify potential connected and disconnected systems to enable targeting of further investigations and management. The stream-aquifer connectivity potential is rated using an index. The input required includes: ƒ Depth to groundwater ƒ Streambed characteristics ƒ Geology ƒ Geomorphology.

The final single index value is categorised into low, medium or high connectivity potential classes based on the output classification classes in the model. These categories can be mapped to spatially represent the estimates of potential connectivity along river reaches. Potentially connected surface and groundwater systems support ecosystems that are potentially dependent on the groundwater to some degree. Indirectly studies have developed methods of mapping springs, and therefore developed maps of GDE’s, for example Walsh (2008), utilised GIS data to map jointing patterns within bedrock across New Mexico that are associated with springs. .

Previous work on regional identification of potential accessibility of groundwater to vegetation generally uses a depth of 5 metres as the limit within which dependent plants are likely to access groundwater, although this could be seriously challenged in some Australian landscapes. Depth to groundwater maps can be used to help locate GDEs but these are very difficult to construct with reliable accuracy at regional scales due to landscape complexity and sporadic distribution of groundwater level data.

24 Mapping Terrestrial Groundwater Dependent Ecosystems

6 Development of state-wide potential GDE maps

An integrated method has been developed to map terrestrial GDEs at a regional scale within the State of Victoria. Techniques applicable to regional GDEs identification, discussed above, were screened and indicators chosen to incorporate into a process for evaluating landscapes for the location of potential GDEs. Data analysis was generally performed on areas defined by Catchment Management Authority (CMA) or adjoining CMAs (Figure 2). The overall process used geological, hydrogeological, and ecological data sets to define regions of common physical and climatic properties, expected to have reasonably consistent signals in remote sensing to the use of groundwater. Remote sensing data sets for each region were then processed to define areas with potential GDEs. Next, the terrestrial GDE class was subdivided, based on indications of the landscape degree of groundwater interaction which is a compilation of derived regional maps distinguishing areas of shallow groundwater from areas where the groundwater is only within reach of deep-rooted trees. Other attributes were added to the geographic information system database for the GDE locations. During this process an exploratory approach was used and professional judgement applied to determine the appropriate data cut-off values for the classification.

Figure 2. Catchment Management Authority (CMA) regions for Victoria

6.1 Spatial data sources for GDE mapping

Data sources used in the mapping are summarized below. Published reports and site-specific information were incorporated into the process where possible. The data sources are divided into three groups, based on process stage. The DPI/DSE Corporate Geospatial Data Library (CGDL) contains geographic datasets in ArcInfo™ geographic information system (GIS) format and was the source of many of the data layers used in the analysis (O’Brien 2004). 1) Definition of Climatic Regions a) Thornthwaite’s moisture index (aridity index): 5 km grid (derived data set, see below)

Mapping Terrestrial Groundwater Dependent Ecosystems 25

2) Vegetation classification of known GDEs a) Ecological Vegetation Classes (CGDL) 3) Remote Sensing Identification of GDEs a) Landsat spectral data from summer images developed for the Australian Greenhouse Office (AGO) http://www.climatechange.gov.au/. The Landsat data were provided by AGO at a resolution of 30 m. b) Landsat NDVI: Midsummer (February images for mid-1990s period of average precipitation and post-2002 drought period). The Landsat NDVI data were obtained at a resolution of 30 m from AGO. c) MODIS EVI: MOD13Q1 product available on the Land Processes Distributed Active Archive Centre (LPDAAC) available on the NASA WIST website at http://edcimswww.cr.usgs.gov/pub/imswelcome/. The Enhanced Vegetation Index (EVI) that comes as part of the MOD13Q1 product provides an estimate of photosynthetic activity at a spatial resolution of 250m every 16 days. 4) Land Use filtering of remote sensing output a) Land use 1:100,000 (CGDL) 5) GDE attribution a) Indicative groundwater depth (derived data set – see below) b) Geomorphic Management Unit (CGDL) c) Groundwater Salinity (CGDL; Clark and Harvey 2008) d) Surficial geology (CGDL)

Derived spatial data sets used were: 1) Thornthwaite’s moisture index: 5 km grid i) SILO gridded rainfall and evapotranspiration: Gridded rainfall and potential evapotranspiration surfaces based on interpolated climate data collected by the Bureau of Meteorology were downloaded from the SILO website: http://www.longpaddock.qld.gov.au/silo/index.html. This data has an approximate spatial resolution of 5 km. ii) AGO Landsat summer image mosaic 2) Indicative Groundwater Depth based on: a) Stream base flow indices; these inform on groundwater interaction with streams. Nathan and Weinmann (1993) derived base flow indices for 39 stream basins from data obtained from 117 stream gauges across Victoria. b) Known groundwater interactive environments determined through a literature search c) Soil and/or geomorphologic characteristics (CGDL) d) Hydrological units, wetlands, streams, swamps and vegetation indicative of saturated, inundated environments (CGDL) e) GIS Soil Salinity layer (Victorian Corporate Spatial Data Library; Clark and Harvey 2008), a data set that implicates groundwater interaction with the soil surface. f) Groundwater data i) The State Observation Bore Network (SOBN) (sourced from http://www.vicwaterdata.net/vicwaterdata/home.aspx) ii) The Victorian dryland salinity observation bore network iii) Groundwater Beneficial Uses maps iv) Groundwater Flow System maps 3) Vegetation types associated with known GDEs a) Statewide areas of selected EVC classes b) Literature review of known GDEs 4) Remote Sensing Identification of GDEs a) Unsupervised classification of Landsat Spectral Data

26 Mapping Terrestrial Groundwater Dependent Ecosystems

b) Consistent photosynthetic activity based on: i) MODIS EVI standard deviation c) Contrasting midsummer photosynthetic activity i) Landsat NDVI threshold for selected summer images

6.2 Definition of Aridity Zones

Remote sensed data was analysed within each CMA area. Prior to analysis of the image data, the image data for each CMA region was stratified based on aridity zones because the effect of groundwater use on vegetation spectral response is expected to vary between more and less arid regions. Interpolated climate data grids of annual rainfall and evapotranspiration (Jeffry et al. 2001) were used to calculate Thornthwaite’s moisture index (aridity index) to give an index of the degree of water deficit below water requirement at any given location (Thornwaite 1948). Thornthwaite’s moisture index ranges from -1 for extremely dry/high potential evapotranspiration to +1 for wet/low evapotranspiration. The data has an approximate spatial resolution of 5 km but the resolution does not have a direct impact on the resolution of the GDE data product because the aridity zones were only used to guide stratification during the analysis.

Aridity thresholds for each CMA region were developed based on knowledge of the vegetation communities and cross- checked by visual inspection of the Landsat midsummer NDVI images and the MODUS EVI time series data. These images have been rectified to a national ground control framework and calibrated to a base image (Furby 2002) to ensure that any changes in image spectral values between seasons are due to changes in ground condition rather than changing atmospheric conditions or sun and sensor geometry.

The aridity index ranges used to define aridity zones were not the same for every CMA region. The reason being the distribution of aridity zones are a function of climate that is turn partially a function of topography. Six aridity zones were determined in the Corangamite CMA (Figure 3) region, with the most arid in the central portion of the CMA due to the influenced of the central highlands in the north and the Otway ranges and coastal influences in the south. Within dry flat landscapes such as the Mallee CMA, only one aridity zone existed.

A B Figure 3. Climate zones for the Corangamite CMA region derived from aridity indices (A) and comparing the climate zones to the MODIS EVI time series data for summer 2003 (B)

Mapping Terrestrial Groundwater Dependent Ecosystems 27

6.3 Remote sensing methodology

Three remote sensing measures were combined to identify potential GDEs. They are: 1. MODIS EVI standard deviation 2. Mid summer Landsat NDVI 3. Unsupervised classification of Landsat Spectral Data

The first measure identifies plants with higher photosynthetic activity than surrounding vegetation, persisting into mid-summer. The second measure is areas of relatively consistent photosynthetic activity through the year. The third identifies areas with spectral characteristics similar to vegetation communities with known ability to use groundwater. Each data set thus was divided into areas of potential GDE. An overlay model was used to combine the data sets as described in Section 6.4. The overall process logic is shown in Figure 4.

Define subregions for Develop Groundwater analysis based on Interactive Map showing Thornwaite’s Index of areas with water table Aridity generally < 5 m bgs based on existing GIS data & professional For each subregion: judgement

Calculate EVI Chose EVI s.d. cut-off standard deviation; value: combine (union) divide into 50 equal areas > cut-off classes Apply Groundwater Combine selected Interactive Map layer to classes for each remote subdivide potential sensing data set and run terrestrial GDEs into Classify midsummer Chose NDVI cut-off overlay model for a deep vs. shallow-rooted NDVI into 50 equal value: combine (union) model region (e.g. CMA areas classes for each areas > cut-off regions). Then chose image date cut-off value of model output: areas>cut-off value = potential Define attributes for Perform unsupervised Chose unsupervised terrestrial GDEs each GDE polygon classification for classes representative Landsat spectral data of GDEs, based on EVC within each subregion and known GDEs: (50 classes) combine (union) areas for those classes Screen potential terrestrial GDEs by land use classification

Assemble model regions into Statewide GDE layer for GIS

Figure 4. Process flow for development of potential terrestrial GDE map layer

6.3.1 Contrasting mid-summer photosynthetic activity

Areas with persistent green vegetation in mid-summer when precipitation is minimal indicate greater access to groundwater, surface water, or soil moisture. NDVI is a measure of the degree of photosynthetic activity by vegetation derived from the ratio of reflectivity in the infrared to the red wavelength regions. NDVI thus is correlated to the water used for photosynthesis. By itself, NDVI does not necessarily indicate groundwater use or dependency.

Data from two years with relatively low rainfall were chosen for analysis in this study, to minimize contributions from vadose zone soil moisture and surface water. One was in the mid 1990s at a time of relatively high groundwater levels and the second post 2000 at a time of relatively low groundwater level within Victoria, after groundwater levels had generally been in decline for at least five years. NDVI values from the Australian Greenhouse Office (AGO) time series data, composed of Landsat imagery as close as possible to mid summer, were obtained. As each AGO Landsat composite image is a mosaic of

28 Mapping Terrestrial Groundwater Dependent Ecosystems

cloud free images collected at different dates, the composite images were also stratified by image acquisition date as well as climatic zone prior to analysis.

It is expected that some ecosystems would only be able to access groundwater when the watertable is high while in other areas relatively high NVDI during drought highlights other communities depending on deeper groundwater for survival. In general the upper 30-50% of NDVI values in a climate zone for each wet and each dry year were selected. All selected areas for both years were treated equally and merged into a single data set. An example of the 1995 and 2002 combined output of mid summer photosynthetic activity in the North Central CMA region is shown in Figure 5.

6.3.2 Consistent annual photosynthetic activity

Vegetation communities in semi-arid or arid regions accessing groundwater for growth are expected to show more consistent growth throughout the year than communities relying on infiltrating precipitation. A time series of data from the MODIS satellite provided measurements of Enhanced Vegetation Index (EVI) every 16 days throughout the year 2003 chosen for analysis. In this instance the EVI was used in preference to the NDVI as Huete et al. (2002) reported that the EVI was more sensitive in high biomass regions than the NDVI and was less influenced by changes in atmospheric conditions. However the MODIS resolution of 250 m is coarser than Landsat.

A stack of 16 day composite EVI images for all of 2003 was created and the standard deviation for the time series calculated for each pixel over the whole year, using the image processing software ENVI. Plotting MODIS time series data can highlight the difference between vegetation that is only seasonally active as it relies on rainfall / surface water and vegetation that is likely to be accessing groundwater and therefore maintaining a relatively constant rate of photosynthetic activity. Figure 6 shows a plot of a EVI for native Eucalyptus camaldulensis (river red gum) with a relatively consistent value through the year compared to a more temporally variable pasture area.

It is important to note that a low EVI standard deviation value does not necessarily mean consistent photosynthetic activity. The standard deviation layer will also include areas of low standard deviation that are not likely to contain a GDE. For example • dry lake beds • anthropogenic structures (roads and buildings etc) • the layer is unable to differentiate photosynthetic activity within forests, such that nearly all forested regions were highlighted.

However, the EVI standard deviation, combined with the NDVI data set, defines low variability, green vegetation. The Landsat NDVI dataset was preferred to simply applying a threshold on the EVI data because of the superior Landsat resolution.

The EVI standard deviation data range for each aridity zone was divided into 50 equally-spaced classes (2% of the data range for each class). All classes were compared with known GDEs and a cut-off threshold determined, separating classes consistent with the potential presence of GDEs from those indicating unlikely presence of GDEs. Pixels with the lowest 40 to 60% of standard deviation classes were generally selected in the CMA wide models. This Boolean classification was then carried forward to integrate with the other remotely sensed data sets to identify potential GDEs.

It was generally observed that the lowest 40-50% of the standard deviation value range encapsulated vegetation around wetlands and along streams and river stretches and deeper rooted vegetation (forested areas), but the selection was based on professional judgement for each aridity zone. For example, within the North Central CMA region the lowest 27 classes were selected in each climatic zone (Figure 7). Areas in white are regions of higher standard deviation and are therefore not included in the model. The white areas are generally grasslands not expected to use groundwater. The regions with low standard deviation (reds and oranges) are consistent with vegetation around major rivers and wetlands, forested regions, and in places irrigation districts, where photosynthetic activity remains constant throughout the year.

Mapping Terrestrial Groundwater Dependent Ecosystems 29

Figure 5. Mid summer NDVI for the combined years of 1995 and 2002 in the North Central CMA region

30 Mapping Terrestrial Groundwater Dependent Ecosystems

Figure 6. Time series of MODIS EVI values showing consistent annual photosynthetic activity of redgum forest in comparison to variable annual photosynthetic activity of pasture

Mapping Terrestrial Groundwater Dependent Ecosystems 31

Figure 7. EVI standard deviation layer for the North Central CMA region, showing the 27 classes with the lowest standard deviation

32 Mapping Terrestrial Groundwater Dependent Ecosystems

6.3.3 Unsupervised classification of Landsat spectral data

Except for to specific locations, such as Murray River flood plain, Eucalypt and Melaleuca wetlands, very little is known of native vegetation groundwater use within the Victorian Landscape. Therefore, insufficient field sites exist to act as training data for remote sensing analyses. The complex and fragmented nature of vegetation communities across the landscape coupled with the lack of detailed and specific training sites, mean that unsupervised classification is suited to the task of breaking the Landsat data into spectrally similar groups (Conrad et al. 2005; Harvey and Hill, 2001). After classification the classes spatially associated with vegetation types suggestive of possible GDEs were identified. Those classes then were considered also to represent potential GDEs in other locations.

Ecological Vegetation Classes (EVCs) thought likely to be dependant on groundwater to some degree were used to identify the Landsat spectral classifications representing potential groundwater dependency. EVC classes selected (Table 3) included generic wetlands, and swamp species such as Eucalyptus camaldulensis and Phragmities australis, halophytic vegetation, amongst others. The distribution of EVCs likely to be partially dependant on groundwater is shown in Figure 8.

Within each climatic zone and image date, the Landsat data was grouped into 50 spectrally similar clusters using a K Means algorithm. In contrast to the other remote sensing layers, the unsupervised classification clusters do not form an ordered continuum of response. Thus any group, not just the highest or lowest numbered ones may relate to the presence of GDEs. The resulting unsupervised classifications were intersected with areas attributed with Ecological Vegetation Classes (EVCs) thought likely to be dependant on groundwater to some degree. The spectral groups that intersected best with the selected EVCs were identified first by area and second by polygon count. The extents of the highest ranked spectral groups were then identified spatially across each climatic zone. Table 3. EVC groups and examples used to develop the indicative response layer EVC Group Examples landscapes Riparian Forests or Woodlands Vegetated streams within the Victorian high country Riparian Scrubs or Swampy Scrubs and Lake Condah Woodlands Riverine Grassy Woodlands or Forests Murray River Red Gums Salt Tolerant and / or succulent Shrublands Lake Tyrrell

6.4 Integration of data sets and development of Model layers

The GDE maps were created by an overlay model within ARCGIS. Because each layer was classified into true/false classes indicating GDE attributes prior to application of the overlay model, the model effectively calculates intersections of the layers. By weighting each data set into distinct ranges, the different intersection areas can be displayed (equivalent to binary arithmetic). An example weighting is shown in Table 4. The intersection of all three remote sensing data sets was used to define potential GDEs in the model. The combination of Contrasting Photosynthetic Activity and Consistent Photosynthetic Activity represents vegetation that is green year-round and includes extensive heavily-treed areas in the highlands that are unlikely to be groundwater dependent but are excluded from the potential GDE maps. It should be noted that some heavily treed riparian areas also appeared in this combination and could show some groundwater dependency. However, the potential GDE map still identifies areas of interest within the riparian zones and those areas are considered the most important targets for future local or catchment area studies.

After integration of the remote sensing data sets, the results were filtered to remove areas of land use that is incompatible with GDEs, such as irrigated agriculture, residential, and industrial areas (Table 5). The classification levels in the statewide 1:100,000 scale land use classification GIS layer was used to mask incompatible land use from the final map layers. Figure 9 shows the land use classification in an area of the North East CMA region and the potential GDEs after application of the land use mask. Note the removal of potential GDEs within Wangaratta and in areas of irrigated agriculture northeast of town.

Mapping Terrestrial Groundwater Dependent Ecosystems 33

Figure 8. Distribution of selected EVC groupings used in the unsupervised classification in Victoria. Black outline defines the Glenelg-Hopkins CMA region.

34 Mapping Terrestrial Groundwater Dependent Ecosystems

Table 4. Example weighting factors for use in model combining remote sensing GDE classes

Contrasting Consistent Unsupervised Total Photosynthetic Photosynthetic Classification Activity Activity of Landsat

(based on (based on std. NDVI) dev. of EVI)

Contrasting Photosynthetic Activity & 1 10 0 011 Consistent Photosynthetic Activity

Contrasting Photosynthetic Activity & 1 0 100 101 Unsupervised Classification

Consistent Photosynthetic Activity & 0 10 100 110 Unsupervised Classification

Contrasting Photosynthetic Activity & 1 10 100 111 Consistent Photosynthetic Activity & Unsupervised Classification

Mapping Terrestrial Groundwater Dependent Ecosystems 35

Figure 9. Application of land use mask to potential terrestrial GDEs. A. Land use classification for section of North East CMA region. B. Potential terrestrial GDEs before and after application of mask.

36 Mapping Terrestrial Groundwater Dependent Ecosystems

Table 5. Land use classes used to exclude areas from classification as potential terrestrial GDEs

Land Use Land Use Description Remove GDEs within Comments Code these areas?

1.0 Conservation and No Natural Areas

2.0 Production from No Relatively Natural Environments

3.0 Production from No May include areas of natural habitat Dryland Agriculture

4.0 Production from Yes Mapped irrigated areas may not Irrigated Agriculture correspond to irrigated paddocks in and Plantations different years

5.1 Intensive Horticulture Yes

5.2 Intensive Production Yes

5.3 Manufacturing and Yes Industrial

5.4 Residential Yes

5.5 Services No Includes some natural recreational areas

5.6 Utilities Yes

5.7 Roads Yes

5.8 Mining Yes

5.9 Waste Treatment and Yes Disposal

6.0 Water No Includes some riparian areas

The application of the land use mask is somewhat problematic due to the nature of the land use classification. Firstly, areas such as irrigated paddocks are dynamic in nature and the land use layer may not identify the irrigated lands present during the time period of the remote sensing data. Secondly, the scale of the land use classification is coarser than the remotely sensed layers. In addition, GDEs may be present in some places with land use designated for removal from the potential GDE maps, such as along road margins, in urban areas, and at the margins of agricultural areas. However, the mask is useful at the CMA scale for identifying the most likely areas of GDE resources.

6.5 Classification and Attribution of GDEs

The remote sensing data sets identify potential GDEs but further geographic information is needed to classify GDEs and provide attributes relating to the GDE type. The state-wide methodology for identifying terrestrial GDEs relies predominantly

Mapping Terrestrial Groundwater Dependent Ecosystems 37

on the spectral response of vegetation that uses water consistently through the year, has high mid-summer water use, and exhibits the characteristics of vegetation communities known or expected to be capable of using groundwater. As such, the areas also include wetland areas and in some cases lakes or other water bodies. Other attributes that are important for evaluation and management of GDEs include the geomorphologic setting, surface geology type, and the presence of saline groundwater.

The association of potential GDEs with wetland areas and water bodies is presented in this report by showing an overlay of the mapped wetland areas with the terrestrial GDE layer. In theory, for any wetland that intersects areas of groundwater discharge, the whole wetland may be groundwater dependent – i.e. the wetland would dry up or be substantially reduced in area if the groundwater source was removed. In the figures shown in Section 7, the water body types are not separated, although certain artificial water bodies such as sewage lagoons are filtered out. It is important to note that the ‘water body’ classification includes areas identified as ‘flat land subject to inundation’. These areas may or may not be persistent wetlands and need to be carefully evaluated in assessing the groundwater dependency of associated ecosystems.

In areas that do not have surface water, the typical depth to groundwater is important for evaluating the types of vegetation that could be using groundwater and the likelihood of groundwater dependency. Groundwater in areas with water tables at depths > ~5m are less likely to be accessed by vegetation as few species are expected to have the ability to develop sufficiently deep roots. The lack of sufficient bore coverage in many areas of the state, particularly in areas of natural vegetation limits the creation and use of accurate statewide maps of depth to groundwater. Fluctuations in watertable depth over time also inhibit the use of these maps in assessment of groundwater dependency. Due to these issues, the available data were integrated into a ‘Groundwater Interactive Map’ (Section 6.5.1) that was developed to show areas where vegetation would likely be able to access groundwater.

The locations of springs are also of interest in assessing terrestrial GDEs, as well as representing GDEs in their own right. Springs indicate surrounding areas may have shallow groundwater (subject to topographic effects) and thus potential GDEs. Spring locations are presented as an overlay on map figures, as appropriate.

6.5.1 Groundwater Interactive Map

By definition all groundwater dependent ecosystems must interact with groundwater. Land surface based GDEs that reside within wetland, river and coastal systems must directly interact with the water table by diffuse or point discharge processes. Determining the landscapes that most likely have wetland, river and coastal systems requires mapping of groundwater discharge into surface systems. However, terrestrial GDEs are slightly different, as the root systems can interact with shallow groundwater systems where groundwater does not contact the land surface.

A groundwater interactive map is similar to a map of depth to watertable, but instead of interpolating groundwater data points across a landscape, a groundwater depth range is assigned to mappable geological and geomorphological units. A groundwater depth range of 0-5m below natural surface has being in this instance. Traditional depth to watertable maps represent a system at a given time, with the accuracy of the map controlled by the distribution of the groundwater monitoring bore network and the timing of the monitoring. In many situations the elevation of the groundwater surface is a function of the climate. The groundwater interactive map created in this project is a more flexible product. It is not designed to define the precise nature of the groundwater surface, but rather provide a spatial sense of where groundwater is likely to exist within the given depth range. The groundwater interactive map does not directly take into account the variation in groundwater levels, but is designed to be inclusive of landscapes that range from permanently artesian (discharging) to landscapes that periodically have perched and/or shallow watertables.

The first step in developing the groundwater interactive map was to collate all mapped geo-spatial data sets that suggest groundwater interaction and or saturated, inundated environments. In general the rivers, streams, wetlands, and low lying regions within Victoria occur upon Quaternary geology (such as alluvial, colluvial and aeolian deposits and the Newer Volcanic Basalt flows). Pre-Quaternary geology generally contains drainage lines of alluvial and colluvial sediments, however, due to the scale of mapped geological landscapes (e.g. Devonian Granites), the integration of additional data sets is required to identify where groundwater may come close to the surface. Additional information was incorporated into this GIS layer from:

• Stream base flow index, that inform on groundwater interaction with streams. Nathan and Weinmann (1993) derived base flow indices for 39 stream basins that contained 117 stream gauges across Victoria • Known groundwater interactive environments determined through literature search

38 Mapping Terrestrial Groundwater Dependent Ecosystems

• Landscapes that have shallow water tables based upon measured groundwater levels • Hydrological units, wetlands, streams, swamps and vegetation indicative of saturated, inundated environments • Saline discharge layer, alternative data set that indicates groundwater interaction with the soil surface.

The resulting groundwater interactive map provides a base layer that represents environments that will most likely contain shallow groundwater, groundwater discharge and thus GDEs. The North Central CMA groundwater interactive map (Figure 10) is provided as an example. Within the mapped groundwater interactive landscape, 87% of all piezometers that have recorded a depth to watertable of <5 m below natural surface fall within the groundwater interactive classed landscapes. The GIM is used to distinguish between areas that may contain shallow rooted GDEs (e.g. marshlands) and those areas that are most likely to have deep rooted vegetation (e.g. woodlands) and, therefore, fall outside the groundwater interactive landscapes.

6.5.2 Geomorphologic, Geologic, and Chemical Attributes

Broad hydrogeological characteristics assigned to potential GDE locations are provided to assist in the discussion of associated hydrological and hydrogeological processes (Table 6). These attributes are viewed as a starting point for the developing of more detailed assessment of GDEs within Victoria. Until more detailed studies are conducted to determine the groundwater use component of native vegetation, regional studies such as this one can only provide generalised outputs.

Statewide Geomorphologic Management Units (GMU), developed by the Department of Primary Industries in consultation with the Geomorphology Reference Group were used as attributes for GDEs (CGDL, O’Brien 2004). The GMU system is a three tier classification with tier one having eight units, tier two having 34 units and tier three having 95 units. The GMU layer was chosen as it has statewide coverage, and its depiction of geomorphology has strong relationships to the groundwater flow system (GFS) approach (Coram et al. 2000; Walker et al. 2003).

Groundwater salinity was used as an attribute, because the quality of groundwater can influence the species that have evolved to use it, such as halophytes within saline springs.

The generalized surfacial geology is provided as additional landscape information for evaluation of potential GDEs. The surficial geology will influence the groundwater flow system and the water quality.

GDE polygons may intersect more than one attribute type so additional columns are included to capture those occurrences.

Mapping Terrestrial Groundwater Dependent Ecosystems 39

Figure 10. Indicative groundwater interactive map (GIM) for the North Central CMA region showing the location of piezometers that have recorded a watertable depth of less than 5 m. Colour coding shows five different percentile ranges indicating percentage of piezometer record for which the water table was less than 5 m below ground surface

40 Mapping Terrestrial Groundwater Dependent Ecosystems

Table 6. Attributes assigned to each potential GDE location Column Attribute Class Symbols or Symbol Column Alias Heading Description Values Description/Meaning Groundwater Water table generally Groundwater 8 expected to be < 5m below surface GRIDCODE interactive accessible by Water table generally designation 9 vegetation > 5m below surface Low Salinity Class (Max L <=1000 TDS) Groundwater GW Salinity Moderate Salinity Class GW_TDS Salinity Level in Class M (Max 1001 to GDE Area 19,999TDS) High Salinity Class (Max H >=20,000 TDS) Sedimentary (includes S Newer Volcanics) Surface Rock Type Surface I Intrusive SUR_GEOLGY (Geology) in GDE Geology Type Area M Metamorphic Unclassified (many are U in water bodies)

Geomorphic Various, all in Each CMA includes a Management Unit the form GMU GMU Class different array of GMU (GMU) in GDE X.X.X Where classes Area X is a #, 0 to 9

A description of the A written specific Geomorphic GMU Class description of the Various pieces DESCRIPTIO Management Unit ‘Description’ GDE’s ‘GMU of text (GMU) Class associated Class’ with each GDE.

Mapping Terrestrial Groundwater Dependent Ecosystems 41

7 CMA region potential GDE map results

This section presents an overview of the potential terrestrial GDE mapping results and some brief examples of areas of possible interest within each CMA region. It should be emphasized that the interpretation is not based on detailed evaluation of the remote sensing response or field checking. Site specific investigation would be needed to establish groundwater dependency. It is hoped this section illustrates the strengths and limitations of the results for initial state-wide management considerations.

7.1 Mallee CMA

The major areas of potential GDE shown in the Mallee CMA region are associated with the mallee vegetation in the west- central part of the CMA region (Figure 11 B). The consistent photosynthetic response of these ecosystems, the lack of surface water, and the presence of shallow groundwater in at least some areas suggests that groundwater may be a significant contributor to ecosystem health. Further east (Figure 11 C) the mapped areas are sparse and, interestingly not strongly associated with defined wetland or surface water bodies, perhaps due to salinity impacts. It is possible that some sparse vegetation between the water bodies/salinas is accessing groundwater but does not sufficiently exhibit the spectral response of green vegetation to be included in the maps. Groundwater levels are known to be high across much of this area, particularly the Raak Plains, west of Hattah (western part of Figure 15 C), which is a regional groundwater discharge zone (Macumber 1980, 1991).

7.2 Wimmera CMA

The Wimmera region presents a transition from the forested Grampian mountains through extensive dryland farming areas to the Mallee in the north. Although the major identified areas of potential GDE are associated with forests and woodlands, more detailed examination shows the greatest presence is associated with water bodies and wetlands within each general area. Of particular note is the association of potential GDEs with the lakes and ponds near Edenhope (Figure 12 B). Groundwater contribution to maintaining those surface water bodies has been established by Fawcett and Huggins (2005). Areas of potential GDE are also found in the Little Desert in the west-central Wimmera. Figure 12 C shows an increasing signature as the Wimmera River is approached near Dimboola. This could be due to increased shallow groundwater occurrence associated with regionally high groundwater pressures near the river.

7.3 Glenelg Hopkins CMA

The Glenelg Hopkins CMA region includes areas of limestone karst with solution enhanced permeability and karst springs ranging to forested upland clastic sediments of the southern Grampians. The greatest extent of potential GDE is found in the west and near the southern Grampians (Figure 13). The major areas correlate well with groundwater interactive landscape areas of generally shallow water table depth. Interpretation of the Grampian landscape is complicated by the effect of the slope-aspect on the remote sensing response. However, the major mapped areas appear along river valleys and at the periphery of the highlands where groundwater discharge may be expected (Figure 13B). Figure 13C shows the correlation of potential GDEs with shallow water table near Portland.

A series of saline wetlands, the Cockajemmy Lakes (Figure 14), was used to test the remote sensing methodology and, with the benefit of good groundwater data, serves to illustrate the potential for application of the mapping to smaller areas. The Cockajemmy Lakes form along the southern margin of a Quaternary age basalt flow (Newer Volcanics). Groundwater pressures measured in piezometers are shallow (<2 m below natural surface) and are artesian around the lake perimeters. Land use around the lakes is dominated by cropping and the region is mostly devoid of trees. The map of potential GDEs within the Cockajemmy Lakes region centres on vegetation growing in association with wetlands. It is therefore very likely the vegetation has a high degree of interaction with groundwater.

42 Mapping Terrestrial Groundwater Dependent Ecosystems

Figure 11. Potential terrestrial GDEs in the Mallee CMA. A. Overview. B. GDEs mapped in area of extensive mallee vegetation. C. Sparse association of mapped GDEs with lakes, salinas and areas of shallow water table

Mapping Terrestrial Groundwater Dependent Ecosystems 43

Figure 12. Potential terrestrial GDEs in the Wimmera CMA. A. Overview. B. Area of known groundwater-surface water interaction near Edenhope. C. Area of potential GDEs near the Wimmera River at Dimboola

44 Mapping Terrestrial Groundwater Dependent Ecosystems

Figure 13. Potential terrestrial GDEs in the Glenelg Hopkins CMA. A. Overview. B. Detail of GDEs shown around the southern Grampians. C. Potential GDEs and spring locations northwest of Portland.

Mapping Terrestrial Groundwater Dependent Ecosystems 45

Figure 14. Map of the Cockajemmi Lakes region showing potential GDEs and the extent of the Quaternary age basalt flow

7.4 North Central CMA

The potential terrestrial GDEs within the North Central CMA region are shown in Figure 15. The major areas of identified potential GDEs are along the Murray River and Kerang Lakes area in the north, and in upland forested areas in the south and central parts of the CMA region. Additional areas are found along other rivers and ephemeral streams.

The detail in Figure 15B (Kerang Lakes area) shows the occurrence of potential GDEs in areas mapped as groundwater interactive along sections of drainage lines and surrounding water bodies. This area is well documented as a regional groundwater discharge zone (e.g. Macumber 1980, 1991, 2004). Although much of the area is designated wetland or subject to inundation, the mapped GDEs are restricted to relatively small parts of the overall wetland area, targeting locations for more detailed evaluation.

46 Mapping Terrestrial Groundwater Dependent Ecosystems

It is felt that much of the upland forest area identified as potential GDE is the result of over-classification due to the consistent green response of the canopy. However Figure 15 C shows the prevalence of mapped springs (Shugg and Brumley 2003, Shugg 2009) in the southern part of the CMA region near Daylesford. These springs tend to occur at the periphery of the mapped GDE areas and indicate the potential for near-surface groundwater sources that could be used by surrounding vegetation or important groundwater contribution to the local surface water.

Figure 15. Potential GDE map for the North Central CMA region. A. Overview of the CMA region. B. Area between Swan Hill and Kerang showing the association of potential GDEs in areas where surface water drainages or wetlands intersect groundwater interactive landscapes. C. Relation of GDEs to spring locations in the vicinity of Daylesford.

Mapping Terrestrial Groundwater Dependent Ecosystems 47

The Tarpaulin Creek region, south of St. Arnaud, was evaluated in detail during development of the GDE mapping process (Figure 16). Interpreting the Tarpaulin Creek region GDE map and accuracy is somewhat problematic as the interaction of groundwater with vegetation is not very evident or well understood due to little groundwater data and the absence of previous research.

Based on the limited groundwater level data and known incidences of salinity in this area (Clark and Harvey 2008), it is reasonable to assume that vegetation around the ephemeral streams will have some connection with groundwater. These areas have been highlighted as potential GDEs that could contain shallow rooted species (Figure 16, cyan). However, landscapes highlighted in yellow are potential GDEs that fall outside the GIM landscapes and are assumed to contain deep rooted vegetation; determining the validity of this assumption requires site specific field assessments.

Figure 16. Map of potential GDEs for the Tarpaulin Creek area, showing landscapes that may contain GDEs with shallow rooted vegetation (cyan) and landscapes which might include deep rooted vegetation and fall outside the indicative groundwater interaction map (yellow)

7.5 Corangamite CMA

The Corangamite CMA region spans a variety of landscapes from forested uplands near Ballarat and in the Otway Ranges to coastal areas (Figure 17). The largest areas of potential terrestrial GDE are found in forested areas near the coast. Water

48 Mapping Terrestrial Groundwater Dependent Ecosystems

tables are generally shallow in this area but interpretation is complicated by the coastal effects of increased precipitation and mitigation of seasonal temperature changes. The areas mapped in the north tend to be associated with upland forests (Figure 17B) but the presence of mapped springs and potential GDEs associated with areas of shallow water table (Nicholson et al. 2006) suggest some further investigation may be warranted. A number of potential terrestrial GDEs are seen in the area of lakes near Colac (Figure 17C), where groundwater-lake interaction and salinity are known to occur (Dahlhaus et al. 2002; Nicholson et al. 2006). It is reasonable to suspect that the vegetation in these areas accesses groundwater and may be of interest for resource management.

Figure 17. Potential terrestrial GDEs in the Corangamite CMA. A. Overview. B. Detail showing areas of potential groundwater dependency east and south of Ballarat. C. Potential terrestrial GDEs around lakes near Colac, including Lake Corangamite

Mapping Terrestrial Groundwater Dependent Ecosystems 49

7.6 Port Phillip and Westernport CMA

The results for the Port Phillip and Westernport CMA region are complicated by the extensive areas of urban development and intensive land use and agriculture. Most of the potential GDEs area is associated with upland forest. Significant area of potential GDEs is identified in Mornington Peninsula and outer suburbs of Melbourne. A few locations of interest along drainages and rivers can also be seen. Little indication of significant area of terrestrial GDEs is seen in the areas with intensive land use and agriculture in the region (e.g. Koo-Wee- Rup area). This probably reflects negative impact of intensive agriculture on the health of GDEs. One example of potential terrestrial GDE is in the area to the south and west of Bacchus Marsh where the mapped potential GDEs are associated with areas of shallow water table near forested uplands (Figure 18B). A second example is along the Yarra River (Figure 18C) near Healesville, where the potential GDEs are associated with cut off river meanders (Figure 19).

7.7 Goulburn Broken CMA The overall pattern seen in the location of potential terrestrial GDEs within the Goulburn Broken CMA region is similar to that of the North Central CMA region (Figure 20). The largest patches are associated with forested areas in the uplands or along the Murray River. In particular, most of the Barmah Forest shows up as potential terrestrial GDE (Figure 20B). This is reasonable given that (i) it occupies a large effluent floodplain of the Murray River on the downthrown side of a major fault (Cadell Fault), (ii) it is known to have good quality groundwater at shallow to moderate depths, and (iii) there is recent, good scientific evidence of groundwater use by the dominant red gum tree species during a dry climate period with limited flooding (Shaun Cunningham, Monash University, pers. comm. 2009). Notable additional areas of potential GDE are seen along the riparian corridors of the Goulburn River and Broken River.

7.8 West Gippsland CMA

Large areas of potential terrestrial GDEs are mapped in the West Gippsland CMA region (Figure 21). However, much of the area is associated with upland or low forest and thus the groundwater usage by vegetation is quite uncertain. Some areas of potential terrestrial GDE are found along waterways and thus may be ecosystems of interest accessing shallow groundwater (e.g. Figure 21B and C). More detailed work is needed to optimize the GDE mapping for this area before making definitive statements regarding groundwater use.

7.9 North East CMA

Some results for the North East CMA region were presented during description of the method development, above. Extensive areas of high elevation forests with high precipitation are present in the North East and East Gippsland CMA regions. These areas are not as subject to limitations of precipitation and surface water as others so they tend to appear similar to potential GDEs. In addition to areas with response in only one of the data sets, areas with response in only the consistent photosynthetic activity (low standard deviation of EVI) and contrasting photosynthetic activity (high midsummer NDVI) were excluded from the potential GDE map in the North East and East Gippsland CMA regions. Those areas correspond to value 011 in Table 4. An additional issue is that the varying slope/aspect in the hilly areas affects the Landsat spectral response used in the unsupervised classification. This leads to differences in classification depending on the direction and steepness of the slope. It would likely be possible to mitigate this effect through stratification of the data according to slope and aspect or illumination angle prior to classification.

Figure 22 shows that large upland areas in the North East CMA region are classified as potential terrestrial GDEs but are thought most likely to be only recording the high rainfall forest cover. However, mapped potential GDEs in the Ovens and Murray River floodplains (e.g. Figure 23 B) are considered to have a greater likelihood of being at least partially dependent on groundwater.

50 Mapping Terrestrial Groundwater Dependent Ecosystems

Figure 18. Potential terrestrial GDEs in the Port Phillip and Westernport CMA. A. Overview. B. Detail showing areas of potential groundwater dependency along streams near Bacchus Marsh. C. Potential GDEs in the Yarra River riparian zone near Healesville

Mapping Terrestrial Groundwater Dependent Ecosystems 51

Figure 19. Satellite image showing potential terrestrial GDEs along Yarra River

52 Mapping Terrestrial Groundwater Dependent Ecosystems

Figure 20. Potential terrestrial GDEs in the Goulburn Broken CMA. A. Overview. B. Detail of GDEs in the Murray Riverfloodplain (large green area is the Barmah Forest), along the lower Goulburn River (lower left) and along other streams including the Broken Creek

Mapping Terrestrial Groundwater Dependent Ecosystems 53

Figure 21. Potential terrestrial GDEs in the West Gippsland CMA. A. Overview. B&C. Detail showing areas of potential groundwater dependency along streams and waterways in the Latrobe Valley (B) and north of Lake Wellington (C)

54 Mapping Terrestrial Groundwater Dependent Ecosystems

Figure 22. Potential terrestrial GDEs in the North East CMA. A. Overview. B. Detail showing areas of potential groundwater dependency along the Murray and Ovens River floodplains

7.10 East Gippsland CMA

The East Gippsland CMA region has similar issues with respect to higher precipitation and forested mountain areas as the North East CMA region. Thus, more restrictive criteria were also applied in the integration of the remote sensing data sets. In addition to areas with response in only one of the data sets, areas with response in only the consistent photosynthetic activity

Mapping Terrestrial Groundwater Dependent Ecosystems 55

(low standard deviation of EVI) and contrasting photosynthetic activity (high midsummer NDVI) were excluded from the potential GDE map in the North East and East Gippsland CMA regions. Those areas correspond to value 011 in Table 4. Still, large areas show scattered response due to the slope/aspect effect on the unsupervised classification as discussed for North East CMA region. The lowland/coastal areas generally show scattered areas of potential GDE that often correspond to open water or more heavily treed areas. However a few areas of potential GDE are seen in areas at the periphery of of drainage lines and water bodies/wetlands (e.g. Figure 23B). Inspection of satellite imagery shows that these are more open vegetation, not forest or woodland. It is reasonable to consider the possibility that they are areas of groundwater use given the setting and commonly shallow water tables.

56 Mapping Terrestrial Groundwater Dependent Ecosystems

Figure 23. Potential terrestrial GDEs in the East Gippsland CMA. A. Overview. B. Detail showing potential groundwater dependency in areas surrounding drainage lines and water bodies and other scattered areas. The large blocks are responding to open water

Mapping Terrestrial Groundwater Dependent Ecosystems 57

8 Conclusions and Recommendations

The GDE mapping method and GIS products produced under this research project are innovative developments for water resource management across the state of Victoria. Protection of GDEs is a stated priority of both National and State groundwater management programs. However, a sound assessment of where GDEs are likely to occur in the landscape has not existed previously. With the GDE mapping data and supporting information now available through this project, state and regional agencies charged with the responsibility of protecting GDEs have a starting point from which to work.

While the maps provide for the first time prediction of where GDE sites are possible in the landscape at a regional scale, it is clear from the study that there are myriad different GDE types and levels of groundwater dependency. Many of the locations identified by the mapping process are likely to contain well adapted, water-efficient, indigenous vegetation that is able to draw on shallow or deep water sources and, so, able to maintain a perennial ET profile. There are locations that would be best described as ‘springs’ or ‘soaks’, places in the landscape where groundwater naturally emerges and supports more abundant, perennial vegetation types that are naturally more dependent upon groundwater. Importantly, the maps show where GDEs are not present in the landscape, ruling out large areas from having to be considered for protection management in relation to groundwater.

The methodology developed is expected to overestimate the extent of terrestrial GDEs. There are locations that appear to fulfil the definition of a GDE (as defined by the mapping model) and are identified as a potential GDE landscape, but may not be using groundwater. Two prominent examples of such are:

1. Riparian zones along sections of rivers and creeks that have deep water tables where the stream is a losing stream and the riparian vegetation is able to access water, either from bank storage, or by intercepting leakage from the stream in the unsaturated (vadose) zone above the groundwater system.

2. Forested regions that are accessing large unsaturated regolith water stores or where the vegetation type does not show significant photosynthetic response to variations in water availability. In general upland forests and woodlands tend to be identified as potential GDEs but it is considered most unlikely that they are universally accessing groundwater.

8.1 Limitations of the maps

In developing this first generation of regional GDE maps for Victoria, several limitations of the method were identified.

The remote sensing analysis aims to identify differences in vegetation response to water availability. It relies on a good understanding of the ground processes and climate so that some inferences can be made about moisture availability. The current maps provide an understanding of landscapes that potentially use groundwater, however without ground data, actual rates of groundwater use cannot be determined. The GIS layer produced does not contain information regarding the degree of dependency on groundwater or the amount of groundwater used. The maps should not be used as an absolute assessment of where GDEs exist or how much groundwater is being used by terrestrial vegetation, unless integrated with site specific hydrogeological and plant water use assessments.

Field checking (qualitative or detailed quantitative measurements made at specific sites) has not yet occurred and is not part of the scope of this project. This is highly recommended and would help greatly to further develop the remote sensing data sets and improve confidence in the output maps. Such work can be strongly aligned to high priority GMAs and Victorian Flagship areas for biodiversity protection.

As discussed within this report, groundwater use by vegetation may be perennial, seasonal, or restricted to dry periods. Even for the same species, the amount of groundwater use will vary depending on specific circumstances. Factors that may affect the amount of groundwater required include the occurrence of summer storms, local groundwater system relationships, groundwater salinities, plant physiology, general health of the vegetation and unsaturated regolith water stores.

It is also possible that due to any number of environmental conditions and influences, a GDE region may be under stress and will not have the normal, constant growth patterns for the year in which the standard deviation was determined.

58 Mapping Terrestrial Groundwater Dependent Ecosystems

Remote sensing cannot see below the uppermost surface, whether this is a thick tree canopy or grasses at ground level. Depending on vegetation density, the response is determined by the uppermost canopies. If any reflected radiation comes from below this canopy, whether it be bare soil or lower levels of vegetation, then it is blended with the uppermost canopy. The current maps need to be used in association with a field based understanding of the vegetation assemblage in question. It is possible that some small surface water features will be masked out along the smaller streams due to overhead canopy and there will, instead, be reliance upon on identifying photosynthetic activity to infer the availability of water.

The current generation of Victorian GDE maps was developed by stratification of climate zones within each CMA boundary, rather than using the biophysical zone boundaries (for example, the basalt plains biophysical zone crosses CMA boundaries). It may be of benefit to undertake a trial that stratifies the climate zones within the biophysical zone boundaries, as each type of land-systems could be expected to have a similar response. By ignoring these differences, we might be mixing responses and losing some details the method could have otherwise identified.

While MODIS is useful for its temporal resolution and shows the variation in vegetation response over time, the 250m spatial resolution means that smaller features cannot be identified. Also, the compositing process introduces some anomalies that should be removed before calculating standard deviation for each pixel (i.e. cloud and high zenith view angle).

In regards to the Landsat data, it was not possible to remove the effect of shadow associated with cloud cover across large areas or the impact of small scale storm events within the Landsat images. This resulted in some areas requiring the use of images flown on different dates, leading to edge effects and linear artefacts in the processed image. Care is needed in the Landsat image date selection process to reduce the Landsat run boundary edge effects.

Forest areas are also problematic, particularly if they are in more elevated and dissected terrain. In some forests, outputs appear to be related to be a product of landscape or slope rather than a vegetation response. This could be an artefact of solar illumination at 10am (approximately the time that the satellite passes over) on the hills and valleys. Alternatively it could be the effect of topography on soil development, where greater moisture availability occurs on southern and eastern slopes, producing higher biomass and leading to formation of a deeper soil profile compared to the drier northern and western slopes. These issues need to be explored in future project development.

8.2 Recommended usage of the maps in current form

The GIS layer produced can be used with other spatial data to gain an understanding of where GDEs may be present. Using the data layer as an initial guide, other data such as air photographs, geology, geomorphology, vegetation type, surface hydrology, groundwater flow system or soil mapping can be used to focus in on specific features of interest. In particular, expert inspection of air photo or satellite imagery and topographic and geology or GFS maps is recommended to screen out upland forested areas that are unlikely to access groundwater. As an initial step, the most likely, and potentially highest-value GDEs are expected to occur in the groundwater interactive landscapes (generally shallow water table) so those areas should be the focus. Figure xx shows a section of the North East CMA region, classified by the groundwater interactive attribute. The shallow water table class excludes areas of upland forest and highlights the areas of greater interest.

One of the most likely scenarios for the need to regulate and protect an identified GDE asset is groundwater extraction. Overlaying the mapping product on GFS and groundwater management data (bore locations, pumping volumes, groundwater depths, management area boundaries etc.) can reveal whether the identified GDE feature is likely to be at risk to groundwater level decline as a result of human activity.

The maps can be used as a field guide for inspectors or diversions officers when assessing groundwater licenses for both regulatory and environmental authorities. For example, in the case of an extraction license application adjacent to an identified GDE, the inspector can ask for an assessment of the potential impact as part of the license requirement. Finding cost- effective and sufficiently reliable means of achieving this within the license approval process will need to be investigated.

The maps provide the first cut prediction of where groundwater is predicted to support surface ecology. They also provide additional information to assist the development of management plans aiming to preserve key environment assets.

Mapping Terrestrial Groundwater Dependent Ecosystems 59

The maps form a basis for more detailed catchment or local scale evaluation of the presence and sensitivity of potential GDEs. The remote sensing methodology can be optimized for particular catchments by more detailed analysis of aridity zones, vegetation characteristics (including tolerances), iterative adjustment of classification levels for the data sets, more detailed land-use analysis, and by development of training data from local field studies.

Figure 24. Potential terrestrial GDEs in a section of the North East CMA region, classified by groundwater interactive landscape. The areas of greatest interest are expected to be in the shallow water table zones

60 Mapping Terrestrial Groundwater Dependent Ecosystems

9 References

Acworth, R., Hughes, C. and Turner, I. (2007) A radioisotope tracer investigation to determine the direction of groundwater movement adjacent to a tidal creek during spring and neap tides. Hydrogeology Journal 15, 281-296. Adam, P. (1994) Saltmarsh and mangrove. In Australian Vegetation, 2nd Edition, ed. R.H. Groves, 395-435. Cambridge Univ. Press. Almendinger, J. and Leete, J. (1998) Regional and local hydrogeology of calcareous fens in the Minnesota River basin, USA. Wetlands 18, 184-202. Bacon, P. E., Stone, C., Binns, D. L., Leslie, D. J., and Edwards, D. W. (1993) Relationships between water availability and Eucalyptus camaldulensis growth in a riparian forest. Journal of Hydrology 150, 541-561. Barton, A., Dalhaus, P., Davies, P. and Cox, J. (2006) Groundwater Dependent Ecosystems in the Corangamite Catchment Management Authority Region. CSIRO, CSIRO Land & Water Science Report 32/06, Canberra, ACT. Batelaan, O., De Smedt, F. and Triest, L. (2003) Regional groundwater discharge: phreatophyte mapping, groundwater modelling and impact analysis of land-use change. Journal of Hydrology 275, 86-108. Bedford, B. and Godwin, K. (2003) Fens of the United States: Distribution, characteristics, and scientific connection versus legal isolation. Wetlands 23, 608-629. Benyon, R. G. and Doody, T. M. (2004) Water use by tree plantations in south east South Australia. CSIRO Forestry and Forest Products, Technical Report no. 148, Mt. Gambier, SA. Benyon, R. G., Theiveyanathan, S. and Doody, T. M. (2006) Impacts of tree plantations on groundwater in south-eastern Australia. Australian Journal of Botany 54, 181-192. Bierwirth, P. N. and Welsh, W. D. (2000) Delineation of recharge beds in the Great Artesian Basin using airborne gamma- radiometrics and satellite remote sensing. Bureau of Rural Sciences Australia, Report for the National Landcare Program, Kingston, ACT. Bokuniewicz, H., Buddemeier, R., Maxwell, B. and Smith, C. (2003) The typological approach to submarine groundwater discharge (SGD). Biogeochemistry 66, 145-158. Boomer, K. and Bedford, B. (2008) Groundwater-induced redox-gradients control soil properties and phosphorus availability across four headwater wetlands, New York, USA. Biogeochemistry 90, 259-274. Boulton, A. J., Humphreys, W. F. and Eberhard, S. M. (2003) Imperilled subsurface waters in Australia: Biodiversity, threatening processes and conservation. Aquatic Ecosystem Health and Management 6, 41-54. Bradford, T., Adams, M., Humphreys, W. F., Austin, A. D. and Cooper, S. J. B. (2010) DNA barcoding of stygofauna uncovers cryptic amphipod diversity in a calcrete aquifer in Western Australia's arid zone. Molecular Ecology Resources 10, 41- 50. Brodie R.S., Green R. and Graham M. (2002) Mapping groundwater-dependent ecosystems: A case study in the fractured basalt aquifers of the Alstonville Plateau, New South Wales, Australia. Proceedings of the IAH 2002 Conference, Darwin, 2002. Brodie, R., Sundaram, B., Tottenham, R., Hostetler, S. and Ransley, T. (2007a) An adaptive management framework for connected groundwater-surface water resources in Australia. Bureau of Rural Sciences, Canberra, ACT. Brodie, R., Hostetler, S. and Slatter, E. (2007b) Q-Lag: A new hydrographic approach to understanding stream-aquifer connectivity. Australian Department of Agriculture, Fisheries, and Forestry, Canberra, ACT. Bruelheide, H., Vonlanthen, B., Jandt, U., Thomas, F. M., Foetzki, A., Gries, D., Wang, G., Zhang, X., and Runge, M. (2010) Life on the edge - to which degree does phreatic water sustain vegetation in the periphery of the Taklamakan Desert? Applied Vegetation Science 13, 56-71. Burnett, W. C., Bokuniewicz, H., Huettel, M., Moore, W. S. and Taniguchi, M. (2003) Groundwater and pore water inputs to the coastal zone. Biogeochemistry 66, 3-33. Canham, C. A., Froend, R. H., and Stock, W. D. (2009) Water stress vulnerability of four Banksia species in contrasting ecohydrological habitats on the Gnangara Mound, Western Australia. Plant, Cell & Environment 32, 64-72. Chapman, J., Lewis, B. and Litus, G. (2003) Chemical and isotopic evaluation of water sources to the fens of South Park, Colorado. Environmental Geology 43, 533-545. Clark, R. M. and Harvey, W. (2008) Dryland salinity in Victoria in 2007: An analysis of data from the soil salinity database and Victorian discharge monitoring network. Victoria Department of Primary Industries, Bendigo, VIC. Clifton, C. A., Cossens, B. and McAuly, C. (2007) A framework for assessing the environmental water requirements of groundwater dependent ecosystems report 1 assessment toolbox. Land and Water Australia, Braddon, ACT.

Mapping Terrestrial Groundwater Dependent Ecosystems 61

Colvin, C., Le Maitre, D. and Hughes, S. (2003) Assessing Terrestrial Groundwater Dependent Ecosystems in South Africa. Water Research Commission, WRC Report No. 1090-2/03. Colvin, C., Le Maitre D., Saayman I., Hughes S. (2007) An Introduction to Aquifer Dependent Eco systems in South Africa. Water Research Commission. WRC Report No. TT301/07. Conrad, J., Low, B., Münch, Z. and Pond, U. (2005) Remote sensing based botany and groundwater dependency study: Northern Sandveld. Department of Water Affairs and Forestry, DWAF Report number RDM/G300/02/CON/0505, Pretoria, South Africa. Cook, P. and O’Grady, A. (2006) Determining soil and ground water use of vegetation from heat pulse, water potential and stable isotope data. Oecologia 148, 97-107. Coram, J., Dyson, P., Houlder, P. and Evans, R. (2000) Australian groundwater flow systems contributing to dryland salinity. Report by the Bureau of Rural Sciences for the National Land and Water Resources Audit. Bureau of Rural Sciences, Canberra, ACT. Costelloe, J. F., Payne, E., Woodrow, I. E., Irvine, E. C., Western, A. W., and Leaney, F. W. (2008) Water sources accessed by arid zone riparian trees in highly saline environments, Australia. Oecologia 156, 43-52. Council of Australian Governments (1994) Water Reform Framework. 25 February 1994 communique. Council of Australian Governments, Hobart, TAZ. Dahl, M., Nilsson, B., Langhoff, J. H. and Refsgaard, J. C. (2007) Review of classification systems and new multi-scale typology of groundwater-surface water interaction. Journal of Hydrology 344, 1-16. Dahlhaus, P., Heislers, D. and Dyson, P. (2002) Corangamite CMA Groundwater Flow Systems. Report No. CCMA 02/02. Danielescu, S., MacQuarrie, K. T. B. and Faux, R. N. (2009) The integration of thermal infrared imaging, discharge measurements and numerical simulation to quantify the relative contributions of freshwater inflows to small estuaries in Atlantic Canada. Hydrological Processes 23, 2847-2859. Danielopol, D. L. (1989) Groundwater fauna associated with riverine aquifers. Journal of the North American Benthological Society 8, 18- 35. De Villiers, S. (2005) Rapid communication. The correct use of Sr isotopes in river-groundwater mixing models: A Breede River case study. Water SA 31, 397-398. Department of the Environment, Water, Heritage and the Arts (2006) Indicator: IW-08 Groundwater used for irrigation State of the Environment 2006. Australian Government, Canberra, ACT. http://www.environment.gov.au/soe/2006/publications/drs/indicator/189/index.html, updated 14-Dec-2007, accessed 19-April-2010. Drexler, J. and Ewel, K. (2001) Effect of the 1997–1998 ENSO-related drought on hydrology and salinity in a micronesian wetland complex. Estuaries and Coasts 24, 347-356. Dulaiova, H., Peterson, R., Burnett, W. C. and Lane-Smith, D. (2005) A multi-detector continuous monitor for assessment of 222Rn in the coastal ocean. Journal of Radioanalytical and Nuclear Chemistry 263, 361-363. Eagleson P. S. (2002) Ecohydrology: Darwinian Expression of Vegetation Form and Function. Cambridge University Press. Eamus, D., Froend, R., Loomes, R., Hose, G. and Murray, B. (2006) A functional methodology for determining the groundwater regime needed to maintain the health of groundwater-dependent vegetation. Australian Journal of Botany 54, 97-114. Eamus, D. (2009) Identifying groundwater dependent ecosystems: A guide for land and water managers. Land and Water Australia, PN30129. Eberhard, S. M., Halse, S. A., Williams, M. R., Scanlon, M. D., Cocking, J. and Barron, H. J. (2009) Exploring the relationship between sampling efficiency and short-range endemism for groundwater fauna in the Pilbara region, Western Australia. Freshwater Biology 54, 885-901. Ellins, K. K., Roman-Mas, A. and Lee, R. (1990) Using 222Rn to examine groundwater/surface discharge interaction in the Rio Grande de Manati, Puerto Rico. Journal of Hydrology 115, 319-341. Ewe, S., Sternberg, L. and Childers, D. (2007) Seasonal plant water uptake patterns in the saline southeast Everglades ecotone. Oecologia 152, 607-616. Farrington. P., Watson, G.D., Bartle, G.A. and Greenwood, E.A.N. (1990) Evaporation from dampland vegetation on a groundwater mound. J. Hydrol. 115:65–75.

Fawcett, J. and Huggins, C. (2005) Preliminary appraisal of a water balance model for Lake Wallace, Edenhope. Victoria Department of Primary Industries, Bendigo, VIC. Fitzpatrick, R. W., Hicks, W. S., Marvanek, S. P., Raven, M. D., Dahlhaus, P. G. and Cox, J. W. (2007) Scoping study of coastal and inland acid sulfate soils in the Corangamite CMA. CSIRO Land and Water Science Report 28/07 Glen Osmond, SA. Fluvial Systems (2006) Lake Condah Water Restoration project. Hydrogeological feasibility study. Technical report prepared for the Glenelg Hopkins Catchment Authority Furby S. (2002) Land Cover Change: Specification for remote sensing analysis. National Carbon Accounting System Technical Report No.9. Autralian Greenhouse Office. Froend, R. H., Heddle, E. M., Bell, D. T. and McComb, A. J. (1987) Effects of salinity and waterlogging on the vegetation of Lake Toolibin, Western Australia. Austral Ecology 12, 281-298. Froend R. H. and Loomes R. C. (2004) Approach to the determination of ecological water requirements of groundwater dependent ecosystems in Western Australia (No. CEM 2004-12). (Centre for Ecosystem Management, Edith Cowan University:Joondalup, WA) Gallardo, A. and Marui, A. (2006) Submarine groundwater discharge: an outlook of recent advances and current knowledge. Geo-Marine Letters 26, 102-113. Gibson, J. J., Price, J. S., Aravena, R., Fitzgerald, D. F. and Maloney, D. (2000) Runoff generation in a hypermaritime bog- forest upland. Hydrological Processes 14, 2711-2730. Gries, D., Zeng, F., Foetzki, A., Arndt, S. K., Brulheide, H., Thomas, F. M., Zhang, X., and Runge, M. (2003) Growth and water relations of Tamarix ramosissima and Populus euphratica on Taklamakan desert dunes in relation to depth to a permanent water table. Plant, Cell & Environment 26, 725-736. Hancock, P. J. and Boulton, A. J. (2009) Sampling groundwater fauna: efficiency of rapid assessment methods tested in bores in eastern Australia. Freshwater Biology 54, 902-917. Harvey, J. W., Newlin, J. T. and Krupa, S. L. (2006) Modeling decadal timescale interactions between surface water and ground water in the central Everglades, Florida, USA. Journal of Hydrology 320, 400-420. Harvey, K.R., Hill, G.J.E. (2001) Vegetation mapping of a freshwater swamp in the Northern Territory, Australia: a comparison of aerial photography, Landsat TM and SPOT satellite imagery. International Journal of Remote Sensing 22(15): 2911-2925 Hatton, T. and Evans, R. (1998) Dependence of ecosystems on groundwater and its significance to Australia. Land and Water Resources Research and Development Corporation (Australia), Occaisional Paper No. 12/98. Canberra, ACT. Hao, X., Li, W., Huang, X., Zhu, C. and Ma, J. (2010) Assessment of the groundwater threshold of desert riparian forest vegetation along the middle and lower reaches of the Tarim River, China. Hydrological Processes 24, 178-186. Huete, A., Didan, K., Miura, T., Rodriguez, E. P., Gao, X. and Ferreira, L. G. (2002) Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sensing of Environment 83, 195-213. Humphreys, W. F. (2006) Aquifers: The ultimate groundwater-dependant ecosystems. Australian Journal of Botany. 54, 115- 132. Jeffrey, S. J., Carter, J. O., Moodie, K. B. and Beswick, A. R. (2001) Using spatial interpolation to construct a comprehensive archive of Australian climate data. Environmental Modelling & Software 16, 309-330. Jolly, I. D., McEwan, K. L. and Holland, K. L. (2008) A review of groundwater-surface water interactions in arid/semi-arid wetlands and the consequences of salinity for wetland ecology. Ecohydrology 1, 43-58. Kemmers, R. H., van Delft, S. P. J. and Jansen, P. C. (2003) Iron and sulphate as possible key factors in the restoration ecology of rich fens in discharge areas. Wetlands Ecology and Management 11, 367-381. Kichner, J. O. G. (1995) Investigation into the contribution of groundwater to the salt load of the Breede River using natural isotopes and chemical tracers. South African Water Research Commission, WRC Report No 344/1/95. Kirchner, J., Moolman, J. H., du Plessis, H. M. and Reynders, A. G. (1997) Causes and Management of Salinity in the Breede River Valley, South Africa. Hydrogeology Journal 5, 98-108. Kishel, H. F. and Gerla, P. J. (2002) Characteristics of preferential flow and groundwater discharge to Shingobee Lake, Minnesota, USA. Hydrological Processes 16, 1921-1934. Lacey, G. C. (1996) Relating base flow characteristics to catchments properties. A scaling approach. Cooperative Research Centre for Catchment Hydrology, 96/8, Canberra, ACT.

Mapping Terrestrial Groundwater Dependent Ecosystems 63

Lambert, M. J. and Burnett, W. C. (2003) Submarine groundwater discharge estimates at a Florida coastal site based on continuous radon measurements. Biogeochemistry 66, 55-73. Lamontagne, S., Leaney, F. and Herczeg, A. (2002) Streamwater Groundwater Interaction: The River Murray at Hattah- Kulkyne Park, Victoria: Summary of Results. CSIRO Land and Water, Technical Report 27/02. Lamontagne, S., Cook, P. G., O'Grady, A., and Eamus, D. (2005a) Groundwater use by vegetation in a tropical savanna riparian zone (Daly River, Australia). Journal of Hydrology 310, 280-293. Lamontagne, S., Leaney, F. W. and Herczeg, A. L. (2005b) Groundwater-surface water interactions in a large semi-arid floodplain: implications for salinity management. Hydrological Processes 19, 3063-3080. Loaiciga, H. A. and Zektser, I. S. (2003) Estimation of Submarine Groundwater Discharge. Water Resources 30, 473-479. Macumber, P. G. (1980) The influence of groundwater discharge on the Mallee Landscape. In Aeolian Landscapes in the Semi- Arid Zone of South Eastern Australia, 67-85. Aust. Soc. Soil Sci. Riverine Branch. Macumber, P. G. (1991) Interaction Between Groundwater and Surface Systems in Northern Victoria. State of Victoria Department of Conservation and Environment, East Melbourne, VIC. Macumber, P. G. (1992) Hydrological processes in the Tyrrell Basin, southeastern Australia. Chemical Geology 96, 1-18. Macumber P.G. (2004) The hydrogeology of the Avoca Marshes region. Technical Report prepared for Dept. Primary Industries, Kerang, October 2004. Mensforth, L. J., Thorburn, P. J., Tyerman, S. D. and Walker, G. R. (1994) Sources of water used by riparian Eucalyptus camaldulensis overlying highly saline groundwater. Oecologia 100, 21-28. Mensforth, L. J. (1996) Water use strategy of Melaleuca halmaturorum in a saline swamp. Ph.D. thesis, University of Adelaide, Adelaide, SA. Mensforth L. J., Thorburn, P. J., Tyerman, S. D., and Walker, G. R. (1994) Sources of water used by riparian Eucalyptus camaldulensis overlying highly saline groundwater. Oecologia 100, 21–28. Metternicht G, Zinck JA (1997) Spatial discrimination of salt and sodium affected soil surfaces. Int J remote Sens 18 (12):2571- 2586. Moore, W. S. (2003) Sources and fluxes of submarine groundwater discharge delineated by radium isotopes. Biogeochemistry 66, 75-93. Morris, J. D. and Collopy, J. J. (1999) Water use and salt accumulation by Eucalyptus camaldulensis and Casuarina cunninghamiana on a site with shallow saline groundwater. Agricultural water management 39, 205-227. Münch, Z. and Conrad, J. (2007) Remote sensing and GIS based determination of groundwater dependent ecosystems in the Western Cape, South Africa. Hydrogeology Journal 15, 19-28. Murray, B. B. R., Zeppel, M. J. B., Hose, G. C. and Eamus, D. (2003) Groundwater-dependent ecosystems in Australia: It's more than just water for rivers. Ecological Management & Restoration 4, 110. Nathan, R.J. and Weinmann, P.E. (1993) Low Flow Atlas for Victorian Streams: drought management plan for Victoria's water resources. Department of Conservation and Natural Resources, Victoria, Long Term Planning Guidelines Background Report L3, Melbourne. Nicholson, C., Dahlhaus, P., Anderson, G., Kelliher, C. and Stephens, M. (2006) Corangamite Salinity Action Plan 2005-2008. Published report for Corangamite CMA. O'Brien, M. (2004) Corporate Geospatial Data Library Catalogue. State of Victoria Department of Primary Industries / Department of Sustainability and Environment, 17th Edition, Melbourne, VIC. O’Grady, A. P., Cook P. G., Howe P., and Werren G. (2006) Groundwater use by dominant tree species in tropical remnant vegetation communities. Australian Journal of Botany, 2006, 54, 155–171. Oliveira, J., Costa, P. and Braga, E. S. (2006) Seasonal variations of 222Rn and SGD fluxes to Ubatuba embayments, São Paulo. Journal of Radioanalytical and Nuclear Chemistry 269, 689-695. Paijmans, K. and Csiro Division of Water Land Resources (1985) Aspects of Australian Wetlands. CSIRO Division of Water and Land Resources, technical paper no. 44, Melbourne, VIC. Palinska, K., Scholz, J., Sterflinger, K., Gerdes, G. and Bone, Y. (1999) Microbial mats associated with bryozoans (Coorong Lagoon, South Australia). Facies 41, 1-14.

Patten, D., Rouse, L., and Stromberg, J. (2008) Isolated Spring Wetlands in the Great Basin and Mojave Deserts, USA: Potential Response of Vegetation to Groundwater Withdrawal. Environmental Management 41, 398-413. Podniesinski, G. and Leopold, D. (1998) Plant community development and peat stratigraphy in forested fens in response to ground-water flow systems. Wetlands 18, 409-430. Radke, L. and Howard, K. (2007) Influence of groundwater on the evaporative evolution of saline lakes in the Wimmera of south-eastern Australia. Hydrobiologia 591, 185-205. Salama, R. B., Farrington, P., Bartle, G. A. and Watson, G. D. (1993) The role of geological structures and relict channels in the development of dryland salinity in the wheatbelt of Western Australia. Aust. J. Earth Sciences 40, 45-56. Salama, R. B., Tapley, I., Ishii, T. and Hawkes, G. (1994) Identification of areas of recharge and discharge using Landsat-TM satellite imagery and aerial photography mapping techniques. Journal of Hydrology 162, 119-141. Schmidt, A., Stringer, C., Haferkorn, U. and Schubert, M. (2009) Quantification of groundwater discharge into lakes using radon-222 as naturally occurring tracer. Environmental Geology 56, 855-863. Scholz, O. (1990) Physiochemistry and vegetation of Piccaninnie Ponds, a coastal aquifer-fed pond in south-east South Australia. Australian Journal of Marine and Freshwater Research 41, 237-246. Semeniuk, V. (1983) Mangrove distribution in Northwestern Australia in relationship to regional and local freshwater seepage. Plant Ecology 53, 11-31. Shugg, A. and Brumley, J. C. (2003) Environmental, hydrogeological and conflict as elements determining the long term viable development of carbonated mineral waters of Hepburn, Victoria. National Environment Conference Brisbane, QLD. Environmental Engineering Society and Institute of Engineers, Australia. Shugg, A. (2009) Hepburn Spa: cold carbonated mineral waters of Central Victoria, South Eastern Australia. Environmental Geology 58, 1663-1673. Siegel, D. I. and P. H. Glaser (1987). Groundwater flow in a bog-fen complex, Lost River peatland, northern Minnesota. Journal of Ecology 75: 743-754. Siegel, D. I., Glaser, P. H., So, J. and Janecky, D. R. (2006) The dynamic balance between organic acids and circumneutral groundwater in a large boreal peat basin. Journal of Hydrology 320, 421-431. Sinclair Knight Merz (2001) Environmental Water Requirements of Groundwater Dependent Ecosystems. Environment Australia, Environmental Flows Initiative Technical Report Number 2, Canberra, ACT. Smith, S., Devitt, D., Sala, A., Cleverly, J. and Busch, D. (1998) Water relations of riparian plants from warm desert regions. Wetlands 18, 687-696. Smith, A. J. and Nield, S. P. (2003) Groundwater discharge from the superficial aquifer into Cockburn Sound Western Australia: estimation by inshore water balance. Biogeochemistry 66, 125-144. Smith, P. L., Williams, R. M., Hamilton, S. and Shaik, M. (2006) A risk-based approach to groundwater management for terrestrial groundwater dependant ecosystems. Murray Darling Basin Conference. Dept. Natural Resources NSW. Stewart, M. K., Morgenstern, U. and McDonnell, J. J. (2010) Truncation of stream residence time: how the use of stable isotopes has skewed our concept of streamwater age and origin. Hydrological Processes 9999. Talsma, T. and Gardner, B. (1986) Soil water extraction by a mixed Eucalypt forest during a drought period. Aust Jour Soil Res. 24 25-32. Taniguchi, M., Turner, J. V. and Smith, A. J. (2003) Evaluations of groundwater discharge rates from subsurface temperature in Cockburn Sound, Western Australia. Biogeochemistry 66, 111-124. Tcherepanov E. N., Zlotnik V. A., Henebry G. M. (2005) Using landsat thermal imagery and GIS for identification of groundwater discharge into shallow groundwater dominant lakes. Int J Remote Sens 26 (17):3649-3661. Thorburn P. J., Hatton T. J. and Walker G. R. (1993a) Combining measurements of transpiration and stable isotopes to determine groundwater discharge from forests. J. Hydrol. 150, 563–587. Thorburn P. J., Walker G. R. and Brunel J-P. (1993b) Extraction of water from Eucalyptus trees for analysis of deuterium and oxygen-18: laboratory and field techniques. Plant Cell and Environment 16, 269–277. Thorburn P. J. and Walker G. R. (1994). Variations in stream water uptake by Eucalyptus camaldulensis with differing access to stream water. Oecologia, 100: 293–301. Thornthwaite, C. W. (1948) An Approach toward a Rational Classification of Climate. Geographical Review 38, 55-94.

Mapping Terrestrial Groundwater Dependent Ecosystems 65

Turnbull, R. L. (2006) A Physical and Hydrogeochemical Investigation of Groundwater-Surface Water Interactions in Lake Corangamite and surrounding lakes, Western District of Victoria, Australia thesis, Monash University, Melbourne, VIC. Tweed, S. O., LeBlanc, M., Webb, J. A. and Lubczynski, M. W. (2007) Remote sensing and GIS for mapping groundwater recharge and discharge areas in salinity prone catchments, southeastern Australia. Hydrogeology Journal 15, 75-96. van Etten, E. and Vellekoop, S. (2009) Response of fringing vegetation to flooding and discharge of hypersaline water at Lake Austin, Western Australia. Hydrobiologia 626, 67-77. Valiela, I., Costa, J., Foreman, K., Teal, J., Howes, B. and Aubrey, D. (1999) Transport of groundwater-borne nutrients from watersheds and their effects on coastal waters. Biodegradation 10, 177-197. Van Tol, G., Van Dobben, H. F., Schmidt, P., and Klap, J. M. (1998). Biodiversity of Dutch forest ecosystems as affected by receding groundwater levels and atmospheric deposition. Biodiversity & Conservation 7, 221. Wagner, F. H. and Bretschko, G., 2003. Riparian Trees and Flow Paths between the Hyporheic Zone and Groundwater in the Oberer Seebach, Austria. International Review of Hydrobiology 88, 129-138. Walker, G., Gilfedder, M., Evans, R., Dyson, P. and Stauffacher, M. (2003) Groundwater Flow Systems Framework-Essential tools for planning salinity management. Murray-Darling Basin Commission, MDBC 14/03, Canberra, ACT. Walsh, P. (2008) A new method for analyzing the effects of joints and stratigraphy on spring locations: a case study from the Sacramento Mountains, south central New Mexico, USA. Hydrogeology Journal 16, 1459-1467. White, J. W. C., Cook, E. R., Lawrence, J. R. and Wallace S. B. (1985) The ratios of sap in trees: Implications for water sources and tree ring ratios. Geochimica et Cosmochimica Acta 49, 237-246. Winter, T. C. (1999) Relation of streams, lakes, and wetlands to groundwater flow systems. Hydrogeology Journal 7, 28-45. Winter, T. C., Harvey, J. W., Franke, O. L. and Alley, W. M. (1998) Ground Water and Surface Water A Single Resource. U.S. Geological Survey, USGS Circular 1139, Denver, CO. Wu, Y., Wen, X. and Zhang, Y. (2004) Analysis of the exchange of groundwater and river water by using Radon-222 in the middle Heihe Basin of northwestern China. Environmental Geology 45, 647-653. Zencich, S. J., Froend, R. H., Turner, J. T., Gailitis, V. (2002) Influence of groundwater depth on the seasonal sources of water accessed by Banksia tree species on a shallow, sandy coastal aquifer. Oecologia 131, 8–19.

Mapping Terrestrial Groundwater Dependent Ecosystems:

Method Development and Example Output

ISBN 978-1-74264-321-2 (print) ISBN 978-1-74264-322-9 (CD-ROM) ISBN 978-1-74264-323-6 (online)