THE EXPERIMENTAL CONTROL OF SPARTINA ANGLICA AND SPARTINA X TOWNSENDII IN ESTUARINE SALT MARSH

MARK E. R. HAMMOND B.Sc. (Hons), M.Res.

FACULTY OF SCIENCE ENVIRONMENTAL STUDIES UNIVERSITY OF ULSTER COLERAINE BT52 1SA NORTHERN IRELAND

THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 2001 Contents Acknowledgements i Declaration on access to contents ii Abstract iii

1. Introduction 1 1.1. Spartina anglica and Spartina x townsendii origin, spread and growth 1 1.2 . The advantages and disadvantages of S. anglica agg. invasion 4 1.2.1. Physical effects of S. anglica agg. invasion 4 1.2.2. Effects of S. anglica agg. invasion on flora 5 1.2.3. Effects of S. anglica agg. invasion on fauna (excluding birds) 5 1.2.4. Effects of S. anglica agg. invasion on birds 6 1.2.5. Economic impacts of S. anglica agg. invasion 7 1.3. Spartina spp. control / eradication treatments 8 1.3.1. Cutting, trampling and burning 8 1.3.1.1. Cutting / grazing 9 1.3.1.2. Trampling / crushing 9 1.3.1.3. Burning / steam treatment 9 1.3.1.4. Digging / bulldozing 10 1.3.2. Smothering, burying, herbicides 10 1.3.2.1. Smothering 10 1.3.2.2. Burying 11 1.3.2.3. Herbicides 11 1.3.2.3.1. Method of herbicide application 13 1.3.2.3.2. Environmental conditions at time of spray 13 1.3.3. Other possible methods for Spartina spp. control 13 1.3.3.1. Biological control 14 1.3.3.2. Coastal engineering 14 1.3.3.3. Other methods 15 1.3.4. Self control by S. anglica agg. / Die-back 15 1.4. Spartina anglica agg. In Northern Ireland 15 1.4.1. Lough Foyle 16 1.4.2. Strangford Lough 16 1.4.3. Northern Ireland concerns about S. anglica agg. spread 20 1.5. Objectives 21 2. Methods 22 2.1. Study area 22 2.2. Experimental lay-out 25 2.2.1. Plot design 26 2.2.3. Treatments 26 2.3.1. Herbicidal treatments 26 2.3.1.1. Herbicide formulation 28 2.3.1.2. Herbicide application 28 2.3.2. Mowing treatments 28 2.3.3. Plastic sheeting 29 2.4. Experimental records 29 2.4.1. Vegetation recording 30 2.4.1.1. Quadrat sampling for stem height, stem density and Domin values30 2.4.1.2. Stem height measurements 30 2.4.1.2.1. Sacrificial pseudoreplication 30 2.4.1.3. Stem density recording 31 2.4.1.4. Domin values 31 2.4.1.5. Dry weight of roots and rhizomes 31 2.4.2. Macro-invertebrate analysis 32 2.5. Statistical analysis 33 2.5.1. Analysis of Variance (ANOVA, F-test) 34 2.5.1.1. Tukey test 34 2.5.1.2. ANOVA assumptions 35 2.5.1.3. Transformations 35 2.5.2. Kruskal-Wallis test 36

3. Results 39 3.1. Live S. anglica agg. stem density 39 3.2. S. anglica agg. stem height 40 3.3. Root dry weight 42 3.4. Puccinellia maritima abundance 42 3.5. Other plant species 43 3.6. Invertebrate abundance 44 4. Discussion 66 4.1. Effects of control treatments on S. anglica agg. 66 4.2. Effects of control treatments on other salt marsh species 70 4.3. Effects of S. anglica agg. control on benthic fauna 75 4.4. Effects of S. anglica agg. control on sediments 82 4.5. Conclusions 84

5. Spartina anglica agg. management strategy 87 5.1. General strategy 87 5.2. Economic and practical restraints 89 5.3. Legal restraints 91 5.4. Alternative Spartina anglica agg. control in areas where legal 92 and economic restraints make eradication improbable 5.5. Other management considerations 93

6. References 87

7. Appendix – Raw data 127 Acknowledgments

This project was funded through a Department of Environment & Heritage Service, Northern Ireland and Department of Education, Northern Ireland, Co-operative Assisted Science and Technology (CAST) award.

I would like to thank Dr. Alan Cooper for his help and guidance during my research; Sam, Robert, Anja, Belinda, Gareth, Julia, George and Paul Hackney for assistance in the field and plant identification; Euan for computer assistance; Tommy for statistical advice; Nigel for photography; Shirley for financial issues; members of the ‘Spartina control group’ especially the late Davey Andrews for assistance given during this study; fellow Spartina researchers, Paul Hedge, Kim Patton, Andrew Bishop, Lisa Lantz, Janie Civille, Mark McCorry, Phil Davey, Alasdair Wells, Willie Shaw, and Paul Gillespie who have given me advice and information during this study; and Pierre for sharing his valuable knowledge and contacts within the field of invasive species.

Thanks also goes to all the friends who have made my three years study a very enjoyable and humorous time. Unfortunately they are too many to mention, but they know who they are.

Special thanks go to my family and to the hummingbird for shared wisdom. Declaration on access to contents

“I hereby declare that with effect from the date on which the thesis is deposited in the library of the University of Ulster, I permit the Librarian of the University to allow the thesis to be copied in whole or in part without reference to me on the understanding that such authority applies to the provision of single copies made for study purposes or for inclusion within the stock of another library. This restriction does not apply to the British Library Thesis Service (which is permitted to copy the thesis on demand for loan or sale under the terms of a separate agreement) nor to the copying or publication of the title and abstract of the thesis. IT IS A CONDITION OF USE OF THIS THESIS THAT ANYONE WHO CONSULTS IT MUST RECOGNIZE THAT THE COPYRIGHT RESTS WITH THE AUTHOR AND THAT NO QUOTATION FROM THE THESIS AND NO INFORMATION DERIVED FROM IT MAY BE PUBLISHED UNLESS THE SOURCE IS PROPERLY ACKNOWLEDGED”

______

Mark Hammond Abstract

Spartina anglica and Spartina x townsendii were introduced into estuarine habitats in Northern Ireland between the 1930s and 1950s. They were planted largely into inter-tidal sediment and have subsequently spread to develop mono-dominant swards at several locations. Sward development reduces the feeding habitat for wildfowl and waders by excluding salt marsh species such as Zostera spp. and physically preventing access to benthic invertebrates. Spartina anglica and Spartina x townsendii are therefore considered to be a pest species requiring control. This thesis investigates methods used for controlling Spartina anglica and Spartina x townsendii by field experiment. Experiments were initiated to study the effectiveness of control methods at two sites, one a S. anglica sward, the other a S. x townsendii sward. The treatments studied included, smothering with black plastic sheeting, applying the herbicides dalapon and glyphosate, and cutting followed by herbicide application. The effects of the control treatments on live S. anglica and S. x townsendii stem density, stem height, dry root weight, and associated flora and benthic fauna was examined. Dalapon applied at a rate of 57kg ha-1 and smothering were the most effective control methods, reducing live S. anglica and S. x townsendii stem density by over 95 %. Glyphosate was relatively ineffective. S. anglica and S. x townsendii re-established via re-growth or seedling colonization in all treatments over the two years following treatment applications, indicating that S. anglica and S. x townsendii eradication would require re-application of control treatments. Cutting treatments increased the abundance of Puccinellia maritima in the S. x townsendii sward one year after stem removal. This suggests that S. anglica and S. x townsendii control treatments may facilitate the growth of other salt marsh species. There were no differences in invertebrate numbers prior to and one year after treatment application. S. anglica and S. x townsendii eradication management strategies should initially focus on eradication of small foci (seedlings – clumps). Monitoring is required to prevent S. anglica or S. x townsendii re-colonization. It is suggested that swards are eradicated in the next phase of the management strategy. It is advisable to use herbicides for eradication when possible due to cost effectiveness and practicality. If herbicide use is prohibited non- herbicidal methods should be used to eradicate small foci. At present it is not practical or cost-effective to remove sward areas with non-herbicidal methods. In such cases it may be possible to covert Spartina spp. sward areas into mixed saltmarsh by using control techniques that facilitate the spread of other saltmarsh species within the sward. 1

1. Introduction

Invasion by non-native introduced species has been recognised as a major threat to conservation of the marine and coastal environment (Carlton 1989, 1996). Fourteen species of marine algae, five diatoms, one angiosperm and thirty invertebrates have been identified as non-native to British marine waters (Eno et al. 1997). These introductions have included species that adversely affect native species and habitats. Control methods have been applied to these detrimental species but, to-date, no non- native marine species has been successfully eradicated from British waters (Eno et al. 1997). Spartina anglica C. E. Hubbard, is an example of an introduced non-native species that has caused changes to native habitats and species. In many areas its invasion is seen as a threat to the conservation (e.g. through loss of bird feeding habitat), and amenity interests of the area. S. anglica has therefore been subject to direct control attempts.

1.1. Spartina anglica and Spartina townsendii origin, spread and growth

Spartina anglica and Spartina x townsendii H. & J. Groves originated at Hythe, Southampton water in the nineteenth century (Hubbard 1957; Marchant 1967; Goodman et al. 1969; Gray et al. 1991). Spartina anglica was the result of chromosome doubling by Spartina x townsendii, the sterile hybrid between the native European Spartina maritima (Curtis) Fernald and the introduced North American Spartina alterniflora Loisel (Marchant 1963, 1967, 1968; Raybould et al. 1991; Ferris et al. 1997). This process is a classic example of allopolyploid speciation. The parent species S. maritima and S. alterniflora have chromosomal counts of 2n = 60 and 2n = 62 respectively. The sterile hybrid S. x townsendii has a chromosome count of 2n = 62 whilst the amphidiploid S. anglica has a count of 2n = 122 – 124. S. x townsendii and S. anglica are morphologically difficult to tell apart, the distinction between them not being recognised until 1957 (Gray et al. 1991)(The term S. anglica agg. will be used in this article when referring specifically to both forms together).

Several S. anglica agg. populations appear distinct suggesting phenotypic variation within the species e.g. dwarf types (Charter and Jones 1951; Marchant 1968). Zonal variation in S. anglica morphology has also been noted in several marshes (Ranwell 2

1967; Long and Mason 1983, Haynes 1984; Marks & Mullins 1984; Marks & Truscott 1985; Hill 1990). Upper marsh plants tend to be taller with larger leaves and infloresences than lower marsh plants (Hill 1990). Other differences are likely to include shoot density, vegetative vigour, time of flowering, seed production, rates of tillering and number of over-wintering shoots (Marks & Mullins, 1984, Marks & Truscott 1985). These zonal differences result in plants from the same zone of different marshes being more similar morphologically than plants in different zones of the same marsh (Thompson 1990). Investigations of variation in isoenzyme levels suggest a lack of genetic variation in S. anglica (Gray et al. 1991). Distinct types and observed zonal differences are therefore likely to be caused by a range of environmental factors and age effects. The lack of genetic variation in S. anglica is likely to be a major factor influencing the evolutionary success of this relatively new species.

In Britain the niche of S. anglica is broadly between Mean High Water Neap tides (MHWN) and Mean High Water Spring tides (MHWS) (Gray et al. 1989, 1995). This comprises a range of low-high elevation estuarine habitat. S. x townsendii has been found on landward edges of S. anglica swards and generally has a preference for higher elevation sites and more stable habitats (Gray et al. 1991). Lower limits have been related to tidal inundations (Spartina anglica agg. have been reported to withstanding periods of 9 hours submergence (Ranwell et al. 1964). Wave action is also suggested as a factor limiting establishment (Morley 1973; van Eerdt 1985; Groenendijk 1986). Upper limits may be caused by lack of immersion (Huckle et al. 2000) or by competition with other species. Successful establishment is also more likely to occur in silt rather that sand sediments (Thompson et al. 1991; Huckle et al. 2000).

The low-lying mud flat S. anglica agg. niche is below the growth of most other halophytes. This enables the extensive spread of S. anglica agg. S. anglica spread occurs in two phases, initial invasion and establishment of seedlings or vegetational fragments on open mudflats, and then expansion of tussocks by radial clonal growth (over 30cm per year in organic mud at Dovey Estuary (Charter & Jones 1957). Spreading tussocks fuse to form clumps that can expand into extensive meadows. S. anglica agg. expansion may experience a lag phase (Gray & Raybould 1997), 3 however when expansions are occurring it can be very rapid. For example at Holes Bay, Poole Harbour, S. anglica introduced in 1899, expanded to cover over 200 ha - more than 60% of the intertidal mud flat - by 1924 (Hubbard 1965; Gray & Raybould 1997). Seed production of S. anglica is variable both temporally and spatially (Gray et al. 1991). It is thought that S. anglica does not form a seed-bank in estuarine substrates. S. x townsendii clonal spread is less vigorous than S. anglica and it is unable to spread by seed due to sterility (Gray et al. 1991).

S. anglica can also colonize upper marsh zones as individual tussocks in many salt marsh vegetation types (Adam 1981). Colonization of salt marsh creeks and pans can also occur.

As S. anglica agg. grows it can accrete large volumes of tidal sediment leading to substantial increases in marsh elevation. Ranwell (1972) reported that accretion rates normally range from 0.2 to 2.0 cm per year. Accumulation rates can be higher than this depending on localised conditions, for example, an area in the Netherlands recorded sediment accumulation of 1.8 m over 22 years (Ranwell 1967). Accumulation occurs due to S. anglica agg. stems reducing the force of tidal waters thus enabling sediment to settle (Gleason et al. 1979; Knutson et al. 1982). The extensive root and rhizome network then binds the deposited mud causing a rise and consolidation of the marsh surface. This is aided by the high tensile root strength of S. anglica agg. (van Eerdt 1985). These properties of sediment accumulation and stabilization made S. anglica agg. valuable species for coastal protection schemes in the twentieth century. The newly accreted sediment is often rich in macronutrients and thus also proved the value of S. anglica agg. for use in reclamation of intertidal mud flats.

Spartina anglica agg. has been introduced for coastal protection, or for land reclamation, into Denmark, Germany, North America, South America, South Africa, Australia, Tasmania, New Zealand, Ireland and China (Ranwell 1967; Nairn 1986; Chung 1990). Only the introductions into South America and South Africa were unsuccessful. Spartina anglica agg. have also been introduced to many areas in Great Britain resulting in approximately 12000 ha by 1967 (Hubbard & Stebbings 1967; Ranwell 1967) (reduced to 10000 ha by 1990 (Charman 1990). The current range of 4

S. anglica is from 48 oN to 57.5 oN in Europe, from 21 oN to 41 oN in China and from 35 oS to 46 oS in Australia and New Zealand (Gray & Raybould 1997). Climate change is likely to influence the future range of S. anglica. Climate changes should therefore be monitored in relation to the distribution of Spartina species.

In many areas the introduction and spread of Spartina species is now seen as a threat to the local environment, whilst in other areas it is still planted due to the benefits gained from its growth.

1.2. The advantages and disadvantages of S. anglica agg. invasion

“The immediate effect of the appearance of this pushful grass on the mudflats of the south coast has been to relieve their bareness and even to beautify them to some extent, and it has no doubt already affected animal life. Physical changes must follow, which, if the grass continues to flourish and spread, will react on the general conditions of the foreshore, resulting probably in the solidification and raising of the mud flat; but the process will take time. Whether the result will in the end be beneficial or to the contrary will depend greatly on local conditions. In any case it will be a change worth watching and studying” Stapf (1908).

1.2.1. Physical effects of S. anglica agg. invasion

S. anglica agg. stems dampen wave energy (Knutson et al. 1982), causing suspended sediment to accrete at stem bases (Gleason et al. 1979). The extensive rhizome system then binds the sediment, leading to a rise in elevation level. For this reason S. anglica agg. has been used world-wide as an agent for coastal protection / stabilisation and land reclamation (Oliver 1925; Allan 1930; Harbord 1949; Ranwell 1967; Chung 1982, 1993). China observed the benefits of S. anglica protection when hit by typhoons. S. anglica areas prevented erosion whereas unvegetated areas were severely scoured (Chung 1993; Qin et al. 1997). Benefits gained by land reclamation can be enhanced by the effect of S. anglica agg. growth on sediment structure and chemistry. S. anglica growth causes amelioration of saline soils thus producing suitable substrates for growth of agricultural crops such as wheat (Chung 1990, 1993). S. 5 anglica agg. growth also causes sediments to have higher organic matter, porosity and pH than barren mud-flat sediment.

The negative effects of such plantings are that salt marsh may be lost due to reclamation schemes (Doody 1990), and that geomorphology may be altered in other areas due to sediment accretions. This is most likely to be a problem in areas with a low sediment budget.

1.2.2. Effects of S. anglica agg. invasion on flora

S. anglica agg. invasion and spread leads to exclusion of native species. Reports refer to replacement or decline of Zostera beds due to the presence of S. anglica agg. (Oliver 1925; Chapman 1959; Ranwell & Downing 1959; Corkhill 1984; Madden et al. 1993; Gray & Raybould 1997). A wasting disease may have added to Zostera decline in the reported time periods (Ranwell & Downing 1960; Madden et al. 1993). Other species to be out-competed be S. anglica include Salicornia spp., Puccinellia maritima and Halimione portalacoides (Nairn 1986; Scholten et al. 1987; Scholten & Rozema 1990).

The growth of S. anglica agg. on the other hand will promote the establishment of other species. Several algal species are associated with S. anglica (Gleason & Zimmerman 1984; Nienhuis 1987) and fungi make use of decaying S. anglica material (Burge & Perkins 1977). The increase in elevation level and sediment stabilisation caused by S. anglica agg. growth may enable native salt marsh species to establish and may enable transitions / successions to other vegetation types (Allan 1930; Ranwell 1961, 1964; Packham & Liddle 1970; Gray & Pearson 1984; Hill 1987; Esselink et al. 1997; Esselink 2000).

1.2.3. Effect of S. anglica agg. invasion on fauna (excluding birds)

The high productivity of Spartina spp. (Lieth & Whittaker 1975; Gallagher et al. 1980; Long et al. 1990) results in a large amount of energy and organic matter entering the ecosystem. This forms the basis of many food webs (Pfeiffer & Wiegert 1981; Little 1990; Currin et al. 1995). 6

Many grazers can include Spartina spp. as a possible source of food. Mules, cattle, buffalo, cows, pigs, goats, sheep, rabbits, geese, rats, deer, donkeys, horses, grasshoppers, leafhoppers, damsel bugs, crabs and nematodes have all been noted grazing on Spartina spp. (Stapf 1914; Crichton 1960; Ranwell 1961; Payne 1972; Jackson et al. 1985, 1986; Chung 1990; Alkemade et al. 1993).

There is conflicting evidence surrounding the effect that S. anglica agg. invasion has on benthic mud-flat communities. Studies have reported a depleted fauna under S. anglica compared with nearby mud flat (Millard & Evans 1984; Haynes 1984). Others suggest that the presence of Spartina spp. may exclude certain species such as the clam Polymesoda carolinana (Capehart & Hackney 1989; Sayce 1990a). Alternatively some studies report that invertebrate communities of S. anglica marsh are richer and more abundant than mud flat areas , and are similar to native salt marsh assemblages (Hedge & Kriwoken 2000). A similar study in China found increased abundance of nerids, sand crabs and molluscs in areas colonised by S. anglica compared to bare mud flat (Lin 1984 cited in Chung 1983). Spartina spp. also provides habitat for many species of fish and invertebrates (Jackson et al. 1985; 1986; Campbell et al. 1990; Doody 1990). Certain invertebrate species such as nematodes seem to be associated with Spartina spp. (Osenga & Coull 1983; Alkemade et al. 1993). At Lindisfarne a study comparing bare mud flat and S. anglica sward reported that Carcinus maenus and Littorina saxalilis were only found within the S. anglica swards (Millard & Evans 1984).

1.2.4. Effects of S. anglica agg. invasion on birds

The loss of feeding habitat used by wintering birds has been one of the major concerns leading to calls for S. anglica agg. eradication. S. anglica agg. invasion into areas of Zostera and Enteromorpha reduce feeding areas for wildfowl such as Brent geese and widgeon (Ranwell & Downing 1959; Nairn 1986; Doody 1990). Percival et al. (1998) state that ‘Loss from the top of the shore through encroachment by Spartina anglica has the greatest effect on the sites capacity to sustain geese and widgeon.’ Waders are also likely to be affected by S. anglica agg. invasion. Dense stands physically stop birds going into the sward thus preventing access to invertebrates in 7 the mud (Prater 1981). Declines of oystercatcher, ringed plover, sanderling and dunlin have been noted in areas of S. anglica expansion (Davis & Moss 1984; Millard & Evans 1984; Nairn 1986, Goss-Custard & Moser 1988, 1990; Simpson 1995). It has not yet been proven that the decreases are due directly to S. anglica expansion or if it is linked to other factors. At Lindisfarne birds including dunlin have been noted to return to areas cleared of S. anglica (Corkhill 1984; Evans 1986). In other areas where S. anglica has died-back birds have failed to return (Goss-Custard & Moser 1988, 1990).

S. anglica agg. provides the benefits of nesting grounds and feeding areas for other bird species. S. anglica can provide a food source for geese in the form of rhizomes and winter buds (Esselink et al. 1997; Chung 1990). In China they feed S. anglica to their domestic geese, and it has been noted that wild geese prefer to feed on introduced S. anglica rather than wheat fields (Chung 1993). Redshanks have been noted feeding among S. anglica clumps even when other areas were available (Millard & Evans 1984)

Spartina spp. can also provide nesting grounds for species such as endangered rails (Harbell & Callaway 1990; Foin & Brenchley-Jackson 1991; Zedler 1993). Rails have been observed in S. x townsendii swards at Ballykelly, N. Ireland (pers. obs.). Ducks have also been observed using the Ballykelly sward as nesting grounds (pers. obs.).

1.2.5. Economic impacts of S. anglica agg. invasion

The threat of Spartina spp. spread into areas where oysters grow has concerned commercial oyster fisheries (Sayce 1990b; Harrington & Harrington 1992; Thom et al. 1997; Hedge & Kriwoken 1997; Kriwoken & Hedge 2000). Spread of S. anglica into amenity areas such as beaches and boat jetties raise concerns such as potential loss of tourism (Ranwell 1967; Truscott 1984; Corkhill 1984; Pringle 1993).

S. anglica agg. has the potential to be used for economic benefits. In China for example, 3.8 million kg of fresh S. anglica was cut in 1983 for use as fish food (Chung 1993). S. anglica agg. could potentially be used for biofuel, green manure / silage, sewage and pollution treatment, paper making, fish food, mushroom culture or 8 animal fodder (Oliver 1925; Hubbard & Ranwell 1966; Ranwell 1967; Chung 1982, 1990, 1993; Callaghan et al. 1984; Scott et al. 1990). The potential of S. anglica agg. in many of these cases surpasses that of other alternatives, S. anglica crude protein and lipid content is better than that of rice straw, corn and barley stalks (Chung 1993). Crops grown using S. anglica as green manure achieve higher yields than using pig manure or chemicals (Chung 1993).

Another possible economic benefit is the use of S. anglica agg. as a health product. Biomineral liquid from S. anglica culms has been put in sodawater, beer, milk, tea, wine and even bathing lotion. It is reported to improve the immune system, has anti- inflammation properties and is a cardiotonic (Qin et al. 1997, 1998). S. anglica flavonoids may also be useful as they reportedly help resistance to encephalon thrombus (Qin et al. 1997, 1998).

1.3. Spartina spp. control / eradication treatments

Attempts are often made to control or eradicate Spartina spp. in areas where the perceived disadvantages of Spartina spp. growth out-weigh the possible benefits.

1.3.1. Cutting, trampling, and burning

Cutting, trampling, and burning are control methods that remove or cause damage to the stems of Spartina spp. Rhizome energy reserves are used during the production of new shoots. Repeated removal or damage weakens the plant and may kill it. Questions surrounding the cost-effectiveness of these methods and practicalities remain. It is not known how large-scale these treatment operations could be, or how many times the control methods would have to be repeated to achieve successful kill.

The methods may be more usefully employed as a precursor to other control methods e.g. cutting and burying (Wijte & Gallagher 1991), Crushing and herbicide (Aberle 1990), Cutting and smothering (Bishop 1995; Lane 1995; Thom et al. 1997) or in combinations with each other such as cutting and trampling (Turner 1987). 9

1.3.1.1. Cutting / grazing

Cutting and grazing have been reported to increase S. anglica agg. stem density (Ranwell 1961; Hubbard 1970). Cutting has previously been tested as a method to remediate Spartina spp. damaged by oil spills (Baker et al. 1990), and has been used to collect S. anglica agg. in silage and biofuel projects (Hubbard & Ranwell 1966; Scott et al. 1990).

Many sites have attempted cutting as a control with little success (Aberle 1990; Bishop 1995; Lane 1995; Major and Grue 1997). Other reports have indicated that continuous cutting reduces Spartina spp. biomass (Turner 1987; Scott et al. 1990). There are no reports of successful eradication using cutting alone.

High grazing levels are likely to have similar effects to cutting (Ranwell 1961). Grazing raises problems relating to trampling, soil compaction, soil erosion and their effect on benthic organisms (Reimold et al. 1975; Shaw & Gosling 1995).

1.3.1.2. Trampling / crushing

Reports about the effectiveness of these methods have been mixed. Some suggest that these methods can successfully kill Spartina spp. (Aberle 1990). It is still unclear how often trampling / crushing is required and what is the required severity of treatment application. Bi-weekly stimulated horse trampling over a period of one year reduced above-ground S. alterniflora biomass by 55% (Turner 1987). Trampling of mixed salt marsh vegetation by students failed to have any impact on S. anglica cover (Headley & Sale 1999).

1.3.1.3. Burning / Steam treatment

Control trials using burning have been ineffective (Aberle 1990; Hedge 1997). Burning has little impact on rhizome biomass and may result in higher stem density (Turner 1987). Burning has been suggested as a suitable remediation technique for S. alterniflora that has been damaged by oil spills (Smith & Proffitt 1999). 10

Steam treatments have been suggested as a control method (Shaw & Gosling 1995; Hedge 1997). Success was reportedly similar to chemical control (Shaw and Gosling 1995).

1.3.1.4. Digging / bulldozing

Digging has proved to be a successful eradication method for removing young seedlings (Corkhill 1984; Way 1987; Aberle 1990; Hedge 1997). Way (1987) suggests that ‘Digging of seedlings and young plants (2-3 years old – 3-10 shoots) can be effective’. The method is labour intensive and requires the removal of all rhizome material to ensure success. In N. Ireland the largest plant to be dug out successfully was 50cm in diameter (Furphy 1970). Attempts to dig up larger clumps have been unsuccessful. Digging is a useful method to eradicate new infestations.

Areas cleared of S. anglica by bulldozer remain suitable habitat for S. anglica re- establishment (Truscot 1984, Way 1987). Bulldozing may also cause unacceptable damage to the shoreline.

1.3.2. Smothering, burying, and herbicides

Reports have indicated that these methods can achieve over 90% Spartina spp. eradication.

1.3.2.1. Smothering

Covering plants with black plastic sheeting blocks sunlight, prevents photosynthesis, and thus leads to plant death. Studies using black plastic to smother Spartina spp. have reported kill rates of up to 99-100% (Aberle 1990, Bishop 1995, Lane 1995).

There are problems associated with smothering treatments. It is likely that smothering will kill all other plant species. The effects of smothering on benthic invertebrates remain unknown. Smothering is a labour intensive method. There is no information on the size of area that could be practically treated using this method. Plastic sheeting 11 in estuarine environments will be prone to damage or removal by wind and waves (e.g. Frenkel 1990).

1.3.2.2. Burying

The removal of live and dead S. alterniflora shoots in winter, followed by blocking the cut shoot ends with sediment, caused death (Wijte & Gallagher 1991). This indicates that blocking the oxygen supply to roots and rhizomes will result in the death of Spartina species. This is the effect that burying is likely to have.

Initial attempts at burying using a rotoburying machine at Lindisfarne resulted in over 95% S. anglica kill (Davey et al. 1996; Anderson & Denny 1998). Digging down to the roots revealed that they were dead and inactive (Davey et al. 1996). Another study at Lindisfarne used a lightweight tracked vehicle to repeatedly drive over S. anglica until it was dislodged and buried in the sediment (Frid et al. 1999). Three years after the treatment stem density was approximately 40 per m2 in buried plots and 80 per m2 in experimental control plots. They suggested that sustained decrease in stem density caused by S. anglica burial was similar to a single application of herbicide.

This method has the disadvantage of requiring vehicles that may not be suitable for use on very soft fine sediments. It may also release rhizome fragments that would colonize elsewhere, or it may cause the release of sediments.

1.3.2.3. Herbicides

Herbicides are the most commonly used Spartina spp. eradication method used due to cost-effectiveness and efficiency. Early trials suggested that dalapon was one of the most effective herbicides for eradicating S. anglica agg., achieving over 90 % kill (Ranwell & Downing 1960; Taylor & Burrows 1968a). Other scientists and field managers (Corkhill 1984; Truscott 1984, Way 1987) confirmed consistently high kill rates. Dalapon however is no longer manufactured. It was considered impractical to use as it had to be applied at high doses and required a large volume of water. 12

Other herbicides have been tried in Spartina spp. eradication trials. To date these include amitrole, bromacil, clethodium, diuron, fenuron, fenoxaprop-ethyl, fluazifop- P, glyphosate, haloxyfop, imazapur, karbulilate, monuron, MSMA, paraquat, quizalofop, and selhoxydim (Ranwell & Downing 1960; Taylor & Burrows 1968a,b; Edward & Davis 1974; Corkhill 1984; Truscott 1984; Aberle 1990; Crockett 1990, 1997; Crothers 1990; Way 1990; Hubbard & Whitwell 1991; Garnett et al. 1992; Kilbride et al. 1995; Bishop 1995; Boekel 1995; Shaw & Gosling 1995; Palmer et al. 1995; Paveglio et al. 1996; Pritchard 1995; Simenstad et al. 1996; Felsot 1997; Major & Grue 1997; Norman & Patten 1997; Shaw & Gosling 1997). Several of these herbicides are reported to be as effective as dalapon. Fluazifop, haloxyfop and imazapyr have each achieved over 90% Spartina spp. kill (Pritchard 1995; Shaw and Gosling 1995).

The most common herbicide tried for Spartina spp. control has been glyphosate. In several areas glyphosate is one of the few herbicides that is licensed for use in estuarine environments e.g. Washington (Aberle 1990) and N. Ireland. Reports on the success of glyphosate have been mixed (Way 1987; Aberle 1990; Garnett et al. 1992; Kilbride et al. 1995; Pritchard 1995; Shaw & Gosling 1995; Crockett 1997). The greatest successes achieving over 75% kill have been obtained using glyphosate along with an added surfactants / adjuvants (Garnett et al. 1992; Kilbride et al. 1995; Crockett 1997). Surfactants / adjuvants are often added to herbicides to help them stick to the leaf surface and thus increase herbicide uptake prior to being washed off. This is especially useful in areas where plants are prone to water inundations such as intertidal areas. In N. Ireland there is no licence for the use of surfactants in estuaries. In such cases herbicides such as glyphosate would have to be applied on its own. Previous work suggests that applications of glyphosate on its own produces poor S. anglica kill rates (Garnett et al. 1992).

The success of any herbicide control programme will be variable due to application rates of herbicides, surfactants used, method of application, and environmental conditions such as tidal regime and weather at time of spraying. 13

1.3.2.3.1. Method of herbicide application

Backpack spraying is commonly used for spot spraying small or easily controlled infestations. Vehicle mounted sprayers may be more practical for use in large infestations or in developed swards. An alternative way to apply herbicides to large areas (10 ha and above) is aerial application by plane or helicopter. Aerial spraying cannot use large volumes of water, therefore high herbicide concentrations are used. Aerial applications will be more prone to herbicide drift than ground applications.

Other possible methods for herbicide application include injecting herbicides into plants or sediments (not practical on large scale), or by wick application. During wick application herbicide is absorbed onto fabric and then wiped onto the plant. This method can use high concentrations of herbicide and can be used in sensitive areas without damage to other species.

1.3.2.3.2. Environmental conditions at time of spray

Ideal spraying maximises the time that the herbicide remains on Spartina spp. before being washed of by tides or rain. This is best achieved by spraying during outgoing neap tides. Spraying at mid-day also reduces the likely-hood of herbicide being washed of by evening dew. Spraying in the morning could result in spraying herbicide onto dew covered Spartina species. Spraying should be avoided during this period as it is preferable to spray onto dry Spartina species. Spraying should also take place on overcast days (Spartina spp. shows leaf rolling on very hot days, and herbicides may crystallize in hot weather) with low wind speeds (exposes underside of leaves yet avoids spray drift) and no rain.

1.3.3. Other possible methods for Spartina spp. control

These methods have not been tried but have been suggested as possible Spartina spp. control methods. 14

1.3.3.1. Biological Control

Spartina spp. are prone to damage by herbivores (Daehler & Strong 1995; Wu et al. 1999). High densities of Prokelisa spp. (planthoppers) have been reported to kill over 90% of S. anglica plants in a greenhouse experiment. S. anglica has not yet been exposed to high Prokelisa levels in the field. Field studies with S. alterniflora, which has evolved with Prokelisa have indicated tolerance of high Prokelisa density (Daehler & Strong 1995). Any other herbivore that feeds on Spartina spp. may be a possible biological control treatment (Payne 1972, Pfeiffer & Wiegert 1981; Aberle 1990; Strong 1990).

Other possible sources of a biological control agent would be microbial pathogens or fungi. An early attempt of biological control involved the transplanting of S. anglica plants from areas of die-back into healthy S. anglica sites (Corkhill 1984). It was thought that die-back might have a pathogen cause. However there was no die-back in the healthy sward resulting from the transplants. Possible agents that would cause damage to Spartina spp. include the fungi Claviceps purpurea, which causes a decline in seed numbers (Thompson 1991) or the Spartina leaf mottle virus (possibly vectored by a mite (Jones 1980).

The specificity of target species of any possible biological agent would have to be tested before any release is sanctioned. There are also unanswered questions regarding resistant genetic variability within S. anglica agg. populations.

It is unlikely that biological control will provide 100% eradication of S. anglica agg. due to population dynamics. It should therefore be considered as a control treatment rather than an eradication method.

1.3.3.2. Coastal engineering

Changes in hydrology may influence Spartina spp. growth. Coastal engineering could artificially increase scour in an area causing erosion, could increase fresh water flow into an area, or could dike and flood Spartina spp. (Way 1987; Aberle 1990). It may be possible that these methods could be used to reduce S. anglica agg. abundance. 15

1.3.3.3. Other methods

Injecting lactic acid into sediments, electroshocking, cutting and coating stems to cut of oxygen supply and other herbicides e.g. pre-emergent and rhizome inhibitors (Way 1987; Aberle 1990).

1.3.4. Self control by S. anglica agg. / Die-back

Control may not be necessary if areas show a decline in S. anglica agg. due to ‘die- back’.

Die-back has caused a decline in S. anglica agg. abundance since the mid 1920’s in most south, and south-eastern coastal estuaries in England, as well as areas in south- west Britain, northern France and south-west Netherlands (Gray et al. 1991, Gray & Raybould 1997). In Holes Bay, Poole Harbour, for example, the 208 ha of S. anglica recorded in 1924 was reduced to about 63 ha by 1984 (Gray & Raybould 1997).

Die-back of S. anglica is due to death caused by soft-rotting of the rhizomes and a gradual decline in vigour of old S. anglica populations (Thompson 1991). The definitive cause of die-back is unknown. It tends however to occur in waterlogged, fine sediments, which induce anaerobiosis and toxic sulphide levels (Goodman et al. 1959; Goodman 1960; Goodman & Williams 1961; Webb et al. 1995).

To-date, there have been no reported cases of S. anglica agg. die-back in Northern Ireland.

1.4. Spartina anglica agg. in Northern Ireland (Fig 1.1)

The first planting of S. anglica occurred at Belfast Lough in 1929 (Praeger 1932), although little if any remains in this area (Bleakley 1979). Further Spartina spp. plantings took place during the 1930’s, 40’s and 50’s in Carlingford Lough, Lough Foyle and Strangford Lough (Bleakley 1979, Carter 1982, Hackney, 1992). Carlingford Lough presently contains seedlings, tussocks and clumps, but swards have yet to develop. S. anglica seedlings, tussocks and clumps have also been 16 recorded in Dundrum Bay but there are no records of introductions being made into this area. Lough Foyle and Strangford Lough contain areas of seedlings, tussocks, clumps and swards.

1.4.1. Lough Foyle (Fig 1.2)

S. anglica agg. was planted at Lough Foyle in 1933 to protect coastal railway lines. 3 Establishment was successful over a distance of /4 miles between Ballykelly and Carrickhue. Initial spread formed a sward that was confined to a strip between these two areas. In 1978 collected S. anglica agg. material from the sward was examined and identified as the sterile hybrid S. x townsendii (Hackney 1980).

A survey in 1997 identified a number of areas that have been colonized by S. x townsendii since the previous survey in 1976 (West to Longfield and as far north-east as the River Roe). S. x townsendii has also spread outside the boundaries of the enclosed sward at Ballykelly. Recent collected material has contained samples that resemble Spartina anglica (Hackney; McCorry pers. comm.). The distinguishing features that differentiate S. anglica and S. x townsendii e.g. ligule length, can overlap between the species. The material collected has features that are on the border between the two species. The material has therefore been sent for chromosomal counting.

1.4.2. Strangford Lough (Fig 1.3)

Initial introductions of Spartina anglica agg. occurred between 1930 and 1940, but it is not known whether these attempts were successful (Carter 1982). In the mid 40’s Spartina anglica was planted at Wood Island, Ardmillan Bay on the western side of the Lough to stabilize a causeway. Plant fragments from this planting were washed into Trench Road Bay. S. anglica spread in this area to cover 13.7 ha by 1975, and despite control measures covered 23.3 ha by 1988. The area is no longer subject to control and is currently expanding. Localised spread in the area can be observed on aerial photographs taken in 1951, 1962, 1969, 1975 (See Kirby 1989). 17

N

Lough Foyle

Belfast Lough

Strangford Lough

Dundrum Bay

040 Carlingford Km Lough

Fig. 1.1 Northern Ireland 18

N

Lough Foyle River Roe

Ballykelly

Carrickhue Longfield River 05 Foyle Km

Fig 1.2 Lough Foyle 19

N

Comber Mount Stewart River Greyabbey Castle Espie Pier

Ballydrain Horse Island

Trench Road Bay

Wood Island

Strangford Lough

Ardmillan Bay Gransha

The Dorn

03Km

Fig 1.3 Strangford Lough 20

S. anglica spread outside the Ardmillan bay area. By 1969 a S. anglica sward had developed at Ballydrain, and tussocks were present on mud and salt marsh north of the Comber River (Furphy 1970). Few S. anglica plants have been recorded south of Ardmillan Bay. At present the main areas affected on the western side of Strangford Lough are Castle Espie Pier, Comber River, Ballydrain Bay and Ardmillan Bay.

S. anglica was planted at Mount Stewart on the eastern side of the Lough in 1952 to protect the upper foreshore (Bleakley 1979). Spread has not been as extensive as on the western side of the Lough, however S. anglica has been recorded from the most northerly tip of the Lough too as far south as The Dorn, and a sward areas have developed at Gransha and Greyabbey.

Since 1992 a fivefold increase in S. anglica has occurred from Comber Estuary to Horse Island suggesting that S. anglica is continuing to expand in the Lough (Andrews 1997).

1.4.3. Northern Ireland concerns about S. anglica agg. spread

The main concern caused by the spread of S. anglica agg. in Northern Ireland is the loss of wildfowl and wader feeding areas. Strangford Lough for example is of international importance for wildfowl and waders. Two-thirds of the Greenland / Canadian Arctic breeding population of Pale-bellied Brent geese (Branta bernicla hrota O. F. Muller), some 13500 birds utilise the Lough (O’Briain & Healy, 1991). Zostera spp. is the preferred food of geese in the Lough (Mathers & Montgomery 1997). S. anglica agg. grows in the many of the same mud flat areas as Zostera. Its spread is therefore seen as a threat to Zostera beds and therefore to wildfowl such as Brent geese and wigeon (Anas penelope (L.)). Similar concerns are raised with the spread of S. anglica agg. and the loss of invertebrate-rich mud flat used for feeding by waders.

Other issues of concern that have been raised regarding the spread of S. anglica agg. in Northern Ireland estuaries include; the loss of high quality salt marsh (a rare habitat in Northern Ireland); invasion and blocking of channels leading to flooding of farmland behind the marshes; destroying amenity of areas. 21

Since the late 1960’s attempts have therefore been made to control and eradicate S. anglica agg.. Methods used have included digging, bulldozing and spraying with dalapon (2,2 dichloropropionic acid). In 1982 a spraying programme at Ardmillan Bay resulted in legal action by a local oyster fisheries. It was claimed that the eradication attempt released sediments that resulted in the smothering of oyster stock (Kirby 1994). Since this time a spraying ban has been imposed in any shellfish designated areas within Northern Ireland. Dalapon is still used in areas where spraying is permitted. Dalapon is however no longer manufactured and the stockpile for use in Northern Ireland will run out in 5 – 10 years. The Environment and Heritage Service, Northern Ireland is keen to find a replacement herbicide for S. anglica agg. control. Glyphosate is, to-date, the only other herbicide for which a licence has been granted for use in estuarine environments in Northern Ireland. The most successful S. anglica agg. control levels have been achieved when Glyphosate is used with a surfactant. Surfactants are currently banned from use in Northern Irish inter-tidal areas. The Environmental and Heritage Service is also wanted to investigate the use of non-herbicidal methods for S. anglica agg. control.

1.5. Objectives The objectives of this study were to – • Asses the effectiveness of the herbicides dalapon and glyphosate in controlling S. anglica agg. by field experiment. • Compare this with alternative methods of control. • Examine the effects of experimental treatments on the associated estuarine flora and fauna. • Evaluate habitat recovery following control. 22

2. Methods

2.1. Study Area This study was carried out in a Spartina x townsendii sward at Ballykelly, Lough Foyle (Fig 1.2) and in a Spartina anglica sward at Ballydrain, Strangford Lough (Fig. 1.3).

Lough Foyle is a 200 km2 marine inlet on the northern end of Northern Ireland. It has a tidal range between 1 - 2 m. The lough receives no Atlantic swell waves. Waves in the lough are locally generated and of low energy due to limited fetch. The lough receives prevailing winds from the south-west and west. Extensive mud flats occur in the southern end and western side of the Lough. The lough has carboniferous bedrock and contains marine and glacial sediments.

The study plots were located in a 1.4 ha Spartina x townsendii sward at Ballykelly (Grid ref. 55, 03 oN – 7, 02.6 oE) (Fig 2.1). The sward is confined to a sheltered bay created by the building of a cross-bank in 1840. S. x townsendii was introduced to the area in the 1930s. The bay is gently sloping (> 1 o slope), and contains silty sediment. The bay is dissected by one large channel (up to 30 m wide by 1.5 m deep) and by several smaller creeks (up to 2 m wide by 1.5 m deep). The plots receive no tidal inundation at Mean High Water Neap tides. During Mean High Water Spring Tides inundation levels range between 32 cm on the lower elevation plots and 20 cm on the higher elevation plots.

S. x townsendii within the study plots had a mean stem density of 232 per m2, a mean stem height of 33.2 cm, and a mean dry root weight of 12.2 per 1177.5 cm3, in July 1998. During the study S. x townsendii was noted to begin growth in April and began flowering in June / July.

The study area contained Puccinellia maritima (median Domin value 3.5), Aster tripolium and Plantago maritima (individuals of both, median Domin value less than 1) prior to the study. Several salt marsh strips were in close proximity to the sward (Fig. 2.1). The Domin values of the species present in these salt marsh areas are provided in Table 2.1. 23

N

Cross Creek Strengthened Bank River Table 2.1 Ballykelly salt marsh Channel vegetation Domin values. Lough Foyle Puccinellia maritima 6 Agropyron pungens 5 Agricultural Land Aster tripolium 5 Plantago maritima 5 Salt Marsh Festuca rubra 5 Vegetation Triglochin maritima 4 Study Chenopodium rubrum 1 Plots Spartina Tussocks Cochleria officinalis 1 + Clumps Glaux maritima 1 Juncus gerardii 1 Spartina townsendii 1 Salt Marsh Vegetation Spergularia media 1 Armeria maritima + Potentilla anserina + Spartina Sward Scirpus maritimus + Ballykelly

0 200 M Carrickhue Railway Agricultural Land

Fig. 2.1 Ballykelly Study Area 24

Table 2.2 Ballydrain salt marsh N vegetation Domin values. Agricultural Land Spartina clumps / tussocks Puccinellia maritima 6 Rocks Plantago maritima 4 Aster tripolium 4 Spartina + Spartina anglica 3 Salicornia Salicornia spp. 2 Area Juncus gerardii 2 Spergularia marina 2 Strangford Spartina Lough Limonium humile 1 Sward Suaeda maritima 1 Study Scirpus maritimus 1 Plots Armeria maritima 1 Creeks Cochleria anglica 1 Salt Marsh Glaux maritima 1 Vegetation Chenopodium rubrum 1 Spartina Sward Juncus maritimus 1 Dam Phragmites australis 1 Triglochin maritima 1

050 Salt Marsh Spartina + Salicornia spp. zone M Vegetation Saprtina anglica 8 Salicornia spp. 8

Fig. 2.2 Ballydrain Study Area 25

Strangford Lough is a 150 km2 marine inlet on the eastern side of Northern Ireland. It has a tidal range between 3.5 m (springs) and 2.0 m (neaps). The lough receives no Irish Sea swell waves. Waves in the lough are locally generated and of low energy due to limited fetch. The lough receives westerly prevailing winds. Extensive mud flats and salt marsh areas occur in the northern end of the lough. The lough has Silurian shale and slate bedrock and contains glacial sediments.

The study plots were located in a 0.15 ha Spartina anglica sward in a sheltered bay at Ballydrain (Grid ref. 54, 31.8 oN – 5, 40.6 oE) (Fig 2.2). S. anglica was introduced to the region in 1945 and by 1969 a sward had developed at Ballydrain. The bay is gently sloping (1 o slope), and contains silty sediment. Channels and creeks up to 3 m wide and 90 cm deep dissect the bay. The plots receive no tidal inundation at Mean High Water Neap tides. During Mean High Water Spring Tides inundation levels range between 67 cm on the lower elevation plots to 51 cm on the higher elevation plots.

S. anglica within the study plots had a mean stem density of 336 per m2, a mean stem height of 23.7 cm, and a mean dry root weight of 11.1 g per 1177.5 cm3, in July 1998. During the study S. anglica was noted to begin growth in April and began flowering in June / July.

The study area contained a low abundance of Puccinellia maritima (Domin value 1) prior to the study. No other species were recorded in the study plots. Several species- rich salt marsh strips were in close proximity to the sward (Fig. 2.2). The Domin values of the species present in these salt marsh areas are provided in Table 2.2.

2.2. Experimental lay-out (Fig 2.3).

Relatively uniform areas of S. anglica agg. sward were selected for trials to avoid gulleys and topographic discontinuations. Plots of 5 x 5 m were laid out in a random block formation. Ranwell & Downing (1960) and Taylor & Burrows (1968a) used similarly sized plots to study S. anglica agg. control. The plots were angled approximately parallel to the shoreline, with a separating distance between plots of 5 m. Previous studies used gaps as little as 1 m between plots (Taylor & Burrows 26

1968a). A 5 m gap reduced the possibility of herbicides being translocated from plot to plot via the underground root system, and also aided in practical working when applying the treatments. Marker poles were inserted as deep as possible into the mud to mark plot corners. A 50 cm buffer zone on the inside of the plots ensured that scouring at the base of the marker poles had no influence on the recorded S. anglica agg.. Six replicate plots of each experimental treatment were applied.

2.2.1. Plot design (Fig 2.3)

Walking was allowed around the outer edge of each plot. Within each plot, two 1 m walking strips were created to be used when applying treatments and when recording. A buffer zone of 50 cm was created around the inner edge of the plot. This area was not used for recording. The remaining areas were divided into thirty-two 0.5 x 0.5 m quadrats. Quadrats with an area of 0.25 m2 have previously been used to sample S. anglica agg. density (Ranwell & Downing 1960; Taylor & Burrows 1968a; Major & Grue 1997). Each quadrat was surrounded on three sides by treated S. anglica agg., and on the other by a walking area.

2.3. Treatments

- Experimental Control. - Dalapon. - Glyphosate. - Sward cut to 10 cm, Dalapon applied after 6 weeks growth (Cut + Dalapon). - Sward cut to 10 cm, Glyphosate applied after 6 weeks growth (Cut + Glyphosate). - Sward cut to 10 cm (Cut). - Sward cut to 10 cm and covered with black plastic sheeting for 6 months (Cut + Smothered).

2.3.1. Herbicidal treatment

Two herbicides were selected for use in this study, dalapon (2,2 dichloropropionic acid), and glyphosate, without any added surfactants. 27 65m ABCDE FG

BDEFAGC

CGFEBAD55m 5m ECBGFDA

FAGCDEB

GFDACBE

Treatments 5m A = Dalapon B = Sward cut and Glyphosate C = Experimental Control D = Glyphosate E = Sward cut and smother F = Sward cut G = Sward cut and Dalapon

1m wide walking zone

5m 50 x 50cm quadrats

50cm buffer zone

Bamboo cane

5m

Fig 2.3 Random block layout and experimental plot design. 28

2.3.1.1. Herbicide formulation

The recommended application rate of 57 kg ha-1 of dalapon was used in this study (Ranwell & Downing 1960; Taylor & Burrows 1968a; Corkhill 1984; Truscott 1984; Way 1987). The form of dalapon available for use was Farmon Dowpon, a wettable powder containing 85 % of the sodium salt of dalapon plus wetter.

To control grasses in the aquatic environment the manufacturers (Monsanto) recommend an application rate of 5.0 l ha-1 of Roundup Biactive (glyphosate), the same rate used by Garnett et al.(1992). Roundup Biactive is an aqueous concentrate containing 360 g l-1 glyphosate acid present as 480 g l-1 of the isopropylamine salt of glyphosate. Glyphosate (Roundup Biactive) was applied at a rate of 5.0 l ha-1 in this study.

2.3.1.2. Herbicide application

Herbicides were applied using a Cooper Pegler CP15 knapsack sprayer. The sprayer was operated at a pressure of 1 bar (15 psi) and was fitted with a red floodjet / deflector nozzle that had a 2.0 m spray width from a nozzle height of 50 cm above the target. The nozzle used produces a wide swathed, evenly deposited pattern of spray with a low risk of drift. It is designed for broadband application of herbicides and is usually used by Environment and Heritage Service and the Northern Ireland National Trust for herbicidal applications to S. anglica agg.. Herbicides were applied at least 6 hours before tidal inundation during neap tides, on cloudy, rainless days with wind speeds of less than 10 km per hour. Spraying occurred in August as advised by Ranwell (1968) and occurred before S. anglica seedheads had developed. Herbicides were applied at Ballydrain on 18th August and at Ballykelly on 27th August 1998.

2.3.2. Mowing Treatments

In this study, mowing S. anglica agg. to 10 cm followed by glyphosate or dalapon application after approximately 6 weeks re-growth was applied as a S. anglica agg. control method. A single mowing treatment was also tested. 29

Mowing was accomplished by using a hand-held brush cutter. Cutting was done to within 10 cm of the substrate. Cut material was raked to one side. This cut material was removed from the area by subsequent tides. Ballydrain plots were cut on 7th July; Ballykelly plots were cut on 15th July 1998. Follow-up herbicide applications were carried out on August 18th at Ballydrain and August 27th 1998 at Ballykelly.

2.3.3. Plastic sheeting

Industrial strength black plastic sheeting was cut into appropriately-sized strips and laid out onto the plots. Galvanised chicken wire was laid on top of the sheeting, extending beyond the edge of the plastic. Galvanised steel wire pegs were staked through the plastic and chicken wire to hold both layers in place. A similar method has been used for research in Tasmania (Bishop 1995). Ballydrain plots were covered on 17th – 20th July 1998. Ballykelly plots were covered on 21st – 24th July 1998. The sheeting was checked monthly to ensure that it was not dislodged by tidal currents. Throughout this study the plastic sheeting remained in place. The plastic sheeting and wire was removed at Ballydrain on the 18th January 1999, and at Ballykelly on the 22nd January 1999.

2.4. Experimental Records

The first programme of experimental recording was carried out at Ballydrain between 3rd – 6th July. At Ballykelly it was carried between 8th – 11th July 1998, both prior to the application of treatments in July/August 1998. This involved recording vegetation and macro-invertebrates. The recording was repeated in July 1999, one year after the application of the treatments. Earlier sampling may not have shown the effectiveness of the treatments, for example, Taylor & Burrows (1968a) noted that herbicides used in S. anglica agg. control may take as long as 8 months to show full effects due to the slow onset of toxicity. In July 2000, Spartina stem density and abundance of other plant species was recorded. 30

2.4.1. Vegetation recording

Four categories of records were measured. - S. anglica agg. stem density - S. anglica agg. stem heights - Dry weight of roots and rhizomes - Domin values of other plant species excluding S. anglica agg..

2.4.1.1. Quadrat sampling for stem height, stem density and Domin values

Every plot contained thirty-two 50 x 50 cm quadrats. Time constraints meant that only five quadrats per plot were used to record stem height, stem density and Domin values. Each quadrat was assigned a number (1-32). Five numbers were randomly drawn for each plot and the corresponding five quadrats were used for recording in each plot. Different quadrats were selected for each years recording.

2.4.1.2. Stem height measurements

A meter rule was dropped into each quadrat three times; once on the right side once on the left side and once in the middle. At the right and left drop the nearest S. anglica agg. stems to the north, south, east and west of the rule were selected for stem height measurement. The rule was placed close to the stem being measured and the reading of the tallest point of the stem was taken as its height. In the central drop the nearest stems to the north and south were measured. This provided 10 stem height measurements per quadrat and a total of 50 records per plot. The mean of the 50 recorded stem heights per plot was used as the figure representing stem height of the plot thus avoiding sacrificial pseudoreplication.

2.4.1.2.1. Sacrificial Pseudoreplication

Multiple samples/measurements taken from the same experimental unit (plot) are useful for increasing the sensitivity of an experiment by increasing the precision of estimates. Sacrificial pseudoreplication occurs when two or more samples/measurements are taken from the same experimental unit (plot) and are 31 treated as independent replicates (Hurlbert 1984). Such samples/measurements are however are not independent as they are obtained from the same experimental unit (plot), and they do not increase the number of degrees of freedom available for testing a treatment effect. Such data should therefore not be used with statistical techniques that assume independence of samples e.g. ANOVA.

The simplest way to avoid sacrificial pseudoreplication is to use only a single datum (mean of samples/measurements) for each experimental unit (plot) and to omit any analysis of the individual samples/measurement data.

2.4.1.3. Stem density recording

The number of live S. anglica agg. stems in each quadrat was counted. The mean of the five stem density counts per plot was used as the figure representing the stem density of the plot.

2.4.1.4. Domin values

Five 50 x 50 cm quadrats were used per plot to estimate, by eye, percentage cover of all plant species present excluding S. anglica agg.. The mean percentage coverage value from the five quadrats was calculated and converted into a Domin value (Kent & Coker 1995). Domin values were used to represent species coverage in each plot.

2.4.1.5. Dry weight of roots and rhizomes

Three quadrats were randomly selected for core sampling (from the remaining quadrats not used for stem density and stem height recording). A core was taken from the central area of each of these quadrats.

Due to time constraints and practicality, samples of roots and rhizomes were extracted from the same core samples that were used for invertebrate sampling (2.4.3.). A soil corer 10 cm diameter (78.5 cm2 area), by 15 cm depth (1177.5 cm3) was used. Goodman et al. (1969) noted that the slender roots of S. anglica have laterals that spread in all directions in the top 10 cm of substrate. It should however be noted that 32

S. anglica agg. is a deep-rooted plant. For example Goodman et al. (1969) noted that the long white unbranched roots of S. anglica descend 1 m or more into the substrate. The cores taken therefore represent a sample of the upper root and rhizome system.

The cores were randomly taken from each plot and placed into labelled plastic bags. They were transported to a freezer before analysis. In the laboratory the samples were thawed and each core was washed and the root component extracted. This was achieved by placing the cores into a 30 cm diameter brass sieve with a mesh size of 1 mm. This mesh size was small enough to retain the animal and plant material but large enough for the sediments to wash through. A jet of water was used to wash of the soil particles. The roots were removed and placed into a container. The roots were then dried in a fan-assisted oven for 24 hrs at a temperature of 85°C. Prior testing indicated no further weight loss after this time. Samples were taken out of the oven in batches of eight and weighed within two minutes. Prior testing indicated that there was no measurable water absorption by the roots within this time. The weight of the roots was recorded using a top-pan balance accurate to 0.1 grams. The mean of the three root weights was used as the figure representing plot root weight.

2.4.2. Macro-invertebrate analysis

Core depths of 10-20 cm are usually used for inter-tidal macro invertebrate sampling (James & Fairweather 1996; Little 2000). A soil corer 10 cm diameter (78.5 cm2 area), by 15 cm depth (1177.5 cm3) was used to sample invertebrates in this study. A similarly-sized core has previously been used to study invertebrates within a S. alterniflora sward (Simenstad et al. 1996). Holme (1964) noted that the majority of benthic animals in mudflats are found in the upper 15 cm of the substrate. Prior testing indicated similar findings at both the Ballydrain and Ballykelly study sites.

The sediments were washed out of the sample core into a 30 cm diameter by 1 mm mesh. After the roots were removed the remaining plant debris containing the macro- invertebrates was washed into foil trays, diluted with water and frozen before further analysis. At a later date the samples were thawed, washed through a 0.025 mm sieve and examined under a binocular microscope. This sieving retains invertebrates that may have disintegrated during the previous processes, yet allows small sediment 33 particles to pass through, enabling easier examination of the samples. All invertebrates were identified (Chinery 1991; Hayward and Ryland 1998; Smith 1989), counted and stored in 70% alcohol.

Annelids and amphipods were easily broken during the sieving and freezing processes, therefore only the number of heads in the samples were counted. There were large numbers of unoccupied shells in the samples. Only shells containing invertebrate bodies were counted.

Invertebrates were generally identified to species level. Taxa that were difficult to distinguish were identified to a broader taxanomic category e.g. sub-order level for Diptera larvae. The mean number of each invertebrate in the three cores was used as the value indicating the number of each invertebrate per plot.

2.5. Statistical analysis

The data collected contains a number of observed values divided amongst seven treatment groups. The null hypotheses states that there is no change in S. anglica agg. stem density; S. anglica agg. stem height; S. anglica agg. root weight; invertebrate numbers; and Puccinellia maritima abundance due to treatment effects. To test these hypotheses we require statistical techniques to find out if there are significant differences amongst treatment groups. To answer these questions the Kruskal-Wallis test and Analysis of Variance (ANOVA) were used. As the data collected contains only one grouping criterion (Treatment) the ANOVA required was one-way analysis of variance (single-classification analysis of variance). ANOVA was used when possible, as it is a more powerful statistical technique. Analysis of variance and the Kruskal-Wallis test are described extensively by Sokal and Rohlf (1998) and Fowler et al. (1998). Statistical techniques used for each analysis are provided in Table 2.3.

All statistical analysis were carried out using the SPSS Version 9 (1999) statistical computer package. 34

2.5.1. Analysis of Variance (ANOVA, F-test)

ANOVA is a technique used to test for differences amongst two or more group means.

The data derived from treatment groups give rise to two sources of variability. 1) Variability around each mean within a treatment group (within-group variance). 2) Variability between the treatment groups due to differences between the means of the populations from which the groups are drawn (between-group variance). ANOVA divides the total variability of a number of groups into these two components. If samples are from populations with equal means and variances, the within-group variance is the same as the between-group variance. If ANOVA shows that this is not the case then the groups have been drawn from a population with different means and / or variances. If the group variances are similar to each other (an assumption tested prior to ANOVA) it is concluded that the discrepancy is due to a difference between treatment group means.

To test which groups are significantly different from each other and whether the means can be divided into groups that are significantly different from each other a posteriori test is required – the Tukey’s honestly significant difference method (T- method, Tukey Test) (Sokal and Rohlf 1998).

2.5.1.1. Tukey test

The Tukey test calculates a minimum significant difference and is used to test the range of all means, and differences between any pair of means (for equal group sizes only). When sample sizes are unequal (e.g. S. anglica agg. stem height data in this study) modifications of values are required to compute the Tukey test. In such cases the SPSS package uses the harmonic mean of the group sizes. The harmonic mean is used to estimate an average group size when the sample sizes of groups are not equal. The harmonic mean is the total number of samples divided by the sum of the reciprocals of the sample sizes. The Tukey test is available on SPSS and is given in detail by Sokal and Rohlf (1998). 35

2.5.1.2. ANOVA assumptions

ANOVA is a parametric test that makes assumptions about the values tested. These assumptions must be met before an ANOVA can be used.

1) Sampling of individuals is random. 2) Independence: treatments are independently and identically distributed. This is achieved by random allocation of treatments among the experimental plots. 3) Homogeneity of Variances (Homoscedasticity – equal scatter): sample groups must have similar variances. This is determined using a Homogeneity of Variance test. This tests whether the largest and smallest variances of the groups are significantly different from each other. This test is available in SPSS and is described by Sokal and Rohlf (1998) and Fowler et al. (1998). 4) Normal probability distribution: the values from each group must be normally distributed. Normal distribution can be tested in many ways (Sokal and Rohlf 1998, Fowler et al. 1998). The test used in this study was the Shapiro-Wilk statistic, which is available in SPSS, and is used for samples with 50 or fewer observations.

If the assumptions for an ANOVA cannot be met, two courses of action are available. A test can be carried out that does not require the rejected assumptions, such as non- parametric (distribution-free) test (Kruskal-Wallis test) in lieu of ANOVA. Alternatively the data to be analysed may be transformed into a new scale that meets the assumptions of the analysis.

2.5.1.3. Transformations

Transformation means the conversion of the raw data values of all observations into a mathematical derivative. Transformations are used to normalise skewed data and to remove the dependency of the variance upon the mean, thus stabilizing the variance.

A natural logarithmic transformation converts the data into logarithms. It is best used when the mean is positively correlated with variance, when the variance of a sample is larger than the mean, or when the data is skewed to the right. If zero counts are 36 included in the data, a value of one should be added to each variate prior to natural logarithmic transformation.

A square-root transformation is an appropriate transformation to use with count data when the variance is about equal to the mean or when there is a Poisson distribution. If there are zero counts in the data a value of 0.5 should be added to each variate prior to square root transformation.

Sokal and Rohlf (1998) and Fowler et al. (1998) describe other transformations.

When transformations are unable to make the data conform to the assumptions of ANOVA a non-parametric test in lieu of ANOVA must be used (Kruskal-Wallis test).

2.5.2. Kruskal-Wallis test

The Kruskal-Wallis test is a non-parametric test used to compare the sum of ranks of three or more groups. The test is a less powerful statistic than analysis of variance because it is not concerned with specific parameters (the mean in an analysis of variance) but only with the distribution of variates. The test however does not require the data to be normally distributed and can be used on any data that allows ranking to occur (interval scale measurements, counts, derived variables or ordinal ranks).

In the test, all variates are pooled into a single sample and ranked. These ranked values then replace the original data for group comparisons. If the groups are not different from each other we expect their rank scores to be similar. If the outcome of the test shows a significant difference between groups, it indicates that the two groups with the highest and lowest sum of ranks are significantly different. There is no post priori test to indicate any further significant differences between the groups. Care must be taken when making inferences about differences between particular pairs of groups or between one sample and the others. When making comparisons ‘common sense’ must therefore be used.

Sokal and Rohlf (1998) and Fowler et al. (1998) demonstrate the use of the Kruskal- Wallis test. 37

Table 2.3 Statistical techniques used for each analysis

Analysis Study area Year Statistical techniques

S. anglica agg. stem density Ballykelly 1998 ANOVA S. anglica agg. stem density Ballykelly 1999 Natural Logarithmic transformation for zero counts ANOVA + Tukey Test S. anglica agg. stem density Ballykelly 2000 Square-root transformation for zero counts ANOVA + Tukey Test S. anglica agg. stem density Ballydrain 1998 ANOVA S. anglica agg. stem density Ballydrain 1999 Kruskal-Wallis test S. anglica agg. stem density Ballydrain 2000 Kruskal-Wallis test

S. anglica agg. stem height Ballykelly 1998 ANOVA S. anglica agg. stem height Ballykelly 1999 ANOVA S. anglica agg. stem height Ballydrain 1998 ANOVA S. anglica agg. stem height Ballydrain 1999 ANOVA

S. anglica agg. root weight Ballykelly 1998 ANOVA S. anglica agg. root weight Ballykelly 1999 ANOVA + Tukey Test S. anglica agg. root weight Ballydrain 1998 ANOVA S. anglica agg. root weight Ballydrain 1999 ANOVA

P. maritima abundance Ballykelly 1998 Kruskal-Wallis test P. maritima abundance Ballykelly 1999 Kruskal-Wallis test P. maritima abundance Ballykelly 2000 Kruskal-Wallis test P. maritima abundance Ballydrain 1998 Kruskal-Wallis test P. maritima abundance Ballydrain 1999 Kruskal-Wallis test P. maritima abundance Ballydrain 2000 Kruskal-Wallis test 38

Table 2.3 continued. Statistical techniques used for each analysis

Analysis Area Year Statistical techniques used

Other species abundance Ballykelly 1998 Kruskal-Wallis tests Other species abundance Ballykelly 1999 Kruskal-Wallis tests Other species abundance Ballykelly 2000 Kruskal-Wallis tests Other species abundance Ballydrain 2000 Kruskal-Wallis test

Invertebrate numbers Ballykelly 1998 Kruskal-Wallis tests Invertebrate numbers Ballykelly 1999 Kruskal-Wallis tests Invertebrate numbers Ballydrain 1998 Kruskal-Wallis tests Invertebrate numbers Ballydrain 1999 Kruskal-Wallis tests 39

3. Results

3.1. Live Spartina anglica agg. stem density

Mean live S. anglica agg. stem density and standard deviation of each treatment group at Ballykelly and Ballydrain in July 1998, 1999 and 2000 are shown in Figures 3.1 – 3.6.

ANOVA showed that there was no significant difference between the mean live S. anglica agg. stem density of the seven treatment groups prior to treatment application at both sites (Table 3.1). The mean live S. anglica agg. stem density was 58 live stems per 0.25 m2 at Ballykelly and 84 live stems per 0.25 m2 at Ballydrain (Figures 3.1 and 3.4).

ANOVA was carried out on 1999 Ballykelly data that had been transformed by a natural Log (variate + 1) transformation. The ANOVA indicated a significant difference (p < 0.001) amongst the mean live S. x townsendii stem densities of the treatment groups (Table 3.1). Tukeys test (Table 3.2) identified two significantly different sub-sets of treatments. The treatments Dalapon, Cut + Dalapon and Cut + Smothered were in one sub-set and had similar live stem density means, all of which were below three live stems per 0.25 m2. These live stem densities were smaller than the other sub-set which contained the treatments Glyphosate, Cut + Glyphosate, Cut, and Experimental Control. This sub-set contained live stem density means of over 47 live stems per 0.25 m2. Within this sub-set the Cut + Glyphosate, and Cut treatments had higher live stem densities than the Glyphosate treatment and the Experimental Control. There were however no significant differences between treatments within each sub-set.

The Ballykelly 2000 live stem density data was transformed using a square root (variate + 0.5) transformation. ANOVA indicated a significant difference (p<0.001) amongst the mean live S. x townsendii stem densities of the treatment groups (Table 3.1). Tukeys test (Table 3.3) identified four different sub-sets of treatments. The Cut + Smothered treatment plots contained the lowest mean live stem density (6 stems per 0.25 m2) and was in a sub-set on its own. The Dalapon and Cut + Dalapon treatments 40 are in another sub-set with mean live stem densities of 25 and 27 per 0.25 m2 respectively. These mean live stem densities were higher than the Cut + Smother treatment, and lower than the other treatments. The Cut, Cut + Glyphosate and the Control treatments are contained in a sub-set with mean live stem densities of over 55 per 0.25 m2. The other sub-set contains the Control and the Glyphosate treatments, both with mean live stem densities of over 80 per 0.25 m2. The Control treatment was contained in two different sub-sets. There were therefore no significant differences between the Control treatment and the other treatments in these two sub-sets (Cut, Cut + Glyphosate and Glyphosate).

The Kruskal-Wallis test was carried out on the 1999 Ballydrain live S. anglica stem density data (Table 3.4). The test indicated a significant difference between the treatments containing the largest and smallest mean rank. These corresponded with the Cut + Dalapon treatment (mean rank 8) and the Cut treatment (mean rank 34). The Cut + Dalapon treatment had similar mean rank scores as the Dalapon and Cut + Smothered treatments (comparable to Ballykelly 1999 Tukey test). The cut treatment had similar mean rank scores as the Glyphosate, Cut and Glyphosate and Control treatments (comparable to Ballykelly 1999 Tukey test).

The Kruskal-Wallis test was carried out on the 2000 Ballydrain live S. anglica stem density data (Table 3.4). The test indicated a significant difference between the treatments containing the largest and smallest mean rank. These corresponded with the Cut + Smother treatment (mean rank 5) and the Cut treatment (mean rank 34). The results are similar to the previous year with the Cut + Dalapon, Dalapon and Cut + Smothering treatments having similar mean ranks (between 5 and 13). The remaining treatments (Cut, Cut + Glyphosate, Glyphosate and Control), had higher mean ranks of between 28 to 34.

3.2. S. anglica agg. stem height

Mean S. anglica agg. stem height and standard deviation of each treatment group at Ballykelly and Ballydrain in 1998 and 1999 are shown in Figures 3.7 – 3.10. 41

ANOVA showed that there were no significant differences between mean S. anglica agg. stem height of the seven treatment groups prior to treatment application at both sites (Table 3.5). Mean S. anglica agg. stem height was 33 cm at Ballykelly and 24 cm at Ballydrain.

By July 1999 ANOVA indicated a significant difference (p < 0.001) in the mean S. anglica agg. stem heights amongst the treatment groups at both Ballykelly and Ballydrain (Table 3.5)

At Ballykelly four significantly different sub-sets were identified by a Tukey test (Table 3.6). All experimental treatment groups contained S. x townsendii with significantly smaller mean stem heights than the Experimental Control (45 cm). The treatment groups that contained S. x townsendii with the smallest mean stem heights were Dalapon, Cut + Dalapon, and Cut + Smother (11 cm, 13 cm and 11 cm respectively). The other treatments contained S. x townsendii with mean stem heights between these two sub-sets. The Cut treatment formed a significantly different sub-set containing S. x townsendii with a mean stem height of 36 cm. The remaining treatments Glyphosate and Cut + Glyphosate formed another significantly different sub-set with a mean S. townsendii stem height of 24 cm and 26 cm respectively.

At Ballydrain three significantly different sub-sets were obtained using Tukeys test (Table 3.7). One sub-set contained the Experimental Control and the Glyphosate treatments. All of the treatment groups contained S. anglica with a significantly smaller mean stem height than the Experimental Control (24 cm), except the Glyphosate treatment (20 cm). There was no significant difference between the mean S. anglica stem height of the Glyphosate treatment and the Cut + Glyphosate, Cut, and Dalapon treatments (16 cm, 16 cm and 18 cm respectively). These treatment groups formed another significantly different sub-set. The remaining Cut + Smother and Cut + Dalapon treatments formed another significantly different sub-set and contained S. anglica with the smallest mean stem heights (both 9 cm). 42

3.3. Root dry weight

Mean root dry weight and standard deviation of each treatment group at Ballykelly and Ballydrain in July 1998 and 1999 are shown in Figures 3.11 – 3.14.

ANOVA showed that there were no significant differences between the mean root dry weights of the seven treatments groups, prior to treatment application at both sites (Table 3.8). The mean root dry weight in July 1998 was 12 g per core at Ballykelly, and 11 g per core at Ballydrain.

By July 1999 (one year after treatment application), ANOVA showed a significant difference (p < 0.05) in mean dry weight amongst the treatment groups (Table 3.8). Tukeys test (Table 3.9) indicated that the mean root dry weight of the Cut + Smothered treatment (10 g per core) was significantly different from the mean dry root weight of the Experimental Control plots (16 g per core). There were no other significant differences between any other combination of treatments. However the mean root dry weights of the Ballykelly treatment groups had all increased over the year, with the exception of the Cut + Smothered treatment, which showed a decline from 11 g to 10 g per core (Figures 3.11 and 3.12).

A similar pattern of change in mean root dry weight was observed at Ballydrain. One year after treatment application there was an increase in mean root dry weight in all treatment groups, except the Cut + Smothered treatment in which mean root dry weight remained at 12 g per core. ANOVA however showed that there was no significant difference between mean root dry weight of the treatment groups (Table 3.8)

3.4. Puccinellia maritima abundance

Median Domin scale and inter-quartile ranges of Puccinellia maritima abundance within each treatment group at Ballykelly and Ballydrain in July 1998, 1999 and 2000 are shown in Figures 3.15 – 3.20. 43

At Ballykelly a Kruskal-Wallis test showed that there was no significant differences in P. maritima abundance between the seven different treatment groups in 1998 prior to treatment application (Table 3.10).

In 1999 (one year after treatment application), a Kruskal-Wallis test indicated a significant difference between the treatments containing the smallest and largest mean rank (Table 3.10). These corresponded with the Dalapon treatment (mean rank 12.08) and the Cut + Glyphosate treatment (mean rank 34.5). Since 1998 the largest reductions in median P. maritima abundance occurred in the Dalapon, Glyphosate and the Cut + Smother treatments (Figures 3.15 and 3.16). In the Experimental Control plots and the Cut + Dalapon plots smaller reductions in median P. maritima abundance were noted. The only treatments in which median P. maritima abundance increased were the Cut + Glyphosate and Cut treatments.

In 2000 (two years after treatment application), a Kruskal-Wallis test indicated that there was no significant difference in P. maritima abundance between the treatment groups (Table 3.10). All of the treatment groups experienced an increase in P. maritima abundance compared with 1999, except the Cut treatment which showed a slight decline and the Experimental Control which remained similar to 1999 levels (Figures 3.16 and 3.17).

At Ballydrain P. maritima was recorded as present in only one experimental plot in 1998, 1999 and 2000. A Kruskal-Wallis test showed no significant differences between P. maritima abundance between treatment groups in 1998, 1999 and 2000 (Table 3.11). Due to the scattered distribution of P. maritima at Ballydrain it is difficult to infer about the effect of the different treatments on P.maritima growth.

3.5. Other plant species

At Ballykelly individuals of Aster tripolium and Plantago maritima were recorded in plots during July 1998 (Table 3.12). An increase in abundance of these species was recorded in July 1999 and July 2000. There were no significant differences in the abundance of A. tripolium and P. maritima between treatment groups. An individual of Chenopodium rubrum was recorded in July 2000. 44

At Ballydrain no other species apart from S. anglica and Puccinellia maritima were recorded in the plots during July 1998 or July 1999. In July 2000 low abundances (maximum 1-4% coverage) of Salicornia spp. were recorded in the Dalapon, Cut + Dalapon, and Cut + Smother plots (Table 3.13). There were no significant differences between Salicornia spp. abundance and treatment group.

3.6. Invertebrate abundance

The most abundant invertebrate species recorded at Ballykelly were Corophium volutator Pallas, Hydrobia ulvae Pennant and Nereis (Hediste) diversicolor O.F. Muller (Table 3.14). Other less abundant invertebrates recorded were Acteon tornatilis Linnaeus, Araneae (Order), Brachycera larvae (Sub-order), Cerastoderma edule Linnaeus, Chaetogammarus marinus Leach, Coleoptera (Order), Cyclorrhapha pupae (Sub-order), Hirudinea (Class), Littorina littorea Linnaeus, Macoma balthica Linnaeus, Miridae (Family), Nematocera larvae (Sub-order) and Retusa obtusa Montagu (Table 3.14). Kruskal-Wallis tests for each invertebrate type showed that there were no significant differences between invertebrate numbers within the seven treatment groups in 1998 prior to treatment application and in 1999, one-year after treatment applications (Table 3.14). (Invertebrate nomenclature obtained from Chinery 1991; Hayward & Ryland 1998; Smith 1989).

The most abundant invertebrates recorded at Ballydrain were Corophium volutator Pallas, Hydrobia ulvae Pennant and Nematocera larvae (Sub-order) (Table 3.15). Other less abundant invertebrates recorded were Brachycera larvae (Sub-order), Carcinus maenas Linnaeus, Chaetogammarus marinus Leach, Cyclorrhapha pupae (Sub-order), Littorina littorea Linnaeus, Macoma balthica Linnaeus, Mytilus edulis Linnaeus, Nereis diversicolor O.F. Muller and Sphaeromatidae (Family) (Table 3.15). Kruskal-Wallis tests for each invertebrate type showed that there were no significant differences between invertebrate numbers within the seven treatment groups in 1998 prior to treatment application and in 1999, one-year after treatment applications (Table 3.15). (Invertebrate nomenclature obtained from Hayward & Ryland 1998; Smith 1989). Figure 3.1 Mean live Spartina x townsendii stem density and standard deviation per 0.25m2 within each treatment group at Ballykelly in July 1998

90 66.1 80 53.9 57.9 57.7 59.8 56.7 55.9 70 55.6 60 50 40 30 20 Live stem density 10 0 Cut Total Cut + Cut + Dalapon Cut + Dalapon Smother Control Glyphosate Glyphosate Experimental Treatment

Figure 3.2 Mean live Spartina x townsendii stem density and standard deviation per 0.25m2 within each treatment group at Ballykelly in July 1999

120 86.7 85.6 100 80 48 47.6 60 40

Live stem density 20 2.1 0.6 2.2 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Figure 3.3 Mean live Spartina x townsendii stem density and standard deviation per 0.25m2 within each treatment group at Ballykelly in July 2000

120 94.5 82.4 100 59.6 80 58.7

60 25.4 40 27.7

Live stem density 20 5.4

0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Figure 3.4 Mean live Spartina anglica stem density and standard deviation per 0.25m2 within each treatment group at Ballydrain in July 1998

160 91.8 100.6 140 89.2 74.7 85 84 78.1 120 68.2 100 80 60 40

Live stem density 20 0 Cut Total Cut + Cut + Dalapon Cut + Dalapon Smother Control Glyphosate Glyphosate Experimental Treatment

Figure 3.5 Mean live Spartina anglica stem density and standard deviation per 0.25m2 within each treatment group at Ballydrain in July 1999

140 76.5 120 100 50.9 80 47.7 40.6 60 40

Live stem density 20 3.9 0.4 0.1 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Figure 3.6 Mean live Spartina anglica stem density and standard deviation per 0.25m2 within each treatment group at Ballydrain in July 2000

140 79.7 120 60.5 100 59.5 80 50.2 60

40 Live stem density 7.1 20 1.8 0.1 0 Cut Cut + Cut + Dalapon Cut + Dalapon Smother Control Glyphosate Glyphosate Experimental Treatment Table 3.1 Live Spartina anglica agg. stem density Analysis of Variance

Study area and Year df F Significance Ballykelly 1998 Between groups 6 0.732 0.627 Ballykelly 1999 Between groups 6 130.368 0.000*** Log (variate + 1) Ballykelly 2000 Between groups 6 35.729 0.000*** Sqroot (variate + 0.5)

Ballydrain 1998 Between groups 6 0.487 0.813 Ballydrain 1999 Not Valid - - - Ballydrain 1999 Not Valid - - - Significance level: * P < 0.05, ** P < 0.01, *** P < 0.001

Table 3.2 Ballykelly 1999 live Spartina x townsendii stem density (log (variate + 1)) Tukey-test sub-sets

Subset at 0.5 significance Treatment N 1 2 Cut + Smother 6 0.6 Dalapon 6 2.2 Cut + Dalapon 6 2.1 Glyphosate 6 47.6 Experimental Control 6 48.0 Cut + Glyphosate 6 85.6 Cut 6 86.7 Significance 0.055 0.117 Un-transformed means for groups in homogenous subsets are displayed.

Table 3.3 Ballykelly 2000 live Spartina x townsendii stem density (square-root (variate + 0.5)) Tukey-test sub-sets

Subset at 0.5 significance Treatment N 1 2 3 4 Cut + Smother 6 5.4 Dalapon 6 25.4 Cut + Dalapon 6 27.7 Cut 6 58.7 Cut + Glyphosate 6 59.6 Control 6 82.4 82.4 Glyphosate 6 94.5 Significance 1.000 0.989 0.290 0.941 Un-transformed means for groups in homogenous subsets are displayed. Table 3.4 Ballydrain 1999 and 2000 live Spartina anglica stem density Kruskal- Wallis test

Treatment Mean Rank 1999 Mean Rank 2000 Cut 34.00 34.00 Cut + Dalapon 7.50 10.75 Cut + Glyphosate 29.50 29.33 Cut + Smother 8.25 4.75 Dalapon 12.75 13.25 Experimental Control 30.00 27.83 Glyphosate 28.50 30.58 Chi-Square 32.059 32.221 Df 6 6 Significance 0.000*** 0.000*** Significance level: * P < 0.05, ** P < 0.01, *** P < 0.001 Figure 3.7 Mean Spartina x townsendii stem height and standard deviation within each treatment group at Ballykelly in July 1998

50 45 40 32.6 34.6 32.9 34.7 32.4 32.9 32 33.2 35 30 25 20 15 10 Stem height (cm) 5 0 Cut Total Cut + Cut + Dalapon Cut + Dalapon Smother Control Glyphosate Glyphosate Experimental Treatment

Figure 3.8 Mean Spartina x townsendii stem height and standard deviation within each treatment group at Ballykelly in July 1999

50 45.2 35.6 40 26.3 23.7 30

20 13.4 11.3 11.4

Stem height (cm) 10

0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Figure 3.9 Mean Spartina anglica stem height and standard deviation within each treatment group at Ballydrain in July 1998

35 24.7 30 23.3 24.6 24.2 24.3 22.8 22.4 23.7 25 20 15 10

Stem height (cm) 5 0 Cut Total Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment

Figure 3.10 Mean Spartina anglica stem height and standard deviation within each treatment group at Ballydrain in July 1999

30 24.3 25 17.5 19.6 20 15.8 15.6 15 8.5 8.5 10

Stem height (cm) 5 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Table 3.5 Mean Spartina anglica agg. stem height Analysis of Variance

Area and Year df F Significance Ballykelly 1998 Between groups 6 0.899 0.506 Ballykelly 1999 Between groups 6 81.378 0.000***

Ballydrain 1998 Between groups 6 0.468 0.828 Ballydrain 1999 Between groups 6 14.327 0.000*** Significance level, * P < 0.05, ** P < 0.01, *** P < 0.001

Table 3.6 Ballykelly 1999 Spartina x townsendii stem height Tukey-test sub-sets

Subset at 0.05 significance Treatment N 1 2 3 4 Cut + Smother 4 11.3 Dalapon 6 11.4 Cut + Dalapon 6 13.4 Cut + Glyphosate 6 23.7 Glyphosate 6 26.3 Cut 6 35.6 Experimental Control 6 45.2 Significance 0.954 0.881 1.0 1.0 Means for groups in homogenous subsets are displayed. The group sizes are unequal. The harmonic mean of the group sizes is used. Type 1 error levels are not guaranteed. Harmonic Mean Sample Size = 5.6.

Table 3.7 Ballydrain 1999 Spartina anglica stem height Tukey-test sub-sets

Subset at 0.05 significance Treatment N 1 2 3 Cut + Smother 3 8.5 Cut + Dalapon 4 8.5 Cut + Glyphosate 6 15.6 Cut 6 15.8 Dalapon 6 17.5 Glyphosate 6 19.6 19.6 Experimental Control 6 24.3 Significance 1.0 0.471 0.266 Means for groups in homogenous subsets are displayed. The group sizes are unequal. The harmonic mean of the group sizes is used. Type 1 error levels are not guaranteed. Harmonic Mean Sample Size = 4.9. Figure 3.11 Mean dry root weight and standard deviation per core within each treatment group at Ballykelly in July 1998

20 18 14.5 12.6 11.8 13.9 12.2 16 10.6 11 14 11.1 12 10 8 6 4 Dry root weight (g) 2 0 Cut Total Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment

Figure 3.12 Mean dry root weight and standard deviation per core within each treatment group at Ballykelly in July 1999

20 15.8 13.8 12.7 14.6 13.6 14 15 10.2

10

5 Dry root weight (g) 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Figure 3.13 Mean dry root weight and standard deviation per core within each treatment group at Ballydrain in July 1998

18 12 16 11.6 11 11.2 11.1 14 11.3 10.5 10.3 12 10 8 6 4

Dry root weight (g) 2 0 Cut Total Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment

Figure 3.14 Mean dry root weight and standard deviation per core within each treatment group at Ballydrain in July 1999

20 14 18 13.4 14.4 11.4 12.4 16 12.1 12 14 12 10 8 6 4 Dry root weight (g) 2 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Table 3.8 Dry root weight Analysis of Variance

Area and Year df F Significance Ballykelly 1998 Between groups 6 2.001 0.092 Ballykelly 1999 Between groups 6 2.515 0.040*

Ballydrain 1998 Between groups 6 0.334 0.914 Ballydrain 1999 Between groups 6 1.236 0.312 Significance level, * P < 0.05, ** P < 0.01, *** P < 0.001

Table 3.9 Ballykelly 1999 dry root weight Tukey-test sub-sets

Subset at 0.5 significance Treatment N 1 2 Cut + Smother 6 10.2 Cut + Dalapon 6 12.7 12.7 Dalapon 6 13.6 13.6 Cut 6 13.8 13.8 Glyphosate 6 14.0 14.0 Cut + Glyphosate 6 14.6 14.6 Experimental Control 6 15.8 Significance 0.100 0.453 Means for groups in homogenous subsets are displayed. Figure 3.15 Median and inter-quartile Puccinellia maritima abundance (Domin scale) within each treatment group at Ballykelly in July 1998

10 9 8 7 6 5 4 3 Domin scale 2 1 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment

Figure 3.16 Median and inter-quartile Puccinellia maritima abundance (Domin scale) within each treatment group at Ballykelly in July 1999

10 9 8 7 6 5 4 3 Domin scale 2 1 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Figure 3.17 Median and inter-quartile Puccinellia maritima abundance (Domin scale) within each treatment group at Ballykelly in July 2000

10 9 8 7 6 5 4 Domin scale 3 2 1 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Figure 3.18 Median and inter-quartile Puccinellia maritima abundance (Domin scale) within each treatment group at Ballydrain in July 1998

2 1.8 1.6 1.4 1.2 1 0.8 0.6 Domin scale 0.4 0.2 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment

Figure 3.19 Median and inter-quartile Puccinellia maritima abundance (Domin scale) within each treatment group at Ballydrain in July 1999

2 1.8 1.6 1.4 1.2 1 0.8 0.6 Domin scale 0.4 0.2 0 Cut Cut + Cut + Dalapon Dalapon Cut + Smother Control Glyphosate Glyphosate Experimental Treatment Figure 3.20 Median and inter-quartile Puccinellia maritima abundance (Domin scale) within each treatment group at Ballydrain in July 2000

0.3

0.25

0.2

0.15

Domin scale 0.1

0.05

0 Cut Cut + Cut + Dalapon Cut + Dalapon Smother Control Glyphosate Glyphosate Experimental Treatment

Table 3.10 Kruskal-Wallis results for Puccinellia maritima abundance at Ballykelly 1998, 1999 and 2000

Mean Rank Treatment Ballykelly Ballykelly Ballykelly 1998 1999 2000 Cut 21.75 29.25 25.00 Cut + Dalapon 14.00 14.08 15.50 Cut + Glyphosate 20.67 34.50 32.08 Cut + Smother 26.83 18.67 21.25 Dalapon 19.17 12.08 19.25 Experimental Control 18.42 19.67 16.17 Glyphosate 29.67 22.25 21.25 Chi-Square 6.795 16.145 7.880 df 6 6 6 Significance 0.340 0.013* 0.247 Significance level: * P < 0.05, ** P < 0.01, *** P < 0.001

Table 3.11 Kruskal-Wallis results for Puccinellia maritima abundance at Ballydrain 1998, 1999 and 2000

Mean Rank Treatment Ballydrain Ballydrain Ballydrain 1998 1999 2000 Cut 21.00 21.00 21.00 Cut + Dalapon 21.00 21.00 21.00 Cut + Glyphosate 21.00 21.00 21.00 Cut + Smother 21.00 21.00 21.00 Dalapon 24.50 24.50 21.00 Experimental Control 21.00 21.00 21.00 Glyphosate 21.00 21.00 24.50 Chi-Square 6.000 6.000 7.880 df 6 6 6 Significance 0.423 0.423 0.247 Significance level: * P < 0.05, ** P < 0.01, *** P < 0.001 Table 3.12. Maximum abundance (Domin scale) of Aster tripolium, Plantago maritima and Chenopodium rubrum recorded in Ballykelly treatment plots in July 1998, 1999 and 2000. Results of Kruskal-Wallis tests are also given.

JULY 1998 JULY 1999 JULY 2000 TREATMENT A. tripolium P. maritima A. tripolium P. maritima A. tripolium P. maritima C. rubrum Cut - - - - 1 1 1 Cut + Dalapon - - 1 1 1 3 - Cut + Smother ------Cut + glyphosate 1 - - - - 1 - Dalapon - 1 - - 1 1 - Experimental control - 1 - 1 - - - Glyphosate - - - 1 1 - - Kruskal-Wallis test 6.0 5.1 12.3 7.0 5.32 9.5 6.0 Significance 0.42 0.53 0.06 0.32 0.50 0.15 0.42 Table 3.13 Maximum abundance (Domin scale) of Salicornia spp. recorded in Ballydrain treatment plots in July 1998, 1999 and 2000. Results of Kruskal-Wallis tests are also given.

Salicornia spp. TREATMENT 1998 1999 2000 Cut - - - Cut + dalapon - - 1 Cut + glyphosate - - - Cut + smother - - 3 Dalapon - - 1 Experimental control - - - Glyphosate - - - Kruscal Wallis test - - 8.5 Significance - - 0.21

Table 3.14 Mean invertebrate abundance per core and Kruskal-Wallis test results for Ballykelly 1998 and 1999

Total Mean (+/- 1 standard deviation) A. tornatilis Araneae Brachycera C. edule C. marinus 1998 0.04 (0.2) 0.04 (0.1) 0.24 (0.4) 0.01 (0.05) 0.11 (0.3) Significance 0.649 0.374 0.164 0.423 0.530 between treatments 1999 0.01 (0.1) 0.02 (0.1) 0.31 (0.4) - 0.06 (0.2) Significance 0.528 0.649 0.091 - 0.220 between treatments

Total Mean (+/- 1 standard deviation) Coleoptera C. volutator Cyclorrhapha Hirundina 1998 - 6.04 (6.8) 0.04 (0.1) - Significance - 0.941 0.778 - between treatments 1999 0.01 (0.1) 1.97 (2.7) 0.06 (0.2) 0.02 (0.1) Significance 0.528 0.186 0.687 0.167 between treatments

Total Mean (+/- 1 standard deviation) H. ulvae L. littorea M. balthica Miridae 1998 7.63 (6.4) 0.02 (0.1) 0.06 (0.2) 0.01 (0.05) Significance 0.404 0.167 0.362 0.423 between treatments 1999 30.77 (22.5) 0.04 (0.1) 0.05 (0.1) - Significance 0.537 0.336 0.720 - between treatments

Total Mean (+/- 1 standard deviation) Nematocera N. diversicolor R. obtusa 1998 0.01 (0.05) 0.48 (0.8) 0.01 (0.1) Significance 0.423 0.845 0.528 between treatments 1999 0.14 (0.3) 0.31 (0.6) 0.15 (0.3) Significance 0.098 0.541 0.507 between treatments Table 3.15 Mean invertebrate abundance per core and Kruskal-Wallis test results for Ballydrain 1998 and 1999

Total Mean (+/- 1 standard deviation) Brachycera C. marina C. marinus C. volutator 1998 0.15 (0.4) 0.04 (0.1) 0.09 (0.3) 2.27 (4.3) Significance 0.357 0.778 0.083 0.961 between treatments 1999 0.31 (0.6) 0.04 (0.1) 0.28 (0.7) 8.97 (14.7) Significance 0.141 0.374 0.105 0.745 between treatments

Total Mean (+/- 1 standard deviation) Cyclorrhapha H. ulvae L. littorea M.balthica 1998 0.04 (0.1) 47.16 (51.2) 0.33 (0.6) 0.09 (0.2) Significance 0.423 0.464 0.462 0.781 between treatments 1999 0.04 (0.1) 18.87 (25.2) 0.06 (0.2) 0.02 (0.1) Significance 0.778 0.291 0.504 0.167 between treatments

Total Mean (+/- 1 standard deviation) M.edulis Nematocera N. diversicolor Sphaeromatidae 1998 0.05 (0.2) 0.75 (1.1) 0.20 (0.3) 0.01 (0.1) Significance 0.477 0.392 0.741 0.423 between treatments 1999 0.01 (0.05) 17.51 (12.8) 0.37 (0.4) - Significance 0.423 0.147 0.576 - between treatments 66

4. Discussion

4.1. Effects of control treatments on Spartina anglica agg.

Ranwell & Downing (1960) treated S. anglica with 22, 34, 45 and 56 kg ha-1 of dalapon at Bridgewater Bay, England in August 1956. They reported 76, 48, 84 and 84% S. anglica kill respectively by 1957. Dalapon applied at 56 kg ha-1 in July and a split dose of 34 and 22 kg ha-1 resulted in 100% kill in subsequent trials. Taylor & Burrows (1968a) reported that dalapon application resulted in 99-100% kill of S. anglica agg. when applied at a rate of 56 kg ha-1 or 112 kg ha-1 in summer months. Two years after application there were still over 98 % reductions in S. anglica agg. stem density from the 112 kg ha-1 treatments and 88 - 94 % reductions from the 56 kg ha-1 dose. High S. anglica kill rates when using dalapon have been noted by other scientists and field managers (Corkhill 1984; Truscott 1984; Way 1987). In this study initial reductions in live stem density reached over 95 % at both sites within a year. A few reports have achieved poor results when using dalapon (Corkhill 1984; Way 1987). In these cases timing of application, wind speed, position of tides, droplet size and a bad batch of herbicide were all mentioned as possible suspects in producing the low kill rate. It has also been noted that herbicides including dalapon are not effective against seedlings (Taylor & Burrows 1968b; Way 1987). It is thought that these seedlings are covered quickly by tides or have leaves that are to small for herbicide penetration.

These studies indicate that when used in suitable conditions, dalapon will cause over 90 % reduction in S. anglica agg. stem density within the first year. Ranwell & Downing (1960) however reported the complete recovery of S. anglica within sprayed areas two years after treatment and Taylor & Burrows (1968a) observed seedling invasions into their controlled areas. In my study, rapid S. anglica agg. re-growth and re-colonisation by seedlings (at Ballydrain) occurred during the second year after dalapon application. Live S. anglica agg. stem density had increased by over 1200 % at Ballykelly and over 200 % at Ballydrain compared to records from the previous years. The substrate conditions remaining after S. anglica agg. eradication by dalapon are therefore suitable for S. anglica agg. re-colonisation. This may be aided by root 67 abundance remaining similar to uncontrolled areas and the rapid breakdown of herbicide residues in the soil (Hutson & Roberts 1987; Howard 1991).

There have also been studies on the effectiveness of glyphosate, the herbicide used most frequently for Spartina spp. control. One report using glyphosate (Rodeo) and R-11 surficant indicated consistent S. alterniflora stem reductions of over 90 % (Crockett 1997). Most other reports from scientists and field managers have shown variability in glyphosate effectiveness with and without adjuvants (Way 1987; Aberle 1990; Garnett et al. 1992; Kilbride et al. 1995; Pritchard 1995; Shaw & Gosling 1995). Garnett et al. (1992) applied 1.8 kg ha-1 glyphosate with and without adjuvant Mixture B (2 % of spray solution) in summer months at 3 sites. At the Dyfi site S. anglica was covered within 5 hours of glyphosate application. Assessment of S. anglica kill was not made due to lack of visible effect. At Lindisfarne glyphosate reduced stem density by 17 % whereas glyphosate + Mixture B resulted in 74 % reduction in stem density. Summer applications of glyphosate at Dee Estuary caused 8 – 55 % reductions in S. anglica stem density compared to 37 – 89 % reduction using Mixture B + glyphosate. Kilbride et al. (1995) reported 84 % reduction in stem density using ground applications of Rodeo (5 % solution) and X-77 spreader (1 % solution) applied at a rate of between 4.7 - 8.8 l ha-1. Helicopter applications (glyphosate 4.7 l ha-1 + X-77 spreader at 0.9 l ha-1) however had no effect on Spartina spp. stem density. Several reports have indicated low levels or no control using glyphosate (Way 1987; Aberele 1990; Pritchard 1995; Shaw & Gosling 1995).

In my study it was noted that glyphosate caused visible browning of the S. anglica agg. leaves, but did not have effects on live stem density the following year. This suggests that glyphosate killed the above ground growth but did not penetrate or kill the S. anglica agg. root and rhizome system. There are no environmental variables that can account for the poor effectiveness of glyphosate reported in this study. Dalapon was applied at the same time as the glyphosate treatments and it proved to be effective. This study combined with previous reports confirms that glyphosate is at present unreliable for use in S. anglica agg. eradication.

Prior cutting would allow plants underneath the main S. anglica agg. canopy to receive herbicide applications. It would also cause S. anglica agg. to use up energy 68 reserves on the production of new shoots. Cutting had no additive effect when applied before herbicides in my study. Preliminary results of an experiment using a single cut followed by Rodeo application suggested a similar outcome (Major & Grue 1997). This indicates that herbicide effectiveness is the main factor influencing reduction of Spartina spp. stem density using these treatments. If the herbicide is effective at penetrating and killing roots and rhizomes, it as unnecessary for all stems to receive herbicide application (Taylor & Burrows 1968b). The death of the root and rhizome system would result in the death of the majority of stems and would prevent re-growth the following year. Only new seedlings would remain.

The single-cut treatments produced the highest live stem density values at each site in my experiment (not significantly different from other treatments). Hubbard (1970) cut S. anglica to the ground every month from June to October 1962. There was an observed increase in stem density by April 1963 (<500 stems m2 in experimental control plots and >1500 stems m2 in the cut plots). The plants were more uniform in height and flowered earlier than untreated areas. Similar observations were made in my study. A single cut will therefore have no impact on S. anglica agg. eradication, and may in-fact be more suitable for rehabilitation of native Spartina spp. e.g. after oil pollution (Baker et al. 1990).

At Ballykelly, live S. x townsendii stem density in cut plots was significantly lower than the experimental control after two years. This may indicate that energy reserves were extensively used up in the year following the cut. It has been suggested that multiple cutting may reduce Spartina spp. vigour and reduce above ground biomass (Turner 1987; Scott et al. 1990; Norman and Patten 1997). Over 90 % kill has been achieved by mowing Spartina spp. to the mud surface 2 - 4 times (Norman & Patten 1997). This high reduction may be due to the cuts being made at the mud surface rather than above the mud surface. The exposed stems would be susceptible to blockage from siltation, which would in-turn reduce the oxygen supply to roots and rhizomes thus hampering growth (Gleason & Zieman 1981; Wijte & Gallagher 1991; Thom et al. 1997). Bimonthly clipping at 10 - 15 cm above the mudline from July to November has been reported to significantly reduce aboveground live Spartina spp. biomass (Turner 1987). Continuous cutting of Spartina spp. for use as a biofuel, and continuous grazing has also been reported to cause reductions in aboveground 69

Spartina spp. biomass (Ranwell 1961; Callaghan et al. 1984; Scott et al. 1990). It has however been noted that continuous cutting has no effect on live rhizome biomass or net above ground primary production (Turner 1987).

The time of cutting will influence the effectiveness of cutting treatments (Callaghan et al. 1984). A cut in June followed by a cut in August reduced Spartina spp. stem density by 95 % and was the most effective cutting regime in a study by Norman & Patton (1997). The most successful results should occur when rhizome biomass is lowest i.e. after main spring growth, midsummer and autumn. Spring growth depletes rhizome biomass (Valiela et al. 1976; Gallagher & Plumley 1979; Smith et al. 1979, Gallagher 1983; Schubauer & Hopkinson 1984). Two-thirds of replenishment of roots and rhizomes occurs in the latter half of the summer and one-third occurs in autumn after flowering and seed fall (Gallagher 1983). Cutting during these times should result in the largest reduction in roots and rhizomes. Cutting just prior to seed-set would also eliminate seed production and help prevent further spread.

These results suggest that continuous cutting will slow down S. anglica agg. growth over time but will not prevent growth or spread of S. anglica agg.. The treatment will however allow other species to spread in the area (Ranwell 1961; Scholten & Rozema 1990). Cutting should therefore be used as a S. anglica agg. control treatment rather than an eradication method.

Smothering is a traditional method used for weed control. Smothering excludes light and thus the plant dies. Ranwell (1961) observed a dense layer of drift to smother an area of S. anglica agg. and prevent re-establishment the following year. Experiments in Tasmania have shown that smothering with black plastic sheeting gives 100 % kill of S. anglica (Lane 1995). The smothering treatment in my experiment also had kill rates close to 100%. Smothering was the only method in this study in which dry root weight was lower or the same as the previous year. This suggests that smothering treatments either increases the rate of root and rhizome decomposition or kills roots and rhizomes more effectively than the other treatments. It also suggested slow decomposition of S. anglica agg. roots and rhizomes after successful control. Hemminga et al. (1998) calculated that decomposition of S. anglica roots and rhizomes vary between 2 to 3.9 years. 70

Most of the studies mentioned failed to achieve 100 % kill of Spartina species. Those that eradicated S. anglica agg. (smothering) in this experiment were re-colonised by S. anglica agg. within two years. Ranwell (1961) reported that S. anglica agg. had established to a higher density than prior to smothering within four years. The re- establishment of S. anglica agg. within 1 or 2 years has been reported in other Spartina eradication experiments (Taylor and Burrows 1968a, Corkhill 1984; Garnett et al. 1992). The substrate and environmental conditions that remain after S. anglica agg. eradication attempts are therefore suitable for Spartina recolonization. Re- establishment may occur from S. anglica agg. fragments, living roots and rhizomes, seed or by vegetative spread from nearby uncontrolled areas. Complete eradication will require repeat applications of eradication methods. Recolonization of S. anglica agg. into eradicated areas is likely to occur if any S. anglica agg. remains within the area.

4.2. Effects of S. anglica agg. control treatments on other salt marsh species

Glyphosate is a broad-range herbicide used for the control of annual and perennial grasses, broad-leaved weeds and woody brush. Dalapon is a more selective herbicide used for the control of annual and perennial grasses and reeds. Any plant species in the S. anglica agg. dominated sward that is within the range of plants affected by either herbicide is likely to be killed. Reductions in Salicornia, Suaeda and Puccinellia abundance have been noted when glyphosate is applied to S. anglica swards (Garnett et al. 1992). In my study both dalapon and glyphosate caused reductions in Puccinellia abundance at Ballykelly. Plants that grow in submerged or damp conditions such as Zostera and Enteromorpha are less likely to be affected by herbicide applications (Bulthuis & Shaw 1993; Corkhill 1984; Palmer et al. 1995). The presence of water in these conditions dilutes the herbicide and reduces absorption into leaves.

Smothering should kill all vegetation due to the exclusion of light. Puccinellia maritima abundance declined in smothered plots at Ballykelly. S. anglica agg. and Salicornia have been killed by algal mats and tidal litter due to the effect of smothering (Ranwell 1961; Jefferies et al. 1981). 71

Cutting or grazing of S. anglica agg. swards may promote the growth of other species such as Puccinellia maritima (Ranwell 1961; Beeftink 1985a; Scholten et al. 1987, Scholten & Rozema 1990; Scott et al.1990). P. maritima abundance increased at Ballykelly in the Cut + Glyphosate and Cut plots over the first year. The reduction of S. anglica agg. height caused by cutting allows increased light penetration within the canopy, thus improving the growth of other light dependent species (Goldberg & Miller 1990; Bertness 1991). P. maritima takes advantage of this opportunity to spread in northern lattitudes by sprouting earlier in the season than S. anglica and increasing tiller density (Scholten et al. 1987; Scholten & Rozema 1990). There was no further increase in P. maritima abundance in the Cut and Cut + Glyphosate plots during the second year of this study. This suggests that the opportunity for P. maritima spread was short-lived. S. anglica agg. growth during the growing season would have increased the height of the S. anglica agg. canopy, reducing light penetration and thus hampering further P. maritima spread and growth.

During the second year of my study several species had established on the Dalapon, Cut + Dalapon and Cut + Smother plots. Puccinellia maritima abundance increased and individuals of Aster tripolium and Plantago maritima had established at Ballykelly. Salicornia spp. had colonized in the Ballydrain plots. The removal of S. anglica agg. by these treatments resulted in areas of bare mud. These patches were suitable for the colonisation by species whose presence had been suppressed by the closed S. anglica agg. canopy. Bare patches in other control experiments have been colonized by Atriplex hastata, Puccinellia maritima, Ruppia maritima, Salicornia spp., Suaeda maritima and Zostera spp. (Ranwell & Downing 1960; Taylor & Burrows 1968a; Garnett et al. 1992; Davey et al. 1996; Anderson & Denny 1998; Frid et al. 1999).

Most studies indicate that there is no long-term accumulation of seeds to form a seed- bank in salt-marsh substrates, with seed shed in autumn almost entirely used up by the following summer (Jefferies et al. 1981; Beeflink 1985b). There is a suggestion that the production of dimorphic seeds by Salicornia spp. may constitute to a seasonal seed bank (Jefferies et al. 1983; Philipupilla & Ungar 1984; Ungar 1987a,b; Hartman 1988). It is unclear how persistent this seed bank is and how long the seeds remain 72 viable. Jefferies et al. (1981) noted no germination from turves collected from a marsh containing Salicornia spp. in late summer and autumn, suggesting seeds were exhausted within a year. It is therefore likely that the majority of establishing plants will colonise via plant fragments, vegetative spread or seed inputs. The abundance of adult plants of each species in surrounding areas will influence the availability of seeds or plant fragments for colonising the bare patches (Rand 2000). The majority of seeds of the main colonizing species in this study (Puccinellia maritima and Salicornia spp.) fall within centimetres of the parent plant (Gray & Scott 1977a; Ellison 1987, Koutstaal et al. 1987). The remaining seeds are carried by tide often in tidal litter to more distant areas (Dalby 1963; Gray & Scott 1977a; Ungar 1978; Beeflink 1985b; Ellison 1987). The fine hairs and hooks present on Salicornia spp. seeds may enable the seeds to attach to floating debris giving it an advantage over other marsh species in colonizing new patches.

Aster triploium and Plantago maritima were noted to establish at Ballydrain at very low densities. The increase in Puccinellia maritima was mainly by vegetative spread from surrounding untreated areas, but seedlings were also noted to have established at low levels. At Ballydrain an individual of Puccinellia maritima was recorded and there was low level colonization by Salicornia species. Jefferies et al. (1981) also noted low vegetation recruitment into disturbed areas. Vegetation was removed from areas within salt marsh containing Salicornia species. In the unvegetated plots Salicornia species established at levels less than 5 per m2. This was low compared to seedling recruitment in undisturbed areas. This suggests that seed input into controlled areas from surrounding salt marsh vegetation is low (Hartman 1988; Rand 2000).

The species that colonise will be dependent upon environmental conditions associated with tidal variation and substrates in relation to the preferred regeneration niche of the individual species (Beeflink 1985a; Rozema et al. 1985). The elevation level of the S. anglica agg. marsh prior to control may have affected the species that colonized the bare areas in my study. Salicornia spp. colonizes bare and disturbed patches in low- mid shore marshes as a pioneer spp. (Adam 1981). It can be in vegetational zones associated with S. anglica agg. and Puccinellia maritima (Goodman et al. 1959; Adam 1981; Beeflink 1985a; Gray 1992). Although their niches overlap, Salicornia spp. usually extends lower down the shore and Puccinellia maritima extends higher 73 up the shore (Taylor & Burrows 1968c; Brereton 1971; Armstrong 1988; Roozen & Westhoff 1985; Hill 1987; Scholten & Rozema 1990; Gray 1992; Rowcliffe et al. 1998). There was a lack of Puccinellia maritima colonisation into the study area at Ballydrain even though it was present in surrounding saltmarsh vegetation. It was observed that P. maritima and Aster tripolium had colonised small micro-hummocks within the marsh. This provides unconfirmed evidence that elevation levels in the study plots are too low for successful colonisation by these species. At Ballykelly species associated with mid-zone marshes were noted to colonise. It is however difficult to make inferences about the low Salicornia spp.colonisation at Ballykelly due to the lack of Salicornia spp. in the surrounding area.

The persistence of a colonising species will be affected by its competitive ability against other saltmarsh species (Pielou & Routledge 1976; Gray & Scott 1977b; Dormann et al. 2000). This will be of particular importance at higher elevations due to the increase in the number of species able to tolerate the associated environmental conditions (tidal factors tend to determine species distributions at lower marsh elevations) (Gray 1992). Competition will also be very important in areas were Spartina re-establishment after control is rapid e.g. in my study.

A number of studies have investigated competitive interactions between S. anglica and Puccinellia. Scholten & Rozema (1990, and Scholten et al. 1987) examined competitive ability between the two species in a transitional zone of S. anglica-P. maritima in a Netherlands saltmarsh (in the Netherlands S. anglica zones tend to give way to P. maritima). The study involved removal experiments in two zones differing in height by 4cm. In the lower-zone plots removal of S. anglica resulted in increased Puccinellia growth, while the removal of P. maritima had no effect on S. anglica. In the upper zone removal of P. maritima resulted in increased S. anglica while removal of Spartina had no effect on P. maritima. In lower plots therefore S. anglica normally out-competes P. maritima whereas 4 cm higher P. maritima wins the competition. They also noted from further experiments that, the suppression of S. anglica by P. maritima is maximal at the beginning of the growing season and decreases during the growing season; suppression of Puccinellia by S. anglica agg. increases as the soil becomes more clayish (also noted by Ranwell 1961); suppression of S. anglica by P. maritima increases with marsh level. 74

Huckle et al. (2000) investigated the effect of salinity, sediment type and waterlogging on growth and interactions between S. anglica and P. maritima in pot experiments. They found that P. maritima had competitive dominance above ground over S. anglica when gradients of sediment type and waterlogging were established. P. maritima dominace was highest in loam rather than sand sediments, and in non- immersed rather than immersed treatments. They suggested that competitive advantage of P. maritima was highest in conditions with the least abiotic stress and was lower or non-existent when stress was increased. It was also noted that P. maritima competition in sandy sediments caused an increase in S. anglica below- ground biomass, suggesting that S. anglica was trying to expand vegetatively into areas without competition. These studies indicate that P. maritima will outcompete S. anglica agg. in Northern latitudes when certain environmental conditions prevail. This is most likely to occur in upper marsh elevations with sandy nutrient-rich sediments.

It is likely that shading from the S. anglica agg. canopy will reduce the survival rate of Salicornia spp. (Beeftink 1985a; Ellison 1987). There have been no studies indicating that Salicornia spp. have any form of competitive effect on S. anglica agg.. It has however been noted to have competitive advantages over other saltmarsh plants such as Puccinellia maritima at low marsh levels (Brereton 1971).

The colonisation and persistence of native salt marsh species in controlled areas will make the S. anglica agg. sward more diverse and may facilitate successional change. Aster tripolium, Atriplex hastata, Elymus pycnanthus, Festuca rubra, Juncus maritimus, Leptocarpus simplex, Phragmites australis, Puccinellia maritima, Scirpus maritimus, and Suadea maritima have been noted to invade and dominate areas of S. anglica (Allan 1930; Ranwell 1961; Ranwell 1964; Packham & Liddle 1970; Adam 1981; Hill 1987; Gray et al. 1991; Esselink et al. 1997; Esselink 2000). Alternatively these species and Agropyron pungens, Agrostis stolonifera, Glaux maritima, Halimione portulacoides and Salicornia europea have been noted to invade S. anglica swards to produce mixed saltmarsh communities (Ranwell 1964; Adam 1981; Gray & Pearson 1984, Esselink et al. 1997). 75

In my study there is evidence that control treatments may result in a mixed S. anglica- Salicornia sward at Ballydrain and a mixed saltmarsh community containing S. anglica agg., Puccinellia maritima, Aster tripolium and Plantago maritima at Ballykelly.

4.3. Effects of S. anglica agg. control on benthic fauna

Laboratory bioassay have studied the effects of the herbicides dalapon and glyphosate on marine invertebrates (Thain & Wilson 1974; Gillespie 1989; Garnett 1991; Anon. 1993; Anon 1996a; Kubena et al. 1997). Many of the tests have been carried out using herbicide mixtures with associated surfactants (Gillespie 1989; Kubena et al. 1997), or have been limited to standard test organisms. The effects on invertebrates found in Northern Ireland estuaries has been investigated during a study conducted on behalf of the Environmental Service of the Department of the Environment, Northern Ireland (Anon, 1996a). Acute aqueous toxicity tests were conducted with preparations containing glyphosate (Roundup Pro Biactive) or 2,2-dichoroproprionic acid (active ingredient in dalapon) on the following species– Cerastoderma edule (cockle), Crangon crangon (brown shrimp), Crassostrea gigas (Pacific oyster), Hydrobia ulvae (mud snail), Littorina littorea (periwinkle) and Neanthes diversicolor (ragworm). Acute sediment phase tests using the same herbicide preparations were conducted upon Arenicola marina (lugworm) and Corophium volutator (amphipod). Tests were carried out using preparations of 360 g l-1 glyphosate (Roundup Pro Biactive) with and without the adjuvant Li700 (2 % w/w), and with 50 g l-1 pure 2,2 dichloropropionic acid Na salt with and without the adjuvant Agral (2 % w/w).

The levels of each preparation used were –

• Glyphosate (Roundup Pro Biactive) at aqueous concentrations up to 10 mg l-1 and sediment concentrations up to 10mg kg-1. • Glyphosate (Roundup Pro Biactive) with the adjuvant LI700 at 2 %, at aqueous concentrations up to 10 mg l-1 and sediment concentrations up to 10mg kg-1. • Dalapon (2,2-dichloropropionic acid) at aqueous concentrations up to 100 mg l-1 and sediment concentrations up to 100mg kg-1. 76

• Dalapon (2,2-dichloropropionic acid) with the adjuvant Agral at 2 %, at aqueous concentrations up to 100 mg l-1 and sediment concentrations up to 100mg kg-1.

The tests did not indicate any acute effects of formulations of either active ingredient. The environmental exposure concentrations of glyphosate and Dalapon in the immediate receiving waters (before substantial dilution) were predicted to be 0.1 and 5 mg l-1 respectively (Anon 1996b). For Roundup Pro Biactive, no acute toxicity was observed at 100 times the predicted near-field environmental concentration. For 2,2- dichloropropionic acid, no acute toxicity was observed at 20 times the predicted near- field environmental concentration (tests were not conducted at 100 times the predicted concentration). It was also reported in the study that Roundup Pro Biactive was highly degradable in an aerobic seawater test (58.4 % biodegradtion after 28 days), while 2,2-dichloropropionic acid proved to be only slightly degradable (10.5 % biodegradtion after 28 days).

The conclusion that the lack of acute effects provides confidence that an adequate safety margin exists for both herbicides is similar to other published work. Kubena et al. (1997) suggested that the survival of the Crassosrea gigas larva (oyster) and juvenile Eohaustorius estuarius and Leptocheirus plunulosus (amphipods) would not be threatened by current application rates of Glyphosate and surficants. Gillespie (1989) concluded the species Galaxias maculatus (whitebait), Helice crassa (mud crab), Chione stutchburyi (cockle), Palaemon affinis (estuarine prawn) and Amphibola crenata and Potamopyrygus esturarinus (mud snails) were not sensitive enough to the Dalapon mixture used to result in high mortality from a spray programme.

The severity of effects on individuals however could vary according to environmental conditions e.g. tidal conditions at time of spraying, by spraying techniques used, or by life history stage of the organism.

There has been limited literature examining the insitu effects of glyphosate, dalapon, or other control methods used in this study on estuarine fauna. 77

Garnett et al. (1992) examined the effects of concentrations (100 – 200 l ha-1) of glyphosate and glyphosate with surficant (50% w/w nonyl phenol ethylene oxide condensate and 50% w/w primary alcohol ethylene oxide condensate) on vegetation and invertebrate communities in three British estuaries (Lindisfarne, Dee Estuary and Dyfi Estuary). At Lindisfarne they found significant reductions in the numbers of the gastropod Hydrobia ulvae and juvenile bivalues, Macoma baltica, 1 day after spraying compared to 1 day before treatment. Populations had increased to similar of higher densities 1 year later. The gastropods, Littorina saxatilis and Littorina littoralis, showed no significant changes and Tubifex spp. (annelids) responses were inconsistant. In the Dyfi Estuary, counts of Corophium volutator declined after treatment but had recovered 7 weeks after spraying. In the Dee Estuary the high degree of natural variation in the untreated replicates made species density changes uninterpretable. These results however must be treated with caution as there are some deficiencies in the survey designs and thus population changes could not be attributed with certainty to chemical effects (as opposed to physical effects resulting from sediment movements, migration associated with tidal patterns, etc.). Simenstad et al. (1996) used a more ridgid methodology to examine the effect of Glyphosate (58 % glyphosate glycine) and the surficant X-77 spreader to invertebrate communities at Willapa Bay, Washington State. Glyphosate was applied by helicopter as Rodeo at 4.7 l ha-1 (20 to 40 times lower than Garnett et al. 1992) mixed with the surficant at a rate of 0.9 l ha-1. There was no indication of short-term or long-term effects on the mudflat community (including Corophium spp. and Chironomid larvae). However as for several species in the Garnett et al. (1992) study, there was high variability in species densities between plots, making inferences about the effects of the herbicides difficult to make.

There is an unpublished account cited in Evans (1986) that compares the densities of benthic invertebrates in S. anglica swards, and areas sprayed with dalapon at Lindisfarne, England. In the study 30m wide strips of S. anglica were sprayed in successive summers starting at the seaward edge of the sward with an application rate of 62.5 kg ha-1 (in aqueous solution at the concentration of 55 g l-1). Comparisons were made between the invertebrate numbers in open mud flat, areas sprayed 3-4 years before, areas sprayed 1 - 2 years before and the remaining S. anglica sward. The results showed that Nereis diversicolor and Hydrobia ulvae reached higher densities 78 further downshore. Corophium volutator and Arenciola marina were absent from Spartina swards and open mud flat but were present in areas previously sprayed with dalapon. The highest numbers of Corophium volutator were found in areas that were sprayed one year previously. These findings must be treated with similar caution as the Garnett et al. (1992) study as the methodology did not take into account zonal environmental variation (e.g. sediment structure, tidal patterns and elevation levels between mud flats and swards). A further U. K. field study applied dalapon as a wettable powder at 31.4 kg ha-1 (in a volume of 280 l ha-1) to areas with and without Spartina clumps in November. Invertebrate species densities were compared before and two weeks after spraying. There were no consistent significant effects on invertebrates at any site. The main species present in the sites included Cerastoderma edule, Macoma balthica, Hydrobia ulvae, Corophium volutator, Nereis diversicolor, Retusa sp., Carcinus maenas, Littorina sp. and Arenicola marina (Thain & Wilson 1974, cited in Murgatroyd et al. 1995).

Toxicity bioassays, my study and previous insitu studies provide no conclusive evidence about the mortality rate within invertebrate communities after eradication treatments are applied to S. anglica agg. swards.

In the event of invertebrate mortality after treatment applications re-colonisation of invertebrates to densities similar to untreated areas is likely to be rapid (within one year in my study and previously mentioned insitu studies). The rate of recolonization will depend upon patch size and the invertebrate taxa involved.

Re-colonisation of adult fauna on small (<2000 cm2) defaunated patches has been observed frequently (Thrush & Roper 1988; Frid 1989; Smith & Brumsickle 1989; Chandrasekara & Frid 1996). The adult fauna may colonise from surrounding undisturbed areas (Sousa 1984a,b 1985) or via the water column (Eckman 1983) during tidal cover. Recovery of larger patches is reliant on fauna colonising by movement over and then through the upper surface of the patch rather than through the sediment (Santos & Simon 1980; Smith & Brumsickle 1989). This is mainly achieved by larval stages rather than adult fauna (Santos & Simon 1980). 79

In this study the main invertebrate taxa within the S. anglica agg. swards were Hydrobia ulvae, Corophium volutator and Nereis diversicolor. H. ulvae and C. volutator have life histories or adaptations that aid in the colonisation of suitable habitats. H. ulvae snails can float on the tides in mucous bubble rafts before sinking and settling as the tide ebbs (McClusky 1971). They also lay eggs that hatch to form larvae. These larvae spend a period of time in plankton before settling and metamorphosing into snails (Little 2000). C. volutator adults crawl over the mud surface and swim during ebbing tides (Little 2000). As neap tides progress to springs, juveniles swim at high waters, ensuring dispersal of populations (Lawrie & Raffaelli 1998). N. diversicolor has less successful dispersal mechanisms. N. diversicolor larvae are non-pelagic and remain near or in the substratum, resulting in little interchange between populations (McClusky 1971).

This suggests that the colonisation of the main taxa found in S. anglica agg. infested areas will rapidly recolonise disturbed areas from nearby undisturbed mudflats containing associated fauna. All three species were noted to colonise an inter-tidal area (3-4 ha-1) that was re-created at the Tees estuary (England) by engineering excavations (Evans et al. 1998). The study initiated tidal flow into an area of reclaimed land that received no tidal inundation for 20 years. Mud flat with an abundance of fauna were less than 200 m away. There were no invertebrate animals present in the study area before the tidal flows began in September 1993. C. volutator had colonised parts of the study area at low densities (up to 300 per m2) within a couple of months. H. ulvae colonised parts of the area (up to 50 per m2) by summer 1994. N. diversicolor was not detected until summer 1995 after further excavation work was carried out in July 1994. By 1996 it had high densities (over 2000 per m2) throughout the site. Densities of H. ulvae and C. volutator had also increased by summer 1996, reaching 2000 and 4000 per m2 respectively in parts of the study site.

The most probable long term change in invertebrate communities associated with S. anglica agg. control will be related mainly to changes in environmental / habitat conditions associated with S. anglica agg. removal.

Changes in Spartina spp. density will affect near-bed flow regimes (Gleason et al. 1979; Knutson et al. 1982, Bruno 2000, Bruno & Kennedy 2000). This will have 80 influences upon grain size, sediment organic matter content, microbial abundance and composition, and pore-water chemistry. Changes in these variables will in turn influence infaunal distributions (Hall 1994; Snelgrove & Butman 1994). Benthic faunal community structure will also depend on the sediment features that will develop after S. anglica agg. removal e.g. mud mounds, erosion cliffs, channel edges and pans (all of which suit different infaunal communities) (McGrorty & Goss- Custard 1987).

High S. anglica agg. stem densities may also provide protection from bird predation. Wader birds can exert heavy predation pressure on mud-flat communities (Long & Mason 1983) e.g. sandpipers may take more than 10000 Corophium in 1 day (Little 2000). Wader birds rarely feed in S. anglica agg. marsh (Tubbs & Tubbs 1980; Millard & Evans 1984; Evans 1986; Goss-Custard & Moser 1990). The removal of S. anglica agg. may create areas of open mud which are suitable for exploitation by birds (Haynes 1984; Evans 1986). Increased predation may effect the abundance of the preferential bird food source.

Changes in root and rhizome abundance caused by decomposition after S. anglica agg. death may also have an impact on infaunal community structure. The dense root and rhizome network produced by Spartina spp. invasions increases habitat complexity and heterogeneity (Lana & Guiss 1991; Flynn et al. 1996). Increases in spatial heterogeneity may correspond with increases in species richness and effects species abundance (Pickett & Cadenasso 1995; Begon et al. 1996). Spartina spp. roots and rhizomes also contribute to sediment oxygenation enabling faunal colonolisation (Teal & Wiser 1966; Osenga & Coull 1983; Lana & Guiss 1991). The decomposition of Spartina spp. roots and rhizomes will have impacts on associated fauna e.g. nematodes (Osenga & Coull 1983) but may encourage colonisation from other species whose occurrence or abundance is limited by the presence of the root system (Teal 1958; Capehart & Hackney 1989).

Spartina spp. are a source of food and detritus material in the estuarine ecosystem. S. anglica agg. removal will therefore have impacts upon energy flow and nutrient recycling within the local area. It will in particular affect fauna associated with the Spartina spp. food chain (Teal 1962; Currin et al. 1995; Little 1990, 2000; Riera et al. 81

1999) e.g. Carcinus maenus (Ropes 1968), and polychaetes including N. diversicolor (Jackson et al. 1985; Alkemade et al. 1993).

Further influences on benthic fauna populations will result from re-colonisation by plant species that are associated with different fauna communities e.g. Salicornia spp. (Talley & Levin 1999), green algal mats (McGrorty & Goss-Custard 1987; Little 2000) or Puccinellia (Little 2000).

All of the main taxa in this study (H. ulvae, C. volutator and N. diversicolor) will be affected if there is an increase in bird predation resulting from the removal of S. anglica agg. and the creation of suitable bird feeding habitat. How each invertebrate species will be effected depends on the species composition of waterfowl in the local area and their preferred foods (Evans 1978, 1979, 1986; Evans et al. 1979; Ambrose 1984; Pienkowski et al. 1984; Reise 1985; McGrorty & Goss-Custard 1987; Darborn et al. 1993; Evans et al. 1998). The nature of future colonisation by native vegetation will also have different effects on each invertebrate species, H. ulvae for example is found at high densities within Enteromorpha and Ulva mats (Nicholls et al. 1981). C. volutator is the most likely species to be affected by changes in sediment structure. H. ulvae and N. diversicolor are found in a wide range of sediments (Heip & Herman 1979; Barnes 1978; Mettam 1981; McGrorty & Goss-Custard). C. volutator are found in fine-grade sediments (Meadows 1964; Boyden & Little 1973). McLusky (1967) stated that for C. volutator ‘the nature of the substrate is a most important factor in controlling distribution’. The affect on the main taxa of Spartina spp. as a food source is unclear, due to the many modes of feeding by the species (Jones & Wolff 1981; Little 2000). Changes in flow regime after S. anglica agg. removal (such as increased water velocity) will also affect the main taxa as they suit sheltered habitats (McGrorty & Goss-Custard 1987).

The results of this study indicate that the habitat structure one year after S. anglica agg. control remains suitable for the colonisation of the main invertebrate taxa found in the S. anglica agg. sward benthos. Recolonization of the main taxonomic groups will be rapid (1 – 3 years) if S. anglica agg. control treatments result in invertebrate mortality. Long-term changes of invertebrate assemblages will not be caused directly by treatment methods but will be caused by environmental change resulting from the 82 removal of S. anglica agg.. Studies should be initiated that investigate changes in invertebrate assemblages after S. anglica agg. eradication.

4.4. Effects of S. anglica agg. control on sediments

Near-bed flow regimes and waves within the estuarine system influence sedimentation, distribution of suspended material, sedimentary organic matter content, grain size, movement of biota, pore water chemistry, microbial abundance and composition and erosion processes (Hall 1994; Snelgrove & Butman 1994; Little 2000). Waves for example can cause sediment loss and erosion by undercutting salt marsh cliffs and by scouring sediment surfaces. Oliver (1925) stated that ‘In Poole Harbour, and doubtless elsewhere, Spartina settlements in exposed positions, especially when the level has been raised above the surrounding muds through silting, become liable to wave erosion. In this way the edge of the settlement, on the exposed side, is subject to undercutting and the mud is washed away from the roots’.

Abiotic stresses resulting from these factors affect the lower limits of salt marsh vascular plants and influence faunal community structure. At the seed stage for example unstable substrates can lead to high levels of seed loss (van Eerdt 1985; Groenendijk 1986). Many vascular plants therefore require stabilized substrates for establishment (Bruno & Kennedy 2000).

Spartina spp. growth dissipates wave and tidal energy, and stabilizes the substratum (Gleason et al. 1979; Knutson et al. 1982; Bruno 2000; Bruno & Kennedy 2000). The Spartina spp. root system is effective in reducing lateral cliff erosion (van Eerdt 1985). The death of roots and rhizomes and subsequent root decomposition will reduce the stability of accreted mud cliffs and could cause increased lateral cliff erosion (van Eerdt 1985; Gabet 1998). Knutson et al. (1982) reported that 50 % of the energy associated with waves was dissipated within the first 2.5 m of a S. alterniflora sward. They also reported that there was virtually no wave energy 30 m into the S. alterniflora sward in their study area. This can facilitate the growth of other salt marsh species. Bruno (2000) noted that salt marsh species growing on cobble beaches required buffering by S. alterniflora before establishment was successful. He also noted that seed retention was 95 % lower in un-buffered areas. When S. anglica agg. 83 is killed above ground-shoots will be removed by tidal and wave action (at Lindisfarne it has been reported that aerial vegetation killed by spraying takes 1 – 2 seasons to be removed (Way 1987). Bruno & Kennedy (2000) removed S. alterniflora shoots and noted a 33 % increase in average flow velocity and an 85 % increase in substrate instability. The likelihood of erosion is thus increased after Spartina spp. removal. Erosion has been noted in many estuaries on the south coast of England where marsh degeneration and S. anglica die-back has occurred (Gray & Pearson 1984; Haynes 1984; Tubbs 1984). S. anglica agg. die-back was observed at Langstone Harbour in the mid 50’s (Tubbs 1984). Within 25 years much of the area had eroded to slumping platforms of mud that were at a lower elevation than the former marsh (Haynes 1984). The erosion was noted to begin at channels and the exposed edge of the marsh. Erosion is aided by internal dissection of the marsh forming further erosion channels. This erosion process leaves small marsh fragments that erode slowly until undulating mud flat is left. In other areas at Langstone bay however, no erosion was observed in areas where S. anglica agg. had been dead for more than eight years. Instead Zostera spp. had established in small hollows that surrounded the remaining dead shoot bases (Tubbs 1984). Similar lack of erosion has been observed in many areas after S. anglica agg. control attempts (Way 1987). This suggests that sward exposure and elevation level will greatly influence erosion after S. anglica agg. control. McGrorty & Goss-Custard (1987) suggest that erosion of accreted sediments in low-level areas would take 2-3 years to erode to mudflat suitable for wading birds after the eradication of S. anglica. It was also suggested however that erosion could continue for 20 years in high marsh levels before a new mudflat is produced.

In my study two swards were examined. The sward at Ballydrain was at a lower elevation and is exposed to higher tidal and wave forces that the Ballykelly sward. Erosion is therefore more likely to occur at the Ballydrain sward. The only other species observed to grow at Ballydrain after S. anglica control was Salicornia spp.. These species require stable substrates for establishment (Bruno 2000; Bruno & Kennedy 2000). It is therefore unlikely that the growth of vegetation after S. anglica agg. removal would be able to prevent erosion if water velocity is high enough to disrupt sediments. At Ballykelly the higher elevation makes the substrate suitable for the growth of several salt marsh species. It has been noted that erosion in these areas 84 could take up to 20 years before mud flat is reached. During this time it is likely that other salt marsh plants would be able to colonize the area and restrict erosion.

4.5. Conclusions

In this study the herbicide glyphosate was relatively ineffective, but the herbicide dalapon applied at a rate of 57 kg ha-1 during summer proved to be effective, reducing live S. anglica agg. density by over 95% in the first year. During the same period, smothering produced similar live stem density levels to the dalapon treatments. By the second year, live S. anglica agg. stems densities within the smothered plots were lower than the dalapon treatments, suggesting a more effective kill. A single cut treatment increased S. anglica agg. stem density in the first year and is therefore not suitable for eradication of S. anglica agg. The single cutting treatment increased Puccinellia maritima abundance. It may be a method used to encourage the spread of other salt marsh plants within S. anglica agg. swards. The use on non-herbicidal methods, such as smothering, may become more important for S. anglica agg. control as environmental and health concerns about herbicide use increases (Felsot 1997; Kriwoken & Hedge 2000), or in areas where herbicide bans have been imposed e.g. shellfish designated areas in Northern Ireland.

Dalapon application and smothering (the most effective eradication methods in this study) did not achieve 100% eradication. Re-growth and seedling establishment of S. anglica agg. occurred. The high kill rates achieved in the first year (over 95%), however, suggests that eradication is feasible with repeated application of control treatments.

Dalapon and smothering caused both S. anglica and S. x townsendii stem density to be reduced over 95% in the first year after treatment application. This suggests that eradication treatments will affect both species similarly. S. x townsendii, however, will probably be easier to eradicate due to a lack of re-establishment from seedlings and less vigorous clonal growth. 85

Site specific environmental factors are likely to influence the success of eradication attempts and re-establishment rates. These factors may have caused the changes of live S. anglica agg. stem density observed in the Experimental Control plots at each study site during the three years of this study. This study was conducted in a relatively uniform sward area. It is possible that habitat heterogeneity will affect the success of eradication attempts. S. anglica agg. in creeks for example may be less accessible to a backpack sprayer or may receive unequal herbicide coverage compared to plants in uniform sward areas. It is therefore important to monitor areas after eradication attempts to ensure that all S. anglica agg. plants in the area have received adequate treatment application.

The elevation of the marsh surface under the S. anglica agg. sward will influence marsh development after eradication attempts. Low elevation S. anglica agg. swards and areas exposed to high tidal and wave velocities are more prone to erosion after S. anglica agg. control attempts. If S. anglica agg. eradication is successful the area is likely to erode to former mud flat levels within 3 years (McGrorty & Goss-Custard 1987), or may be colonized with low marsh-level vascular species such as Salicornia species. This would result in mud flat that is suitable for use as feeding grounds for wildfowl and waders.

A longer period of erosion would be required to obtain original mudflat levels at mid to high elevation marshes where several salt marsh species such as Puccinellia maritima, Aster tripolium and Plantago maritima can grow e.g. up to twenty years in old high-level marsh (McGrorty & Goss-Custard 1987). After eradication the area will be prone to colonization by S. anglica agg. and other salt marsh species. Colonization by these species will reduce erosion and promote stabilization of sediment. If the desired outcome is the original mud-flat level repeated control to eradicate these species would probably be necessary. In these mid-high elevation swards it may be possible to use control techniques to encourage succession by, and growth of, other salt marsh species in order to obtain a mixed salt marsh community. This would provide habitats for associated S. anglica agg. animal communities (e.g. rails and ducks use the Ballykelly sward for nesting, personal observations), yet would create a mosaic of habitats suitable for use by other species (e.g. Linz et al. 1996). The plant species colonizing controlled S. anglica agg. areas in this study 86

(Aster tripolium, Plantago maritima, Puccinellia maritima and Salicornia spp.) are suitable food sources for Brent geese (Rowcliffe 1995; Rowcliffe et al. 1998). Current Spartina marshes in Strangford Lough and Lough Foyle may therefore become useful food-sources for geese if edible salt marsh species are encouraged to grow within them. This method would however lead to the long-term loss of feeding ground for waders.

The success of this approach depends on environmental variables at the control site and level of seed input. The lack of seed arriving into the controlled areas is likely to be a major factor in hampering the conversion of S. anglica agg. swards into mixed salt marsh communities (Hartman 1988; Rand 2000). There have been no studies that examine attempts to increase the abundance of other species within S. anglica agg. swards. It may however be possible to overcome the lack of seed input into the area using species transplants or seed additions (Knutson & Woodhouse 1982; Kruczynski 1982; Webb 1982; Woodhouse & Knutson 1982; Anon 1989; Burchett et al. 1998). The damaging effect of increased flow velocity after the control of Spartina species on seedling retention and establishment can be overcome by leaving a strip of Spartina spp. at the seaward edge of the sward (Bruno 2000; Bruno & Kennedy 2000). Another possible obstacle to this method is future climatic change. The niche of native salt marsh plants and Spartina spp. plants are affected by climatic and latitudinal limits. Changes in climate and sea-level resulting from ‘Global Warming’ will therefore result in changes to the potential distribution of these species and will affect competition between species. S. anglica for example would be able to begin spring growth earlier in the year and stop growth later in the year if temperatures increase. This extension in the growing period would reduce the time available for Puccinellia maritima to spread within swards. This alternative control method requires further investigation to evaluate its possible success.

The eradication treatments in this study had no affect on invertebrate numbers one year after application. Eradication treatments are unlikely to cause any long-term changes to invertebrate numbers. Changes of invertebrate numbers are more likely to be caused by habitat / environmental changes resulting from Spartina spp. removal. 87

5. Spartina anglica agg. management strategy

The 95% reduction in live stem density caused by dalapon application or smothering in this study suggests that eradication of S. anglica agg. is feasible if treatment applications are repeated. An effective management strategy is required to successfully eradicate Spartina species (see Figure 5.1).

5.1 General strategy

Moody & Mack (1988) state that “Under a wide variety of circumstances, the area occupied through the growth of satellite foci eventually exceeds the range occupied by the spread of a main focus. After the crossover time, the importance of the main focus to the overall invasion will continue to diminish unless severe control is directed at the satellites. But such a belated control effort is likely to be both impractical and ineffective. Furthermore, any benefit derived by reducing the rate of satellite establishment through curbing the expansion of the main focus is generally not as important in curbing the overall invasion as eliminating satellites. This eventually dominant contribution by the satellites to the overall invasion occurs even if the main focus is initially much larger and grows moderately faster than the satellites.”

It is therefore suggested that initial attention should focus on eradication of S. anglica agg. seedlings, tussocks and clumps and the prevention of S. anglica agg. establishment into new areas. Once achieved yearly monitoring and removal of S. anglica seedlings is required to keep these areas free from S. anglica re-establishment.

The next phase of eradication should focus on eradication of sward areas. Low elevation swards should be eradicated first followed by mid-high elevation swards. Eradication of low elevation swards and subsequent erosion is likely to result in mud flat that is suitable for use as feeding grounds for wildfowl and waders within 3 years (McGrorty & Goss-Custard 1987). High-level marsh may require a period of up to 20 years to erode to low-elevation mud flat after S. anglica agg. eradication (McGrorty & Goss-Custard 1987). During this time the area will be prone to colonization by other 88 Is the invasion of Spartina species perceived as a threat requiring control

No Yes

Do nothing or use control methods Are there legal restraints preventing to convert sward areas into mixed extensive use of herbicides saltmarsh, thus establishing a more diverse ecosystem. No Yes

Are there adequate economic Are there adequate economic resources for complete eradication resources for Spartina control

No Yes No Yes

Use herbicides to Eradicate small Spartina Focus on removal of Remove small Spartina foci eradicate small foci (seedlings – small Spartina foci using using non-herbicidal methods Spartina foci clumps). Then eradicate non-herbicidal methods (use herbicides when possible). (seedlings – clumps). low elevation swards (use herbicides when Eradicate sward areas where Monitor and repeat followed by mid-high possible). Monitor and herbicide use is permitted and eradication treatments elevation swards. repeat eradication prevent expansion of other to ensure no Spartina Monitor throughout and treatments to ensure no swards. re-establishment. If repeat eradication Spartina re-establishment. Monitor and repeat eradication possible prevent sward treatments to prevent If possible prevent sward treatments to ensure no expansion. Spartina re- expansions. Spartina re-establishment. establishment. Convert remaining sward areas into mixed saltmarsh.

Figure 5.1. Potential Spartina anglica agg. management strategies in relation to legal restraints and economic resources. 88 89 salt marsh species. There is therefore a chance that the area will develop into salt marsh rather than mudflat.

Several issues are likely to constrain the effectiveness of this general eradication strategy, such as, limitations of economic resources, practical, and legal restraints.

5.2. Economic and practical restraints

Herbicide application is the most frequently used S. anglica agg. control method due to its practical ease of use and cost effectiveness. Non-herbicidal methods tend to be considerably more expensive and are often impractical to use (see Table 5.1). It is estimated, for example, that it costs £160 to spray a hectare of Spartina spp. with glyphosate (Norman & Patten 1997) whilst smothering the same area with plastic sheeting would cost £7300 (McCluggage & Renwik 2000, pers. comm.). There is also the practical problem of securing the plastic sheeting to the sediment, especially in areas with strong tidal currents. The sheeting also requires routine monitoring to ensure that it is not dislodged and washed away. Non-herbicidal methods such as smothering are therefore not suitable for large scale eradication programmes due to high costs and practicalities. They may however be suitable for use on a small scale, especially in areas where legal restraints prohibit herbicide use. Herbicide use is the most suitable method for eradicating the majority of Spartina spp. infestations. Dalapon proved to be a suitable herbicide in this study, but it is no longer manufactured and is therefore unlikely to have a large contribution to S. anglica agg. eradication in the future. Fluazifop (Fusilade) and Haloxyfop (Gallant) are herbicides that may become suitable replacements for dalapon in Northern Ireland. Fusilade is currently used in Tasmania whilst Gallant is currently used in New Zealand for control of Spartina species (Prichard 1995; Shaw & Gosling 1995; Hedge pers. comm.; McCluggage pers. comm.). These herbicides would require to be registered for use in aquatic systems before they could be used in Northern Ireland. 90

Treatment Estimated Cost Source** Practical Problems Per Hectare* Mow £ 84.00 N & P 1997 Cutting machinery required on soft sediment Mow and glyphosate (weed wipe) £ 116.00 N & P 1997 Machinery required on sediment, heterogeneity of area Glyphosate weed wipe £ 84.00 N & P 1997 Machinery required, heterogeneity of area Glyphosate hand spray £ 158.00 N & P 1997 Limited number of suitable spraying days Glyphosate aerial spray £ 45.00 N & P 1997 Number of spraying days, require plane / helicopter Haloxyfop stem injection £ 33333.00 M & R 2000 Very labour intensive (every 3rd stem injected) Smothering with plastic sheeting £ 7333.00 M & R 2000 Requires securing in tidal area, regular monitoring Cutting and burning regrowth £ 9667.00 M & R 2000 Regrowth requires regular burning Hand pulling 3p – 9p per stem N & P 1997 Labour intensive, only on plants 1 – 3 years old

* Figures converted from American and Australian Dollars.

** Norman & Patton 1997; McCluggage & Renwick 2000

Table 5.1. Estimated cost of Spartina spp. control treatments and associated practical problems. 91

5.3. Legal restraints

Legal restraints often impact upon Spartina spp. eradication attempts. In Northern Ireland, for example, there is currently a ban on the use of herbicides in shellfish designated areas. This is the result of a legal dispute that occurred after an experiment to eradicate S. anglica in 1980 (Kirby 1994). A local oyster farmer settled out of court after claiming that the removal of S. anglica resulted in the liberation of silt, which subsequently smothered and killed his oysters. In these areas only de-minimus herbicide application is permitted. Other areas in Northern Ireland come under European environmental designations e.g. Strangford Lough is a Marine Nature Reserve and other estuarine areas are Areas of Special Scientific Interest (ASSI). In such areas the implementation of most eradication methods including smothering with sheeting require government consent. Within Northern Ireland the consent of landowners must also be gained for eradication to proceed. The landowner can refuse to give consent (especially when chemical control is going to be adopted on their land). Another legal restraint is the required registering / licensing of herbicides for use in the aquatic environment. At present only glyphosate and dalapon are registered for use in N. Ireland estuaries (dalapon on an experimental permit). There is currently no registration of the preferred alternatives fluazifop or haloxyfop. Registration would require a number of toxicity studies on estuarine invertebrates, and S. anglica agg eradication field trials. Mathematical modelling of herbicide dispersal within inter- tidal areas is also required before herbicides can be used in Northern Ireland estuaries. The registering process is likely to be long and expensive especially if governmental environment agencies have to pay for registration fees themselves (herbicide companies may be unwilling to pay the costs needed to register their herbicides for use in aquatic systems).

Herbicides should be used where possible due to cost effectiveness and ease of use. Non-herbicidal methods can normal be used in areas where herbicides are not permitted as consent for their use is usually easier to obtain. This, however, is likely to limit the size of area that can be eradicated as non-herbicidal methods are best suited to small sites. Legal and practical restraints combined with limited economic resources and may mean that completed eradication Spartina spp. within an estuarine 92 system is not feasible. In such cases an alternative control strategy is required rather than the general strategy suggested earlier.

5.4. Alternative Spartina anglica agg. control in areas where legal and economic restraints make eradication improbable

The control strategy should focus attentions on containing the current S. anglica agg. population. To contain the population, small foci such as seedlings, tussocks and clumps should be eradicated. If possible sward expansion should also be prevented. Most resources should be used in eradicating these foci and subsequent monitoring and removal of S. anglica agg. seedlings / fragments to ensure that these areas remain free from S. anglica agg.. Non-herbicidal methods should be applied to these small foci if herbicides are not permitted. Seedlings and tussocks up to 50 cm2 can be successfully dug out, providing that all root and rhizome material is collected. Smothering is likely to be successful at eradicating seedling, tussocks and clumps. Sward areas should be left (except for preventing seaward spread). Sediment will continue to accumulate in the sward areas thus raising elevation levels. Eventually elevation levels required for the growth of several mid-high elevation salt-marsh plants will be reached. These plants will colonise into the sward from surrounding salt marsh areas. The Spartina spp. sward will then, over time, develop into mixed salt marsh communities. This would provide habitat for animal communities that live within Spartina spp. swards, and would create a mosaic of habitats suitable for use by other animal species.

There may be an opportunity to use control techniques in the sward areas that encourage succession by, and growth of, other salt marsh species in order to speed up the development of the mixed salt marsh community e.g. cutting. The lack of seed arriving into the controlled areas is likely to be a major factor in hampering the conversion of S. anglica agg. swards into mixed salt marsh communities (Hartman 1988; Rand 2000). It may however be possible to overcome the lack of seed input into the area using species transplants or seed additions, and by leaving a strip of seaward Spartina spp. to dissipate wave energy, thus leading to increased seed retention (Bruno 2000). 93

5.5. Other management considerations

Sediment release after S. anglica agg. eradication may cause concern to some parties e.g. oyster farmers. To lessen the impact of sediment release a strip of Spartina spp. can be left at the seaward edge of the sward (Bruno 2000; Bruno & Kennedy 2000). The reduced flow velocity caused by the Spartina spp. strip will enable other saltmarsh species to colonize the eradicated area. These species will in-turn stabilize sediments and will reduce sediment erosion when the Spartina spp. strip is eradicated.

If seagrass beds are the desired outcome after S. anglica agg. eradication, seagrass re- vegetation planting techniques can be employed in the eradicated area (Ranwell et al. 1974; Philips 1982).

Successfully eradicated areas and areas that have yet to be colonized by S. anglica agg. can remain free from infestation by the yearly monitoring and removal of any established seedlings. This process could be aided by the production of Geographical Information System (G.I.S.) maps showing the potential spread of Spartina spp. in particular areas (Daehler & Strong 1996). This could be achieved by using S. anglica agg. niche models e.g. the Gray et al. (1985 & 1995) Spartina anglica Niche Model that can predict the potential spread of S. anglica in estuaries from Poole Harbour to Morecambe Bay in south and west Britain. From these maps potential S. anglica agg. areas that are not infested by S. anglica agg. could be identified and a monitoring strategy initiated.

Construction works built on inter-tidal areas, such as piers and tidal barrages, can also cause changes to tidal flows and therefore have the potential to alter the amount of inter-tidal area suitable for S. anglica agg. colonization (Gray et al. 1995). Any future developments on inter-tidal areas should therefore include evaluation of the potential spread of S. anglica agg. in planning considerations.

The niche of Spartina anglica agg. and other saltmarsh plants will be affected by climatic and latitudinal limits. Changes in climate and sea-level resulting from ‘Global Warming’ will therefore result in changes to the potential distribution of Spartina species. Possible changes are worth further research and future monitoring. 94

6. References

Aberle, B. (1990). The biology, control, and eradication of introduced Spartina (cordgrass) worldwide and recommendations for its control in Washington. Report to Washington State Department of Natural Resources, Washington.

Adam, P. (1981). The vegetation of Britsh salt marshes. New Phytologist 88, 143-196.

Alkemade, R., Wielemaker, A. & Hemminga, M. A. (1993). Correlation between nematode abundance and decomposition rate of Spartina anglica leaves. Marine Ecology Progress Series 99, 293-300.

Allan, H. H. (1930). Spartina townsendii. A valuable grass for reclamation of tidal mud-flats. New Zealand Journal of Agriculture 40, 189-196.

Ambrose, W. G. (1984). Influences of predatory polychaetes and epibenthic predators on the structure of a soft-bottom community in a Maine estuary. Journal of Experimental Marine Biology and Ecology; An Annual Review 81, 115-145.

Anderson, G. Q. A. & Denny, M. J. H. (1998). Effects of rotoburying Spartina anglica at Lindisfarne NNR. Report to English Nature.

Andrews, D. (1997). Spartina survey 1997 - Preliminary results. Strangford Lough, National Trust Report.

Anon. (1989). Review of coastal revegetation techniques. Report by Nature Conservancy Council and Environmental Advisory Unit, Liverpool University.

Anon. (1993). Noxious emergent plant. Element E. Environmental effects of Glyphosate and 2,4-D. Report by Ebasco Environmental for Washington State Department of Ecology, Washington. 95

Anon. (1996a). Toxicity studies on herbicides of potential use in Spartina control. Report by Environment and Resource Technology Ltd for the Department of the Environment Northern Ireland, Environment Service Countryside and Wildlife, Belfast.

Anon. (1996b). Spartina control in Strangford Lough. Mathmatical Modelling of herbicide dispersal (Doctor's Bay - Horse Island). Report by Kirk McClure Morton for the Department of the Environment, Northern Ireland, Environmental Service Countryside and Wildlife, Belfast.

Armstrong, W. (1988). Life in the Humber, (A) Salt-marshes. In: A dynamic estuary. Man, nature and the Humber. (ed. N. V. Jones), pp. 46-57. Hull University Press, Hull.

Baker, J. M., Oldham, J. H., Wilson, C. M., Dicks, B., I., L. D. & Levell, D. (1990). Spartina anglica and oil: spill and effluent effects, clean-up and rehabilitation. In: Spartina anglica - a research review. (ed. A. J. Gray and P. E. M. Benham), pp. 52- 62. ITE research publication no. 2., Natural Environment Research Council and HMSO, London.

Barnes, R. S. K. & Greenwood, J. G. (1978). The response of the intertidal gastropod Hydrobia ulvae (Pennant) to sediments of differing particle size. Journal of Experimental and Marine Biology. 31, 43-54.

Beeftink, W. G. (1985a). Vegetation study as a generator for population biological and phsiological research on salt marshes. Vegetatio 62, 469-486.

Beeftink, W. G. (1985b). Population dynamics of annual Salicornia species in the tidal salt marshes of the Ooterschelde, the Netherlands. Vegetatio 61, 127-136.

Begaon, M., Harper, J. L. & Townsend, C. R. (1996). Ecolody: Individuals, Populations and Communities. Blackwell Science, Oxford. 96

Bertness, M. D. (1991). Interspecific interactions among high marsh perennials in a New England saltmarsh. Ecology 72, 125-137.

Bishop, A. (1995). Assessing options for Spartina containment in Tasmania. In: How green is your mudflat? Proceedings of the Australaisian conference on Spartina control. (ed. J. E. Rash, R. C. Williamson and S. J. Taylor), pp. 14-19. Department of Conservation and Natural Resources, Yarram, Victoria.

Bleakley, B. (1979). Spartina - an unwelcome immigrant. Irish Hare 2, 10 - 11.

Boekel, R. (1995). Spartina control in the Barwon river estuary. In: How green is your mudflat? Proceedings of the Australaisian conference on Spartina control (ed. J. E. Rash, R. C. Williamson and S. J. Taylor), pp. 63-65. Department of Conservation and Natural Resources, Yarram, Victoria.

Boyden, C. R. & Little, C. (1973). Faunal distributions in soft sediments of the Severn Estuary. Estuarine and Coastal Marine Science 1, 203-223.

Brereton, A. J. (1971). The structure of the species populations in the initial stages of salt-marsh succession. Journal of Ecology 59, 321-338.

Bruno, J. F. (2000). Facilitation of cobble beach plant communities through habitat modification by Spartina alterniflora. Ecology 81, 1179 - 1192.

Bruno, J. F. & Kennedy, C. W. (2000). Patch-size dependent habitat modification and facilitation on New England cobble beaches by Spartina alterniflora. Oecologia 122, 98 - 108. 97

Bulthuis, D. A. & Shaw, T. C. (1993). Effects of application of glyphosate on the eelgrasses Zostera marina and Zostera japonica in Padilla Bay, Washington. Padilla Bay National Estuarine Research Reserve Technical Report No. 8., Washington State Department of Ecology, Mount Vernon, Washington.

Burchett, M. D., Allen, C., Pulkownik, A. & Macfarlane, G. (1998). Rehabilitation of saline wetland, Olympics 2000 site, Sydney (Australia) - II: Saltmarsh transplantation trials and application. Marine Pollution Bulletin 37, 526 - 534.

Burge, M. N. & Perkins, E. J. (1977). Studies in the distribution and biological impact of the effluent released by Albright and Wilson, LTD. Whitehaven. 1. An investegation of the fungi assocaited with decay of Spartina (cord grass) culms at seaside, Auchencairn Bay, Solway Firth, and the influence of Marchon Effluent. Report by Cumbria Sea Fisheries Committee, Carlisle.

Callaghan, T. V., Scott, R., Lawson, G. J. & Mainwaring, A. M. (1984). An experimental assessment of native and naturalized species of plants as renewable sources of energy in Great Britain. II. Cordgrass Spartina anglica. Report by Institute of Terrestrial Ecology, Natural Environment Research Council, Merlwood, Cumbria.

Cambell, D. E., Odum, H. T. & Knox, G. A. (1990). Organization of a new ecosystem: exotic Spartina alterniflora marsh in New Zealand. In: Spartina workshop record (ed. T. F. Mumfored, P. Peyton, J. R. Sayce and S. Harbell), pp. 24-25. University of Washington and Washington Sea Grant Programme, Seattle, Washington.

Capehart, A. A. & Hackney, C. T. (1989). The potential role of roots and rhizomes in structuring salt-marsh benthic communities. Estuaries 12, 119-122.

Carlton, J. T. (1989). Man's role in changing the face of the ocean: biological invasions and implications for conservation of near-shore environments. Conservation Biology 3, 265 - 273. 98

Carlton, J. T. (1996). Marine bioinvasions: the alteration of marine ecosystems by nonindigenous species. Oceanography 9, 36 - 43.

Carter, R. W. G. (1982). The coast. In: Northern Ireland: Environment and Natural Resources (ed. J. G. Cruickshank and D. N. Wilcock), pp. 119-138. Queens University of Belfast and University of Ulster, Belfast.

Chandrasekara, W. U. & Frid, C. L. J. (1996). The effects of relic fauna on initial patch colonisation in a British saltmarsh. Netherlands Journal of Aquatic Ecology 30, 49-60.

Chapman, V. J. (1959). Studies in saltmarsh ecology. IX. Changes in saltmarsh vegetation at Scolt Head Island. Journal of Ecology 47, 619-639.

Charman, K. (1990). The current status of Spartina anglica in Britain. In: Spartina anglica - a research review. (ed. A. J. Gray and P. E. M. Benham), pp. 11-14. ITE research publication no. 2. Natural Environment Research Council and HMSO, London.

Charter, E. H. & Jones, H. (1951). New forms of Spartina townsendii (Groves). Nature 168, 126.

Charter, E. H. & Jones, H. (1957). Some observations on Spartina townsendii H. and J. Groves in the Dovey Estuary. Journal of Ecology 45, 157-167.

Chinery, M. (1991). A field guide to the insects of Britain and Northern Europe. Harper Collins Publishers, London.

Chung, C. H. (1982). Low marshes, China. In: Creation and Restoration of Coastal Plant Communities (ed. R. R. Lewis III), pp. 131-145. CRC Press, Florida. 99

Chung, C. H. (1990). Twenty-five years of introduced Spartina anglica in China. In: Spartina anglica - a research review. (ed. A. J. Gray and P. E. M. Benham), pp. 72- 76. ITE research publication no. 2, Natural Environment Research Council and HMSO, London.

Chung, C. H. (1993). Thirty years of ecological engineering with Spartina plantations in China. Ecological Engineering 2, 261-289.

Corkhill, P. (1984). Spartina at Lindisfarne NNR and details of recent attempts to control its spread. In: Spartina anglica in Great Britain (ed. P. Doody), pp. 60-63. Focus on nature conservation No. 5, Nature Conservancy Council, Attingham.

Crichton, O. W. (1960). The marsh crab. Estuarine Bulletin 5, 3-10.

Crockett, R. P. (1990). A preliminary review of smooth cordgrass Spartina alterniflora, control with rodeo. In: Spartina workshop record (ed. T. F. Mumford, P. Peyton, J. R. Sayce and S. Harbell), pp. 33. Washington State University and Washington Sea Grant Program, Seattle, Washington.

Crockett, R. P. (1997). A historical perspective of glyphosate use to control Spartina alterniflora in Willapa Bay. In: Second International Spartina Conference Proceedings. (ed. K. Patten), pp. 83-84. Washington State University, Olympia, Washington.

Crothers, K. J. (1990). Control and eradication methods for Spartina in Southland, New Zealand. In: Spartina workshop record (ed. T. F. Mumford, P. Peyton, J. R. Sayce and S. Harbell), pp. 31-32. Washington State University and Washington Sea Grant Program, Seattle, Washington.

Currin, C. A., Newell, S. Y. & Paerl, H. W. (1995). The role of standing dead Spartina alterniflora and benthic microalgae in salt marsh food webs: considerations based on multiple stable isotope analysis. Marine Ecology Progress Series 121, 99- 116. 100

Daehler, C. C. & Strong, D. R. (1995). Impact of high herbivore densities on introduced smooth cordgrass, Spartina alterniflora, invading San Francisco Bay, California. Estuaries 18, 409-417.

Daehler, C. C. & Strong, D. R. (1996). Status, Prediction and prevention of introduced cordgrass Spartina spp. invasions in Pacific estuaries, USA. Biological Conservation 78, 51-58.

Dalby, D. H. (1963.). Seed dispersal in Salicornia pusilla. Nature 199, 197-198.

Darborn, G. R., Amos, C. L., Brylinsky, M., Chritian, H., Drapeau, G., Fass, R. W., Grant, J., Long, B., Paterson, D. M. & Perillo, G. M. E. (1993). An ecological cascade effect: migratory birds affect stability of intertidal sediments. Limnology and Oceanography 38, 225-231.

Davey, P., Venters, M. & Bacon, J. (1996). Spoiling Spartina, a muddy problem solved? Enact 4, 8-9.

Davis, P. & Moss, D. (1984). Spartina and waders - the Dyfi Estuary. In: Spartina anglica in Great Britain. (ed. P. Doody). pp. 37-40. Focus on nature conservation no. 5, Nature Conservancy Council, Attingham.

Doody, J. P. (1990). Spartina - friend or foe? A conservation viewpoint. In: Spartina anglica - a research review. (ed. A. J. Gray and P. E. M. Benham), pp. 77-79. ITE research publication no. 2, Natural Environment Research Council and HMSO, London.

Dormann, C. F., Van Der Wal, R. & Bakker, J. P. (2000). Competition and herbivory during salt marsh succession: the importance of forb growth strategy. Journal of Ecology 88, 571-583. 101

Eckman, J. E. (1983). Hydrodynamic processes affecting benthic recruitment. Limnology and Oceanography 28, 241-257.

Edwards, A. C. & Davis, D. E. (1974). Effects of herbicides on the Spartina salt marsh. In: Ecology of Halophytes (ed. R. J. Reimold and W. H. Queen). pp. 531-563 Academic Press, London.

Ellison, A. M. (1987). Effects of competition, disturbance, and herbivory on Salicornia europaea. Ecology 68, 576-586.

Eno, N. C., Clark, R. A. & Sanderson, W. G. (1997). Non-native marine species in British waters: a review and directory. Joint Nature Conservation Committee, Peterborough.

Esselink, P., Helder, G. J. F., Aerts, B. A. & Gerdes, K. (1997). The impact of grubbing by Greylag Geese (Anser anser) on the vegetation dynamics of a tidal marsh. Aquatic Botany 55, 261-279.

Esselink, P., Zijlstra, W., Dijkema, K. S. & van Diggelen, R. (2000). The effects of decreased management on plant-species distribution patterns in a salt marsh nature reserve in the Wadden Sea. Biological Conservation 93, 61-76.

Evans, P. R. (1978). Reclamation of intertidal land: some effects on Shelduck and water populations in the Tees estuary. Verhandlung ornithologische Gesellschaft Bayern 23, 147-168.

Evans, P. R. (1979). Adaptations shown by foraging shorebirds to cyclic variations in the activity and availability of their invertebrate pray. In: Cyclic Phenomena in Marine Plants and Animals. (ed. E. Naylor and R. G. Hartnoll), pp. 357-366. Pergamon Press, Oxford. 102

Evans, P. R., Herdson, D. M., Knights, P. J. & Pienkowski, M. W. (1979). Short-term effects of reclamation of part of Seal Sands, Teesmouth, on wintering waders and shelduck. Oecologia 41, 183-206.

Evans, P. R. (1986). Use of herbicide 'Dalapon' for control of Spartina encroaching on intertidal mudflats: Beneficial effects on shorebirds. Colonial waterbirds 9, 171-175.

Evans, P. R., Ward, R. M., Bone, M. & Leakey, M. (1998). Creation of temperate- climate intertidal mudflats: Factors affecting colonization and use by benthic invertebrates and their bird predators. Marine Pollution Bulletin 37, 535-545.

Felsot, A. S. (1997). Risk evaluation of chemical technologies used in Spartina control. In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 71-72. Washington State University, Olympia.

Ferris, C., King, R. A. & Gray, A. J. (1997). Molecular evidence for the maternal parentage in the hybrid origin of Spartina anglica (C. E. Hubbard). Molecular Ecology 6, 186-187.

Flynn, M. N., Tararam, A. S. & Wakabara, Y. (1996). Effects of habitat complexity on the structure of macrobenthic association in a Spartina alterniflora marsh. Revista Brasileira Oceanografia 44, 9-21.

Foin, T. C. & Brenchley-Jackson, J. L. (1991). Simulation model evaluation of potential recovery of endangered light-footed clapper rail populations. Biological Conservation 58, 123-148.

Fowler, J., Cohen, L. & Jarvis, P. (1998). Practical Statistics for Field Biology. John Wiley & Sons, Chichester. 103

Frenkel, R. E. (1990). Spartina in Oregon. In: Spartina workshop record (ed. T. F. Mumford, P. Peyton, J. R. Sayce and S. Harbell), pp. 26. Washington State University and Washington Sea Grant Program, Seattle, Washington.

Frid, C. L. J. (1989). The role of recolonisation processes in benthic communities with special reference to the interpretation of predator-induced effects. Journal of Experimental Marine Biology and Ecology 126, 163-171.

Frid, C. L. J., Chandrasekara, W. U. & Davey, P. (1999). The restoration of mud flats invaded by common cord-grass (Spartina anglica, CE Hubbard) using mechanical disturbance and its effects on the macrobenthic fauna. Aquatic conservation: Marine and Freshwater Ecosystems 9, 47-61.

Furphy, J. S. (1970). The distribution and treatment of Spartina anglica (Cord-grass) in Nothern Ireland 1969. Report to Nature Reserves Committee, Department of Environment., Belfast.

Gabet, E. J. (1998). Lateral migration and bank erosion in a salt marsh tidal channel in San Francisco Bay, California. Estuaries 21, 745 - 753.

Gallagher, J. L. & Plumley, F. G. (1979). Underground biomass profiles and productivity in atlantic coastal marshes. American Journal of Botany 66, 156-161.

Gallagher, J. L., Riemold, R. J., Linthurst, R. A. & Pfeiffer, W. J. (1980). Aerial production, mortality, and mineral accumulation-export dynamics in Spartina alterniflora and Juncus roemerianus plant stands in a Georgia slat marsh. Ecology 61, 303-312.

Gallagher, J. L. (1983). Seasonal patterns in recoverable underground reserves in Spartina alterniflora Loisel. American Journal of Botany 70, 212-215. 104

Garnett, R. P. (1991). The control of Spartina on salt marshes and mudflats using Roundup Pro herbicide. An overview of efficacy and ecotoxicology. Monsanto internal report.

Garnett, R. P., Hirons, G., Evans, C. & O'Connor, D. (1992). The control of Spartina (cord-grass) using glyphosate. Aspects of Applied Biology 29, 359-364.

Gillespie, P. A. (1989). Bioassay of acute toxicity of herbicide mixture Dalapon/Weedazol 4L on estuarine fauna. Environmental Toxicology and Chemistry 8, 809-815.

Gleason, M. L., Elmer, D. A., Pien, N. C. & Fisher, J. S. (1979). Effects of stem density upon sediment retention by salt marsh cordgrass, Spartina alterniflora Loisel. Estuaries 2, 271-273.

Gleason, M. L. & Zieman, J. C. (1981). Influence of tidal inundation on interval oxygen supply of Spartina alterniflora and Spartina patens. Estuarine and Coastal Shelf Science 13, 47-57.

Gleason, D. F. & Zimmerman, R. J. (1984). Herbivory potential of postlarval brown shrimp associated with salt marshes. Journal of Experimental Marine Biology and Ecology 84, 235-246.

Goldberg, D. E. & Miller, T. E. (1990). Effects of different resource additions on species diversity in an annual plant community. Ecology 71, 213-225.

Goodman, P. J., Braybrooks, E. M. & Lambert, J. H. (1959). Investigations into 'die- back' in Spartina townsendii agg. I. The present status of Spartina townsendii in Britain. Journal of Ecology 47, 651-677. 105

Goodman, P. J. (1960). Investigations into 'die-back' in Spartina townsendii agg. ii. The morphological structure and composition of the Lymington sward. Journal of Ecology 48, 711-724.

Goodman, P. J. & Williams, W. T. (1961). Investigations into 'die-back' in Spartina townsendii agg. III. Phsiological correlates of 'die-back'. Journal of Ecology 49, 391- 398.

Goodman, P. J., Braybrooks, E. M., Marchant, C. J. & Lambert, J. M. (1969). Biological flora of the British Isles. Spartina. Journal of Ecology 57, 285-313.

Goss-Custard, J. D. & Moser, M. E. (1988). Rates of change in the number of dunlin, Calidris alpina, wintering in British estuaries in relation to spread of Spartina anglica. Journal of Applied Ecology 25, 95-109.

Goss-Custard, J. D. & Moser, M. E. (1990). Changes in the numbers of dunlin (Calidris alpina) in British estuaries in relation to changes in the abundance of Spartina. In: Spartina anglica: A research Review. (ed. A. J. Gray and P. E. M. Benham), pp. 69-71. ITE Research Publication no.2, Natural Environment Research Council and HMSO, London.

Gray, A. J. & Scott, R. (1977a). Puccinellia maritima (Huds.) Parl. biological flora of the British Isles. Jornal of Ecology 65, 699-716.

Gray, A. J. & Scott, R. (1977b). The ecology of Morecambe Bay. VII. The distribution of Puccinellia maritima, Festuca rubra and Agrostis stolonifera in the saltmarshes. Journal of Applied Ecology 14, 229-241.

Gray, A. J. & Pearson, J. M. (1984). Spartina marshes in Poole Harbour, Dorset, with particular reference to Holes Bay. In: Spartina anglica in Great Britain. (ed. P. Doody), pp. 11 - 14. Focus on nature conservation no. 5, Nature Conservancy Council, Attingham. 106

Gray, A. J., Clarke, R. T., Warman, E. A. & Johnson, P. J. (1989). Prediction of marginal vegetation in a post-barrage environment. Institute of Terrestrial Ecology, Wareham.

Gray, A. J., Marshall, D. F. & Raybould, A. F. (1991). A century of evolution in Spartina anglica. Advances in Ecological Research 21, 1-62.

Gray, A. J. (1992). Saltmarsh plant ecology: zonation and succession revisited. In: Saltmarshes. Morphodynamics, conservation and engineering significance. (ed. J. R. L. Allen and K. Pye), pp. 63-80. Cambridge University Press, Cambridge.

Gray, A. J., Warman, E. A., Clarke, R. T. & Johnson, P. J. (1995). The niche of Spartina anglica on a changing coastline. Coastal Zone Topics: Process Ecology and Management 1, 29-34.

Gray, A. J., Raybould, A. F. & Brown, S. L. (1997). The environmental impact of Spartina anglica; past, present and predicted. In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 39-45. Washington State University, Olympia.

Gray, A. J. & Raybould, A. F. (1997). The history and evolution of Spartina anglica in the British Isles. In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 13-16. Washington State University, Olympia.

Groenendijk, A. M. (1986). Establishment of a Spartina anglica population on a tidal mudflat : a field experiment. Journal of Environmental Management 22, 1 - 12.

Hackney, P. (1980). A first record of Spartina x townsendii H. and J. Groves in NE Ireland. Irish Naturalist Journal 20, 46. 107

Hackney, P. (1992). Stewart & Corry's flora of the north-east of Ireland. Vascular plant and charophyte sections. The Institute of Irish Studies and The Queens University of Belfast, Belfast.

Hall, S. J. (1994). Physical disturbance and marine benthic communities: life in unconsolidated sediments. Oceanography and Marine Biology: An Annual Review 32, 179-239.

Harbell, S. & Callaway, J. (1990). Oregon and California. In: Spartina workshop record (ed. T. F. Mumford, P. Peyton, J. R. Sayce and S. Harbell), pp. 40-41. University of Washington and Washington Sea Grant Programme, Seattle, Washington.

Harbord, W. L. (1949). Spartina townsendii: a valuable grass on tidal mud flats. New Zealand Journal of Agriculture 78, 507-508.

Harrington, L. M. B. & Harrington, J. A. (1992). Spartina invasion in the Palix estuary. Papers and Proceedings of Applied Geography 15, 56-60.

Hartman, J. E. (1988). Recolonization of small disturbance patches in a New England salt marsh. American Journal of Botany 75, 1625-1631.

Haynes, F. N. (1984). Spartina in Langstone Harbour, Hampshire. In: Spartina anglica in Great Britain. (ed. P. Doody), pp. 5-10. Focus on nature conservation no. 5, Nature Conservancy Council, Attingham.

Hayward, P. J. & Ryland, J. S. (1998). Handbook of the marine fauna of north-west Europe. Oxford University Press, Oxford.

Headley, A. D. & Sale, F. (1999). The impact of trampling by student groups on saltmarsh vegetation. Field Studies 9, 513-530. 108

Hedge, P. & Kriwoken, L. (1997). Managing Spartina in Victoria and Tasmania, Australia. In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 93-96. Washington State University, Olympia.

Hedge, P. T. (1997). Strategy for the management of rice grass, Spartina anglica, in Tasmania, Australia. Report by Rice Grass Advisory Group, Department of Primary Industries, Water and Environment, Hobart.

Hedge, P. & Kriwoken, L. K. (2000). Evidence for effects of Spartina anglica invasion on benthic macrofauna in Little Swanport estuary, Tasmania. Australian Ecology 25, 150-159.

Heip, C. & Herman, R. (1979). Production of Neris diversicolor O.F. Muller (Polychaete) in a shallow brackish-water pond. Estuarine and Coastal Marine Science 8, 297-305.

Hemminga, M. A., Kok, C. J. & de Munck, W. (1988). Decomposition of Spartina anglica roots and rhizomes in a salt marsh of the Westerschelde Estuary. Marine Ecology - Progress Series 48, 175 - 184.

Hill, M. I. (1987). Vegetation change in communities of Spartina anglica C.E. Hubbard in salt marshes of north-west England. Report by Nature Conservancy Council, Peterborough.

Hill, M. I. (1990). Population differentiation in Spartina in the Dee estuary – common garden and reciprocal transplant experiments. In: Spartina anglica a research review. (ed. A. Gray and P. E. M. Benham), pp. 15-19. ITE research publication no. 2. Natural Environment Research Council and HMSO, London.

Holme, A. (1964). Methods of sampling the benthos. Advances in Marine Biology 2, 171-260. 109

Howard, P. H. (1991). Handbook of environmental fate and exposure data for organic chemicals. Volume III - Pesticides. Lewis Publishers Inc., Michigan.

Hubbard, C. E. (1957). In: report of the British Ecological Society Symposium on Spartina. Journal of Ecology 45, 612-616.

Hubbard, J. C. E. (1965). Spartina marshes in Southern England. VI. Pattern of invasion in Poole Harbour. Journal of Ecology 53, 799-813.

Hubbard, J. C. E. & Ranwell, D. S. (1966). Cropping Spartina marsh for silage. Journal of the British Grassland Society 21, 214-217.

Hubbard, J. C. E. & Stebbings, R. E. (1967). Distribution, dates of origin, and acreage of Spartina townsendii (s.l.) marshes in Great Britain. Proceedings of the Botanical Society of the British Isles 7, 1-7.

Hubbard, J. C. E. (1970). Effects of cutting and seed production in Spartina anglica. Journal of Ecology 58, 329-334.

Hubbard, J. & Whitwell, T. (1991). Ornamental grass tolerance to post-emergence grass herbicides. Hortscience 26, 1507-1509.

Huckle, J. M., Potter, J. A. & Marrs, R. H. (2000). Influence of environmental factors on the growth and interactions between salt marsh plants: effects of salinity, sediment and waterlogging. Journal of Ecology 88, 492-505.

Hurlbert, S. H. (1984). Pseudoreplication and design of ecological field experiments. Ecological Monographs 54, 187 - 211.

Hutson, D. H. & Roberts, T. R. (1987). Herbicides. Volume 6, Progress in pesticide biochemistry and technology. John Wiley and son, Chichester. 110

Jackson, D., Mason, C. F. & Long, S. P. (1985). Macro-invertebrate populations and production on a salt-marsh in east England dominated by Spartina anglica. Oecologia 65, 406-411.

Jackson, D., Harkness, D. D., Mason, C. F. & Long, S. P. (1986). Spartina anglica as a carbon source for salt-marsh invertebrates: a study using 13C values. Oikos 46, 163- 170.

James, R. J. & Fairweather, P. G. (1996). Spatial variation of intertidal macrofauna on a sandy ocean beach in Australia. Estuarine and Coastal Shelf Science 43, 81 - 107.

Jefferies, R. L., Davy, A. J. & Rudmik, T. (1981). Population biology of the salt marsh annual Salicornia europaea agg. Journal of Ecology 69, 17-31.

Jefferies, R. L., Jensen, A. & Bazely, D. (1983). The biology of the annual Salicornia europaea agg. at the limits of its range in Hudson Bay. Canadian Journal of Botany 61, 762-773.

Jones, P. (1980). Leaf mottling of Spartina species causes by a newly recognised virus, spartina mottle virus. Annals of Applied Biology 94, 77 - 81.

Jones, N. V. & Wolff, W. J. (1981). Feeding and survival strategies of estuarine organisms. Plenum Press, London.

Kent, M. & Coker, P. (1995). Vegetation description and analysis: a practical approach. John Wiley & Sons Inc., Chichester.

Kilbride, K. M., Paveglio, F. L. & Grue, C. E. (1995). Control of smooth cordgrass with Rodeo in a southwestern Washington estuary. Wildlife Society Bulletin 23, 520- 524. 111

Kirby, R. (1989). Tidal flat instability and fine-grained sediment transport at Ardmillan Bay, Strangford Lough in relation to Spartina eradication and oyster trestle burial. Report to Department of the Environment for Northern Ireland, Belfast.

Kirby, R. (1994). Sediments 2 - Oysters 0: The case histories of two legal disputes involving fine sediment and oysters. Journal of Coastal Research 10, 466 - 487.

Knutson, P. L., Brochu, R. A., Seelig, W. N. & Inskeep, M. (1982). Wave damping in Spartina alterniflora marshes. Wetlands 2, 87-104.

Knutson, P. L. & Woodhouse, W. W. J. (1982). Pacific coastal marshes. In: Creation and Restoration of Coastal Plant Communities. (ed. R. R. Lewis), pp. 111-130. CRC Press, Inc., Florida.

Koutstaal, B. P., Markusse, M. M. & De Munk, W. (1987). Aspects of seed dispersal by tidal movements. In: Vegetation Between Land and Sea. (ed. A. H. L. Huiskes, C. W. Blom and J. Rozema), pp. 226-232. Dr. W. Junk publishers, Dordrecht, Netherlands.

Kriwoken, L. K. & Hedge, P. (2000). Exotic species and estuaries: managing Spartina anglica in Tasmania, Australia. Ocean and Coastal Management 43, 573-584.

Kruczynski, W. L. (1982). Salt marshes of the northeastern Gulf of Mexico. In: Creation and Restoration of Coastal Plant Communities. (ed. R. R. Lewis), pp. 71-88. CRC Press, Inc., Florida.

Kubena, K. M., Grue, C. E. & DeWitt, T. H. (1997). Rounding up the facts about Rodeo: Development of concentration-response relationships for selected non-target species. In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 73-75. Washington State University, Olympia Washington. 112

Lana, P. C. & Guiss, C. (1991). Influence of Spartina alterniflora on structure and temporal variability of macrobenthic associations in a tidal flat of Paranagua Bay (southeastern Brazil). Marine Ecology Progress Series 73, 231-244.

Lane, D. (1995). The occurance and impact of Spartina in the Rubicon estuary, Tasmania. In: How green is your mudflat? Proceedings of the Australasian conference on Spartina control (ed. J. E. Rash, R. C. Williamson and S. T. Taylor), pp. 20-25. Department of Conservation and Natural Resources, Yarram, Victoria.

Lawrie, S. M. & Raffaelli, D. G. (1998). In situ swimming behaviour of the amphipod Corophium volutator (Pallas). Journal of Experimental Marine Biology and Ecology 224, 237-251.

Lieth, H. & Whittaker, R. H. (1975). Primary productivity of the biosphere. Springer, New York.

Lin, D. F. (1984). Ecological significance of Spartina planting on tideland. Journal of Ecology , 7-9.

Linz, G. M., Blixt, D. C., Bergman, D. L. & Bleier, W. J. (1996). Response of ducks to glyphosate - induced habitat alterations in wetlands. Wetlands 16, 38 - 44.

Little, C. (1990). Animals of the Severn Estuary salt marshes. Proceedings of the Bristol Naturalists' Society 50, 83-94.

Little, C. (2000). The biology of soft shores and estuaries. Oxford University Press, Oxford.

Long, S. P. & Mason, C. F. (1983). Saltmarsh Ecology. Blackie, Glasgow. 113

Long, S. P., Dunn, R., Jackson, D., Othman, S. B. & Yaakub, M. H. (1990). The primary productivity of Spartina anglica on an east amglican estuary. In Spartina anglica - a research review (ed. A. J. Gray and P. E. M. Benham). ITE research publication no. 2, Natural Environment Research Council and HMSO, London.

Madden, B., Jennings, E. & Jeffrey, D. W. (1993). Distribution and ecology of Zostera in Co Dublin. The Irish Naturalist's Journal 24, 303-310.

Major, W. & Grue, C. E. (1997). Control of Spartina alterniflora in Willapa Bay, Washington: Efficacy of mechanical and chemical controls techniques, and their off target impacts on eelgrass (Z.japonica). In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 76-81. Washington State University, Olympia, Washington.

Marchant, C. J. (1963). Corrected chromosome numbers for Spartina x townsendii and its parents. Nature 199, 929.

Marchant, C. J. (1967). Evolution in Spartina (Gramineae). I. The history and morphology of the genus in Britain. Botanical Journal of the Linnean Society 60, 1- 24.

Marchant, C. J. (1968). Evolution in Spartina (Gramineae). II. Chromosomes, basic relationships and the problem of S. x townsendii agg. Botanical Journal of the Linnean Society 60, 381-409.

Marks, T. C. & Mullins, P. H. (1984). Population studies on the Ribble estuary. In: Spartina anglica in Great Britain (ed. P. Doody), pp. 20-25. Focus on nature conservation No. 5, Nature Conservancy Council, Attingham.

Marks, T. C. & Truscott, A. J. (1985). Variation in seed production and germination of Spartina anglica within a zoned saltmarsh. Journal of Ecology 73, 695-705. 114

Mathers, R. G. & Montgomery, W. I. (1997). Quality of food consumed by over wintering pale-bellied brent geese Branta bernicla hrota and wigeon Anas penelope. Biology and Environment: Proceedings of the Royal Irish Academy 97b, 81 - 89.

McCluggage, T. & Renwick, R. (2000). Spartina trials in Northland. In: Biosecurity – A wider perspective. Plant and Animal Biosecurity Conference 19-21 July 2000. New Zealand Biosecuity Institute. Auckland, (in Press).

McClusky, D. S. (1967). Some effects of salinity, moulting and growth of Corophium volutator (Amphipod). Journal of the Marine Biological Association of the United Kingdom 47, 607-617.

McClusky, D. S. (1971). Ecology of Estuaries. Heinemann Educational Books Ltd, London.

McGrorty, S. & Goss-Custard, J. D. (1987). A review of the rehabilitation of areas cleared of Spartina. Report to the Nature Concervancy Council, Institute of Terrestrial Ecology (Natural Environment Research Council).

Meadows, P. S. (1964). Subsrate selection by Corophium species: particle size of substrates. Journal Of Animal Ecology 33, 387-394.

Mettam, C. (1981). Survival strategies in estuarine nereids. In: Feeding and survival strategies of estuarine organisms. (ed. N. V. Jones and W. J. Wolff). pp. 65-77, Plenum Press, London.

Millard, A. V. & Evans, P. R. (1984). Colonisation of mudflats by Spartina anglica: some effects on invertebrate and shorebird populations at Lindisfarne. In: Spartina anglica in Great Britain. (ed. P. Doody), pp. 41-48. Focus on nature conservation No. 5. Nature Conservancy Council, Attingham. 115

Moody, M. E. & Mack R. N. (1988). Controlling the spread of plant invasions: the importance of nascent foci. Journal of Applied Ecology 25, 1009-1021.

Morley, J. V. (1973). Tidal immersion of Spartina marsh at Bridwater Bay, Somerset. Journal of Ecology 61, 383-386.

Murgatroyd, C., Hedgecott, S. & Johnson, I. (1995). Effects of herbicides used in Spartina control (Dalapon and Glyphosate) on saltwater flora and fauna in Strangford Lough - a literature review. Report by the Department of Environment for Northern Ireland, Belfast.

Nairn, R. G. W. (1986). Spartina anglica in Ireland and its potential impact on wildfowl and waders - a review. Irish Birds 3, 215-228.

Nicholls, D. J., Tubbs, C. R. & Haynes, F. N. (1981). The effect of green algal mats on intertidal macrobenthic communities and their predators. Kieler Meeresforschungen 5, 511-520.

Nienhuis, P. H. (1987). Ecology of salt-marsh algae in the Netherlands, A review. In: Vegetation between land and sea (ed. A. H. L. Huiskes, C. W. P. M. Blom and J. Rozema), pp. 66-83. Dr. W. Junk Publishers, Dordrecht.

Norman, M. & Patten, K. (1997). Cost-efficacy of integrated Spartina control practices in Willapa Bay, Washington. In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 89-92. Washington State University, Olympia, Washington.

O'Briain, M. & Healy, B. (1991). Winter distribution of brent geese Branta bernicla hrota in Ireland. Ardea 317 - 326.

Oliver, F. W. (1925). Spartina townsendii: its mode of establishment, economic uses and taxonomic status. Journal of Ecology 13, 74 - 91. 116

Osenga, G. A. & Coull, B. C. (1983). Spartina alterniflora (Loisel) root structure and meiofaunal abundance. Journal of Experimental Marine Biology and Ecology 67, 221-225.

Packham, J. R. & Liddle, M. J. (1970). The Cefni salt marsh, Anglesey. Field Studies 3, 331-356.

Palmer, D., Parry, G., Hart, C., Greenshields, P., Crookes, D., Lockett, M. & Pritchard, G. (1995). Toxicity of Fusilade to seagrass and near-shore marine fauna. In: How green is your mudflat? Proceedings of the Australasian Conference on Spartina Control (ed. J. E. Rash, R. C. Williamson and S. J. Taylor), pp. 30-39. Department of Conservation and Natural Resources, Yarram, Victoria.

Paveglio, F. L., Kilbride, K. M., Grue, C. E., Simenstad, C. A. & Fresh, K. L. (1996). Use of rodeo and X-77 spreader to control smooth cordgrass (Spartina alterniflora) in a southwestern Washington estuary: I. Environmental fate. Environmental Toxicology and Chemistry 15, 961-968.

Payne, K. (1972). A survey of the Spartina feeding insects in Poole Harbour, Dorset. Entomologist's monthly magazine 108, 66-79.

Percival, S. M., Sutherland, W. J. & Evans, P. R. (1998). Intertidal habitat loss and wildfowl numbers: applications of a spatial depletion model. Journal of Applied Ecology 35, 57-63.

Pfeiffer, W. J. & Wiegert, R. G. (1981). Grazers on Spartina and their predators. In: The ecology of a salt marsh. (ed. L. R. Pomerroy and R. G. Weigert), pp. 87-112. Ecological studies 38. Springer - Verlag, New York. 117

Philipupilla, J. & Ungar, I. A. (1984). The effect of seed dimorphism on the germination and survival of Salicorina europaea L. Populatus. American Journal of Botany 71, 542-549.

Phillips, R. C. (1982). Seagrass Meadows. In: Creation and Restoration of Coastal Plant Communities (ed. R. R. Lewis), pp. 173 - 201. CRC Press, INC., Florida.

Pickett, S. T. A. & Cadenasso, M. L. (1985). Landscape ecology: spatial heterogeneity in ecological systems. Science 26, 331-334.

Pielou, E. C. & Routledge, R. D. (1976). Salt marsh vegetation: latitudinal gradients in the zonation patterns. Oecologia 24, 311-321.

Pienkowski, M. W., Ferns, P. N., Davidson, N. C. & Worrall. (1984). Balancing the budget: measuring the energy intake and requirements of shorebirds in the field. In: Coastal Waders and Wildfowl in Winter. (ed. P. R. Evans, J. D. Goss-Custard and W. G. Hale), pp. 29-56. Cambridge University Press, Cambridge.

Praeger, R. L. (1932). Some noteworthy plants found in or reported from Ireland. Proceedings of the Royal Irish Academy 41b, 95 - 124.

Prater, A. J. (1981). Estuary birds of Britain and Ireland. Poyser, Calton.

Pringle, A. W. (1993). Spartina anglica colonisation and physical effects in the Tamar Estuary, Tasmania 1971-91. Papers and Proceedings of the Royal Society of Tasmania 127, 1-10.

Pritchard, G. H. (1995). Herbicide trials on Spartina. In: How green is your mudflat? Proceedings of the Australasian Conference on Spartina Control. (ed. J. E. Rash, R. C. Williamson and S. J. Taylor), pp. 66. Department of Conservation and Natural Resources., Yarram, Victoria. 118

Qin, P., Xie, M., Jiang, Y. & Chung, C. H. (1997). Estimation of the ecological- economic benifits of two Spartina alterniflora plantations in North Jiangsu, China. Ecological Engineering 8, 5-17.

Qin, P., Xie, M. & Jiang, Y. (1998). Spartina green food ecological engineering. Ecological Engineering 11, 147-156.

Rand, T. A. (2000). Seed dispersal, habitat suitability and the distribution of halophytes across a salt marsh tidal gradient. Journal of Ecology 88, 608-621.

Ranwell, D. S. & Downing, B. M. (1959). Brent Goose (Branta bernicla (L.)) winter feeding pattern and Zostera resources at Scolt Head Island. Norfolk. Animal Behaviour 7, 42-56.

Ranwell, D. S. & Downing, B. M. (1960). The use of Dalapon and substituted urea herbicides for control of seed-bearing Spartina (cord grass) in inter-tidal zones of estuarine marsh. Weeds 8, 78-88.

Ranwell, D. S. (1961). Spartina salt marshes in southern England. I. The effects of sheep grazing at the upper limits of Spartina marsh in Bridgwater Bay. Journal of Ecology 49, 325-340.

Ranwell, D. S., Bird, E., Hubbard, J. C. E. & Stebbings, R. E. (1964). Spartina salt marshes in Southern England. V. Tidal submergence and chlorinity in Poole Harbour. Journal of Ecology 52, 627-641.

Ranwell, D. S. (1964). Spartina salt marshes in southern England. III. Rates of establishment, succession and nutrient supply at Bridgwater Bay, Somerset. Journal of Ecology 52, 95-105. 119

Ranwell, D. S. (1967). World resorces of Spartina townsendii (sensu lato) and economic use of Spartina marshland. Journal of Applied Ecology 4, 239-256.

Ranwell, D. S. (1968). Spartina in Northern Ireland; History of introduction and present known distribution of Spartina anglica in Northern Ireland. Report to the Nature Conservancy, Coastal Ecology Research Station and Nature Reserves Committee, Northern Ireland Parliament, Belfast.

Ranwell, D. S. (1972). Ecology of salt marshes and sand dunes. Chapman and Hall, London.

Ranwell, D. S., Wyer, D. W., Boorman, L. A., Pizzey, J. M. & Waters, R. J. (1974). Zostera transplants in Norfolk and Suffolk, Great Britain. Aquaculture 4, 185 - 198.

Raybould, A. F., Gray, A. J., Lawrence, M. J. & Marshall, D. F. (1991). The evolution of Spartina anglica C. E. Hubbard: origin and genetic variability. Biological Journal of the Linnean Society 43, 111-126.

Reimold, R. J., Linthurst, R. A. & Wolf, P. L. (1975). Effects of grazing on salt marsh. Biological Conservation 8, 105-125.

Reise, K. (1985). Tidal Flat Ecology: an Experimental Approach to Species Interactions. Springer-Verlag, Heidelberg.

Riera, P., Stal, L. J., Nieuwenhuize, J., Richard, P., Blanchard, G. & Gentil, F. (1999). Determination of food sources for benthic invertebrates in a salt marsh (Aiguillon Bay, France) by carbon and nitrogen stable isotopes: importance of locally produced sources. Marine Ecology Progress Series 187, 301-307.

Roozen, A. J. M. & Westhoff, V. (1985). A study on long-term salt-marsh succession using permanent plots. Vegetatio 61, 23-32. 120

Ropes, J. W. (1968). The feeding habits of the green crab, Carcinus maenas. U.S. Fish and Wildlife Service Fishery Bulletin 67, 183-203.

Rowcliffe, J. M. (1995). Cyclic winter grazing pattern in brent geese and the regrowth of salt marsh grass. Functional Ecology 9, 931 - 941.

Rowcliffe, J. M., Watkinson, A. R. & Sutherland, W. J. (1998). Aggregative responses of brent geese on salt marsh and their impact on plant community dynamics. Oecologia 114, 417-426.

Rozema, J., Bijwaard, P., Prast, G. & Brockman, R. (1985). Ecophysiological adaptions of coastal halophytes from foredunes and saltmarshes. Vegetatio 61, 23-32.

Santos, S. L. & Simon, J. L. (1980). Marine soft-bottom community establishment following annual defaunation: Larval or adult recruitment. Marine Ecology Progress Series 2, 235-241.

Sayce, K. (1990a). Species displaced by Spartina in the Pacific Northwest. In: Spartina workshop record (ed. T. F. Mumford, P. Peyton, J. R. Sayce and S. Harbell), pp. 26-27. University of Washington and Washington Sea Grant Programme, Seattle, Washington.

Sayce, J. R. (1990b). Spartina in Willapa Bay: a case history. In: Spartina workshop record (ed. T. F. Mumford, P. Peyton, J. R. Sayce and S. Harbell). pp. 27- 29. University of Washington and Washington Sea Grant Program, Seattle, Washington.

Scholten, M., Blaauw, P. A., Stroetenga, M. & Rozema, J. (1987). The impact of competitive interactions on the growth and distribution of plant species in salt marshes. In: Vegetation between land and sea. Structure and processes (ed. A. H. L. Huiskes, C. W. P. M. Blom and J. Rozema), pp. 270-281. Dr. W. Junk Publishers, Lancaster. 121

Scholten, M. & Rozema, J. (1990). The competitive ability of Spartina anglica on Dutch salt marsh. In: Spartina anglica a research review. (ed. A. Gray and P. E. M. Benham), pp. 39-47. ITE research publication no. 2. Natural Environment Research Council and HMSO, London.

Schubauer, J. P. & Hopkinson, C. S. (1984). Above and belowground emergent macrophyte production and turnover in a coastal marsh ecosystem, Georgia. Limnology and Oceanography 29, 1052-1065.

Scott, R., Callaghan, T. V. & Lawson, G. J. (1990). Spartina as a biofuel. In Spartina anglica - a research review. (ed. A. J. Gray and P. E. M. Benham), pp. 48-51. ITE research publication no. 2. Natural Environment Research Council and HMSO, London.

Shaw, W. B. & Gosling, D. S. (1995). Spartina control in New Zealand - An Overview. In: How green is your mudflat? Proceedings of the Australasian Conference on Spartina Control. (ed. J. E. Rash, R. S. Williamson and S. J. Taylor), pp. 43-60. Department of Conservation and Natural Resources., Yarram, Victoria.

Shaw, W. B. & Gosling, D. S. (1997). Spartina ecology, control and eradication - recent New Zealand experience. In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 32-38. Washington State University, Olympia.

Simenstad, C. A., Cordell, J. R., Tear, L., Weitkamp, L. A., Paveglio, F. L., Kilbride, K. M., Fresh, K. L. & Grue, C. E. (1996). Use of Rodeo and X-77 spreader to control cordgrass (Spartina alterniflora) in a southwestern Washington estuary: 2. Effects on benthic microflora and invertebrates. Environmental Toxicology and Chemistry 15, 969-978. 122

Simpson, J. (1995). Wading birds of Anderson inlet and the work of the Victorian wader study group. In How green is your mudflat? Proceedings of the Australasian conference on Spartina control (ed. J. E. Rash, R. C. Williamson and S. J. Taylor), pp. 67. Department of Conservation and Natural Resources, Yarram, Victoria.

Smith, K. K., Good, R. E. & Good, N. F. (1979). Production dynamics for above and belowground components of a New Jersey Spartina alterniflora tidal marsh. Estuarine and Coastal Marine Science 9, 189-201.

Smith, C. R. & Brumsicle, J. B. (1989). The effects of patch size and substrate isolation on colonization models and rates in an intertidal sediment. Limnology and Oceanography 34, 1263-1277.

Smith, K. G. V. (1989). An introduction to the immature stages of British flies. Diptera larvae, with notes on eggs, puparia and pupae. Royal Entomological Society of London and the British Museum, London.

Smith, D. L. & Proffitt, C. E. (1999). The effects of crude oil and remediation burning on three clones of smooth cordgrass (Spartina alterniflora Loisel.). Estuaries 22, 616- 623.

Snelgrove, P. V. R. & Butman, C. A. (1994). Animal sediment relationships revisited: cause versus effect. Oceanography and Marine Biology: An Annual Review 32, 111- 177.

Sokal, R. R. & Rohlf, F. J. (1998). Biometry: the principles and practice of statistics in biological research., Third edition. W.H. Freeman and Company, New York.

Sousa, W. P. (1984a). The role of disturbance in natural communities. Annual Review of Ecological Systems 15, 353-391.

Sousa, W. P. (1984b). Intertidal mosaics: patch size, propagule availability, and spatially variable patterns of succession. Ecology 65, 1918-1935. 123

Sousa, W. P. (1985). Disturbance and patch dynamics on rocky intertidal shores. In: The ecology of natural disturbance and patch dynamics. (ed. S. T. A. Pickett and P. S. White), pp. 101-124. Academic Press, London.

Stapf, O. (1908). Spartina townendii. The Gardeners' Chronicle , 33-35.

Stapf, O. (1914). Distribution and ecology of Spartina townsendii. Journal of Ecology 2, 192-193.

Strong, D. R. (1990). Insect herbivores that feed on Spartina alterniflora. In: Spartina workshop record (ed. T. F. Mumford, P. Peyton, J. R. Sayce and S. Harbell), pp. 20. University of Washington and Washington Sea Grant Program, Seattle, Washington.

Talley, T. S. & Levin, L. A. (1999). Macrofaunal succession and community structure in Salicornia marshes of southern California. Estuarine, Coastal and Shelf Science 49, 713-731.

Taylor, M. C. & Burrows, E. M. (1968a). Chemical control of fertile Spartina townsendii (S.L.) on the Cheshire shore of the Dee Estuary. Weed Research 8, 170- 184.

Taylor, M. C. & Burrows, E. M. (1968b). Chemical control of fertile Spartina townsendii (S.L.) on the cheshire shore of the Dee Estuary. II. Response of Spartina to treatment with paraquat. Weed Research 8, 185-195.

Taylor, M. C. & Burrows, E. M. (1968c). Studies on the biology of Spartina in the Dee estuary, Cheshire. Journal of Ecology 56, 795-809.

Teal, J. M. (1958). Distribution of fiddler crabs in Georgia salt marshes. Ecology 39, 185-193. 124

Teal, J. M. (1962). Energy flow in the salt marsh ecosystem of Georgia. Ecology 43, 614-624.

Teal, J. M. & Wieser, W. (1966). The distribution and ecology of nematodes in a Georgia saltmarsh. Limnology and Oceanography 11, 217-222.

Thain, J. E. & Wilson, K. W. (1974). The toxicity of some herbicides used for Spartina control. ICES Fisheries Improvement Committee paper C. M.

Thom, R., Cordell, J., Simenstad, C., Luiting, V. & Borde, A. B. (1997). Autoecology of Spartina in Willapa Bay, Washington: benthic metabolism and below-ground growth. In: Second International Spartina Conference Proceedings (ed. K. Patten), pp. 18-20. Washington State University, Olympia, Washington.

Thompson, J. D. (1990). Morphological variation among natural populations of Spartina anglica. In: Spartina anglica a research review. (ed. A. Gray and P. E. M. Benham), pp. 26-33. ITE research publication no. 2. Natural Environment Research Council and HMSO, London.

Thompson, J. D. (1991). The biology of an invasive plant. What makes Spartina anglica so successful? Bioscience 41, 393-401.

Thompson, J. D., McNeilly, T. & Gray, A. J. (1991). Population variation in Spartina anglica C. E. Hubbard. III. Response to substrate variation in a greenhouse experiment. New Phytologist 117, 141-152.

Thrush, S. F. & Roper, D. S. (1988). Merits of macrofaunal colonisation of intertidal mudflats for pollution monitoring; preliminary study. Journal of Experimental Marine Biology and Ecology 116, 219-233. 125

Truscott, A. (1984). Control of Spartina anglica on the amenity beaches of Southport. In: Spartina anglica in Great Britain. (ed. P. Doody), pp. 64-69. Focus on nature conservation No. 5. Nature Conservancy Council, Attingham.

Tubbs, C. R. & Tubbs, J. M. (1980). Wader and Shelduck feeding distribution in Langstone Harbour, Hampshire. Bird Study 27, 239-248.

Tubbs, C. (1984). Spartina on the south coast: an introduction. In: Spartina anglica in Great Britain. (ed. P. Doody), pp. 3-4. Focus on nature conservation no. 5. Nature Conservancy Council, Attingham.

Turner, M. G. (1987). Effects of grazing by feral horses, clipping, trampling, and burning on a Georgia salt-marsh. Estuaries 10, 54-60.

Ungar, I. A. (1978). Halophyte seed germination. Botanical Review 44, 233-305.

Ungar, I. A. (1987a). Population ecology of halophytes seeds. Botanical Review 53, 301-334.

Ungar, I. A. (1987b). Population characteristics, growth, and survival of the halophyte Salicornia europea. Ecology 68, 569-575.

Valiela, I., Teal, J. M. & Persson, N. Y. (1976). Production and dynamics of experimentally enriched salt marsh vegetation: belowground biomass. Limnology and Oceanography 21, 245-252. van Eerdt, M. M. (1985). The influence of vegetation on erosion and accretion in salt marshes of the Oosterschelde, the Netherlands. Vegetatio 62, 367-373.

Way, L. (1987). A review of Spartina control methods. Report by the Nature Conservancy Council. 126

Way, L. (1990). Spartina anglica control programs in the United Kingdom. In: Spartina workshop record (ed. T. F. Mumford, P. Peyton, J. R. Sayce and S. Harbell), pp. 30-31. Washington State University and Washington Sea Grant Program, Seattle, Washington.

Webb, J. W. J. (1982). Salt marshes of the western Gulf of Mexico. In: Creation and Restoration of Coastal Plant Communities. (ed. R. R. Lewis), pp. 89-110. CRC Press, Inc., Florida.

Webb, E. C., Mendelssohn, I. A. & Wilsey, B. J. (1995). Causes for vegetation dieback in a Louisiana salt marsh: A bioassay approach. Aquatic Botany 51, 281-289.

Wijte, A. H. B. M. & Gallagher, J. L. (1991). The importance of dead and young live shoots of Spartina alterniflora (Poaceae) in a mid-latitude salt marsh for overwintering and recoverability of underground reserves. Botanical Gazette 152, 509-513.

Woodhouse, W. W. J. & Knutson, P. L. (1982). Atlantic coastal marshes. In: Creation and Restoration of Coastal Plant Communities. (ed. R. R. Lewis), pp. 45-70. CRC Press, Inc., Florida.

Wu, M.-Y., Hacker, S., Ayres, D. & Strong, D. R. (1999). Potential of Prokelisis spp. as biological control agents of English Cordgrass, Spartina anglica. Biological Control 16, 267-273.

Zedler, J. B. (1993). Canopy architecture of natural and planned cordgrass marshes: selecting habitat evaluation criteria. Ecological Applications 3, 123-138. 127

7. Appendix – Raw data

Appendix: Study plots raw data - S. anglica agg. stem density per 0.25 m2

BALLYKELLY BALLYDRAIN TREATMENT 1998 1999 2000 1998 1999 2000 Dalapon 60.4 1 4.6 128.4 5.4 10.4 Dalapon 60 1.4 48.2 103.2 9.6 5.2 Dalapon 69.8 1 15.6 95.4 .4 .4 Dalapon 70 5.6 43.6 36.2 0 0 Dalapon 46.8 3.2 21.6 145.4 7.8 23.2 Dalapon 51.8 .8 18.8 42.2 0 3.4 Cut + Glyphosate 78.2 78.6 59.2 121.2 74.4 116.4 Cut + Glyphosate 53 108.4 83.6 72.8 80.2 76.4 Cut + Glyphosate 33.4 70 61.6 21.4 15 17.4 Cut + Glyphosate 62.8 111.6 39.4 81.2 74.8 65.2 Cut + Glyphosate 55.6 55.2 56.6 34.8 30 42.6 Cut + Glyphosate 40.4 90 57 78 30.8 44.8 Experimental Control 50.6 48 92.6 144 86.2 61.2 Experimental Control 65 46.2 93.4 116 62.4 69.2 Experimental Control 57.2 51.6 54.2 76.6 44.4 54.2 Experimental Control 40.8 28.6 72 93.4 33 52.4 Experimental Control 63.6 52 86.4 124.4 38.2 27.6 Experimental Control 63.2 61.8 95.8 49.4 22 36.4 Glyphosate 61.2 50.8 98.4 109.6 49.4 48.6 Glyphosate 57.6 52.8 67.8 82.4 38.2 50 Glyphosate 67.2 59 93.6 54.8 18.2 23.2 Glyphosate 46 62.2 103.2 136.8 66.8 84.4 Glyphosate 44.2 39.8 96.6 30 33 78 Glyphosate 59.2 21 107.4 96.4 37.8 72.8 Cut + Smother 53.2 .2 19.8 27.6 0 0 Cut + Smother 63.8 0 1 72 0 0 Cut + Smother 62.6 2.2 0 43.4 .2 0 Cut + Smother 50.2 .2 2.8 155.6 .4 0 Cut + Smother 53.8 .8 4.6 47.8 0 .2 Cut + Smother 49.8 0 4.4 101.6 0 0 Cut 38.4 77.6 64.6 51.6 65.8 83.4 Cut 56.2 51.8 63.6 107 116.4 105 Cut 68.6 86.2 79.6 141.4 34.2 79.8 Cut 58.2 102.2 59.4 65.4 114 44.4 Cut 69.4 86 51.6 53.6 111.4 129.8 Cut 55.2 116.5 33.2 49.8 17.4 35.8 Cut + Dalapon 49.8 2.8 29.6 68.2 0 1.6 Cut + Dalapon 56.2 2 34.4 80.4 0 0 Cut + Dalapon 55.4 2.8 26.8 51.4 0 .6 Cut + Dalapon 66.6 .6 21 117.4 0 4.2 Cut + Dalapon 92 .6 31 136 0 .2 Cut + Dalapon 76.4 3.8 23.6 82 2.6 4.4 128

Appendix: Study plots raw data - S. anglica agg. stem height (cm)

BALLYKELLY BALLYDRAIN TREATMENT 1998 1999 1998 1999 Dalapon 33.01 11.24 26.26 16.33 Dalapon 30.23 13.70 26.20 12.73 Dalapon 33.28 10.70 29.13 26.25 Dalapon 31.57 12.98 21.57 13.70 Dalapon 34.20 11.66 21.67 19.27 Dalapon 29.81 8.40 20.19 16.50 Cut + Glyphosate 37.30 23.57 24.18 17.89 Cut + Glyphosate 33.22 18.89 24.03 13.99 Cut + Glyphosate 32.67 24.46 20.46 16.52 Cut + Glyphosate 31.18 22.14 27.89 19.29 Cut + Glyphosate 30.18 24.34 20.92 13.39 Cut + Glyphosate 32.91 29.08 19.55 12.67 Experimental Control 30.03 43.08 27.77 29.67 Experimental Control 32.84 45.52 25.96 28.65 Experimental Control 39.09 48.20 32.38 20.54 Experimental Control 34.08 45.87 24.74 23.64 Experimental Control 30.24 43.72 22.43 21.79 Experimental Control 31.34 44.66 21.45 21.74 Glyphosate 34.55 24.18 25.37 21.17 Glyphosate 37.93 17.79 25.97 20.84 Glyphosate 35.43 25.97 21.68 20.05 Glyphosate 31.13 29.17 21.20 22.05 Glyphosate 35.63 26.83 21.13 15.79 Glyphosate 33.78 33.61 19.01 17.64 Cut + Smother 37.75 10.90 26.56 - Cut + Smother 37.25 10.20 25.57 10.10 Cut + Smother 36.97 - 23.71 5.95 Cut + Smother 34.02 9.72 25.98 9.34 Cut + Smother 31.86 14.50 25.59 - Cut + Smother 29.61 - 19.98 - Cut 35.02 34.16 27.87 18.84 Cut 28.97 36.38 27.40 18.47 Cut 30.81 26.03 29.12 17.06 Cut 31.64 38.72 26.11 14.60 Cut 38.17 36.95 19.06 13.64 Cut 30.82 41.42 18.61 12.66 Cut + Dalapon 31.02 14.28 28.31 - Cut + Dalapon 31.15 14.37 26.80 9.53 Cut + Dalapon 31.38 13.60 20.97 9.87 Cut + Dalapon 31.59 11.13 21.79 5.77 Cut + Dalapon 32.61 14.42 20.57 - Cut + Dalapon 36.78 12.33 21.08 8.96 129

Appendix: Study plots raw data - Dry root weight (g) per core

BALLYKELLY BALLYDRAIN TREATMENT 1998 1999 1998 1999 Dalapon 13.9 17.8 12.0 16.9 Dalapon 9.3 11.8 10.8 15.2 Dalapon 12.8 12.7 10.7 13.6 Dalapon 7.4 11.7 12.9 10.7 Dalapon 12.8 13.9 11.6 13.4 Dalapon 14.6 13.7 9.5 10.5 Cut + Glyphosate 14.0 15.1 13.6 14.1 Cut + Glyphosate 16.7 17.2 9.4 15.1 Cut + Glyphosate 16.8 14.7 10.3 11.6 Cut + Glyphosate 15.4 14.7 9.2 8.8 Cut + Glyphosate 14.0 11.9 10.6 11.5 Cut + Glyphosate 9.8 13.7 14.0 13.1 Experimental Control 14.9 18.9 9.2 14.1 Experimental Control 12.5 13.6 11.2 18.0 Experimental Control 15.9 17.3 9.2 12.2 Experimental Control 13.4 12.2 9.5 14.4 Experimental Control 11.5 17.4 10.8 13.0 Experimental Control 14.9 15.1 13.1 14.9 Glyphosate 11.3 13.2 10.7 12.7 Glyphosate 12.3 15.9 8.5 18.1 Glyphosate 10.1 9.9 9.7 11.5 Glyphosate 9.7 14.3 9.3 11.6 Glyphosate 10.9 14.6 10.4 16.8 Glyphosate 12.5 15.8 13.0 13.2 Cut + Smother 8.4 8.0 9.5 15.1 Cut + Smother 10.0 10.4 10.9 9.3 Cut + Smother 13.0 11.9 17.4 12.2 Cut + Smother 14.2 9.2 6.0 12.1 Cut + Smother 8.5 13.4 15.5 9.5 Cut + Smother 12.0 8.0 12.6 13.7 Cut 9.8 8.3 7.0 13.4 Cut 9.2 11.5 16.3 11.5 Cut 16.2 13.6 14.6 10.7 Cut 12.4 15.8 11.2 11.2 Cut 12.2 17.5 11.4 14.3 Cut 15.5 16.2 9.0 11.7 Cut + Dalapon 4.2 6.2 10.2 11.7 Cut + Dalapon 8.4 13.6 9.8 6.3 Cut + Dalapon 13.2 10.2 14.3 9.9 Cut + Dalapon 13.0 13.7 10.9 10.3 Cut + Dalapon 10.6 16.2 12.6 17.0 Cut + Dalapon 13.9 16.2 7.9 13.3 130

Appendix: Study plots raw data - Puccinellia maritima abundance (Domin scale)

BALLYKELLY BALLYDRAIN TREATMENT 1998 1999 2000 1998 1998 2000 Dalapon 3 0 1 1 0 0 Dalapon 1 4 0 0 7 0 Dalapon 1 1 5 0 0 0 Dalapon 3 0 6 0 0 0 Dalapon 5 0 5 0 0 0 Dalapon 4 0 1 0 0 0 Cut + Glyphosate 8 7 8 0 0 0 Cut + Glyphosate 3 5 4 0 0 0 Cut + Glyphosate 0 8 4 0 0 0 Cut + Glyphosate 9 4 10 0 0 0 Cut + Glyphosate 3 8 9 0 0 0 Cut + Glyphosate 0 3 5 0 0 0 Experimental Control 1 4 0 0 0 0 Experimental Control 1 1 1 0 0 0 Experimental Control 9 0 8 0 0 0 Experimental Control 1 8 0 0 0 0 Experimental Control 4 0 5 0 0 0 Experimental Control 2 1 1 0 0 0 Glyphosate 3 1 3 0 0 0 Glyphosate 8 3 7 0 0 0 Glyphosate 6 1 4 0 0 0 Glyphosate 4 3 4 0 0 1 Glyphosate 4 1 3 0 0 0 Glyphosate 6 3 1 0 0 0 Cut + Smother 7 1 8 0 0 0 Cut + Smother 7 1 5 0 0 0 Cut + Smother 4 0 3 0 0 0 Cut + Smother 4 5 1 0 0 0 Cut + Smother 0 1 0 0 0 0 Cut + Smother 5 1 5 0 0 0 Cut 0 1 1 0 0 0 Cut 2 0 1 0 0 0 Cut 4 8 3 0 0 0 Cut 3 5 7 0 0 0 Cut 4 6 6 0 0 0 Cut 10 10 10 0 0 0 Cut + Dalapon 0 1 0 0 0 0 Cut + Dalapon 0 4 0 0 0 0 Cut + Dalapon 0 0 0 0 0 0 Cut + Dalapon 4 0 3 0 0 0 Cut + Dalapon 2 1 5 0 0 0 Cut + Dalapon 5 0 7 0 0 0 131

Appendix: Study plots raw data - Other plant species abundance (Domin scale) Ballykelly

Aster tripolium Plantago maritima C. rubrum TREATMENT 1998 1999 2000 1998 1999 2000 2000 Dalapon 0 0 0 1 0 1 0 Dalapon 0 0 0 0 0 0 0 Dalapon 0 0 0 0 0 0 0 Dalapon 0 0 0 0 0 0 0 Dalapon 0 0 1 0 0 0 0 Dalapon 0 0 0 0 0 0 0 Cut + Glyphosate 0 0 0 0 0 0 0 Cut + Glyphosate 1 0 0 0 0 1 0 Cut + Glyphosate 0 0 0 0 0 0 0 Cut + Glyphosate 0 0 0 0 0 0 0 Cut + Glyphosate 0 0 0 0 0 0 0 Cut + Glyphosate 0 0 0 0 0 0 0 Experimental Control 0 0 0 0 0 0 0 Experimental Control 0 0 0 1 1 0 0 Experimental Control 0 0 0 0 0 0 0 Experimental Control 0 0 0 0 0 0 0 Experimental Control 0 0 0 0 0 0 0 Experimental Control 0 0 0 0 0 0 0 Glyphosate 0 0 0 0 1 0 0 Glyphosate 0 0 1 0 0 0 0 Glyphosate 0 0 0 0 0 0 0 Glyphosate 0 0 0 0 0 0 0 Glyphosate 0 0 0 0 0 0 0 Glyphosate 0 0 0 0 0 0 0 Cut + Smother 0 0 0 0 0 0 0 Cut + Smother 0 0 0 0 0 0 0 Cut + Smother 0 0 0 0 0 0 0 Cut + Smother 0 0 0 0 0 0 0 Cut + Smother 0 0 0 0 0 0 0 Cut + Smother 0 0 0 0 0 0 0 Cut 0 0 0 0 0 1 0 Cut 0 0 0 0 0 0 0 Cut 0 0 0 0 0 0 0 Cut 0 0 0 0 0 0 0 Cut 0 0 0 0 0 0 0 Cut 0 0 1 0 0 0 1 Cut + Dalapon 0 0 0 0 0 1 0 Cut + Dalapon 0 0 0 0 0 0 0 Cut + Dalapon 0 0 0 0 0 0 0 Cut + Dalapon 0 1 0 0 1 1 0 Cut + Dalapon 0 1 1 0 1 0 0 Cut + Dalapon 0 0 1 0 0 3 0 132

Appendix: Study plots raw data - Other plant species abundance (Domin scale) Ballydrain

Salicornia spp. TREATMENT 1998 1999 2000 Dalapon 0 0 1 Dalapon 0 0 1 Dalapon 0 0 0 Dalapon 0 0 0 Dalapon 0 0 0 Dalapon 0 0 0 Cut + Glyphosate 0 0 0 Cut + Glyphosate 0 0 0 Cut + Glyphosate 0 0 0 Cut + Glyphosate 0 0 0 Cut + Glyphosate 0 0 0 Cut + Glyphosate 0 0 0 Experimental Control 0 0 0 Experimental Control 0 0 0 Experimental Control 0 0 0 Experimental Control 0 0 0 Experimental Control 0 0 0 Experimental Control 0 0 0 Glyphosate 0 0 0 Glyphosate 0 0 0 Glyphosate 0 0 0 Glyphosate 0 0 0 Glyphosate 0 0 0 Glyphosate 0 0 0 Cut + Smother 0 0 1 Cut + Smother 0 0 3 Cut + Smother 0 0 0 Cut + Smother 0 0 0 Cut + Smother 0 0 0 Cut + Smother 0 0 0 Cut 0 0 0 Cut 0 0 0 Cut 0 0 0 Cut 0 0 0 Cut 0 0 0 Cut 0 0 0 Cut + Dalapon 0 0 1 Cut + Dalapon 0 0 0 Cut + Dalapon 0 0 0 Cut + Dalapon 0 0 0 Cut + Dalapon 0 0 0 Cut + Dalapon 0 0 0 133

Appendix: Study plots raw data - Invertebrate numbers per core, Ballykelly 1998

TREATMENT A. tornatilis Aranea Brachycera C. edule Dalapon0000 Dalapon0000 Dalapon0000 Dalapon0000 Dalapon0000 Dalapon0000 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 .3 .3 0 Cut + Glyphosate .3 0 .3 0 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 0 .3 0 Experimental Control 0 .3 0 0 Experimental Control 0 0 .3 0 Experimental Control 0 0 1 0 Experimental Control 0 0 0 0 Experimental Control 0 .3 .3 0 Experimental Control 0 0 0 0 Glyphosate .7 0 .3 0 Glyphosate 0 .3 0 0 Glyphosate 0 0 .3 0 Glyphosate 0 0 0 0 Glyphosate 0 0.3 .3 0 Glyphosate 0 0 0 0 Cut + Smother 0 0 .3 0 Cut + Smother 0 0 1 0 Cut + Smother 0 0 0 0 Cut + Smother .7 0 0 0 Cut + Smother 0 0 .7 .3 Cut + Smother 0 0 0 0 Cut 0 0 0 0 Cut 0 .3 .3 0 Cut 0 0 2 0 Cut 0 0 .3 0 Cut 0 0 .3 0 Cut 0 0 .7 0 Cut + Dalapon 0 0 .3 0 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 .3 0 Cut + Dalapon 0 0 .3 0 134

Appendix: Study plots raw data - Invertebrate numbers per core, Ballykelly 1998

TREATMENT C. marinus C. volutator Cyclorrhapha H. diversicolor Dalapon 0 12.7 .3 .7 Dalapon .3 16.3 0 1 Dalapon 0 .7 0 0 Dalapon 0 0 0 .3 Dalapon .7 2.7 0 0 Dalapon 0 1.0 0 0 Cut + Glyphosate 0 15.7 0 1.3 Cut + Glyphosate 1.3 8.3 0 1 Cut + Glyphosate 0 1 .3 .3 Cut + Glyphosate 0 2 0 0 Cut + Glyphosate .3 0 0 0 Cut + Glyphosate 0 1 0 0 Experimental Control 0 4.3 0 1.7 Experimental Control 0 0 0 0 Experimental Control 0 .3 0 0 Experimental Control 0 13 0 1.3 Experimental Control 0 10.3 0 1.3 Experimental Control 0 16.3 0 .3 Glyphosate 0 14.3 0 .3 Glyphosate 0 6.6 0 .7 Glyphosate 0 0 0 0 Glyphosate 0 0 0 0 Glyphosate 0 0 0 0 Glyphosate 0 2 0 0 Cut + Smother 0 23.7 0 .3 Cut + Smother .3 4 0 0 Cut + Smother 0 15.3 0 4.3 Cut + Smother .3 3.7 0 .7 Cut + Smother 0 0 0 0 Cut + Smother 0 .7 0 0 Cut .3 15.7 0 .3 Cut 0 10. 0 1 Cut 0 12.3 0 .3 Cut 0 0 0 0 Cut 0 .7 0 0 Cut .3 1.3 .7 0 Cut + Dalapon .3 7 0 1.7 Cut + Dalapon 0 7.7 .3 0 Cut + Dalapon .7 20.7 0 1.3 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 2 0 0 135

Appendix: Study plots raw data - Invertebrate numbers per core, Ballykelly 1998

TREATMENT H. ulvae L. littorea M. balthica Miridea Dalapon9000 Dalapon 1.7 0 0 0 Dalapon3000 Dalapon 9 0 .3 0 Dalapon .3 0 0 0 Dalapon 4 0 .3 0 Cut + Glyphosate 36.3 0 0 0 Cut + Glyphosate 11 0 0 0 Cut + Glyphosate 4.3 0 0 0 Cut + Glyphosate 2.3 0 0 0 Cut + Glyphosate 16 0 0 0 Cut + Glyphosate 19.3 0 0 0 Experimental Control 3.7 0 0 0 Experimental Control 5 0 0 0 Experimental Control 6.7 0 0 0 Experimental Control 8.7 0 0 0 Experimental Control 2 0 0 0 Experimental Control 10 0 .3 0 Glyphosate 3.3 0 0 0 Glyphosate 10.7 0 0 0 Glyphosate 8.7 .3 0 0 Glyphosate 8.7 0 0 0 Glyphosate 14.7 0 0 0 Glyphosate 1.3 0 0 0 Cut + Smother 2 0 0 .3 Cut + Smother 8.3 .3 0 0 Cut + Smother 6 .3 1 0 Cut + Smother 10.6 0 .3 0 Cut + Smother 7.3 0 0 0 Cut + Smother 15 0 0 0 Cut 4.7 0 0 0 Cut 3.7 0 0 0 Cut 6.7 0 0 0 Cut 4.6 0 0 0 Cut 3 0 0 0 Cut 7.7 0 .3 0 Cut + Dalapon 17 0 0 0 Cut + Dalapon 4.7 0 0 0 Cut + Dalapon 9 0 0 0 Cut + Dalapon 5 0 0 0 Cut + Dalapon 4.6 0 0 0 Cut + Dalapon .7 0 0 0 136

Appendix: Study plots raw data - Invertebrate numbers per core, Ballykelly 1998

TREATMENT Nematocera R.obtusa Dalapon 0 0 Dalapon 0 0 Dalapon 0 0 Dalapon 0 0 Dalapon 0 0 Dalapon 0 0 Cut + Glyphosate 0 0 Cut + Glyphosate 0 0 Cut + Glyphosate 0 .3 Cut + Glyphosate 0 0 Cut + Glyphosate 0 0 Cut + Glyphosate 0 0 Experimental Control 0 0 Experimental Control 0 0 Experimental Control 0 0 Experimental Control 0 0 Experimental Control 0 0 Experimental Control 0 .3 Glyphosate 0 0 Glyphosate 0 0 Glyphosate 0 0 Glyphosate 0 0 Glyphosate 0 0 Glyphosate .3 0 Cut + Smother 0 0 Cut + Smother 0 0 Cut + Smother 0 0 Cut + Smother 0 0 Cut + Smother 0 0 Cut + Smother 0 0 Cut 0 0 Cut 0 0 Cut 0 0 Cut 0 0 Cut 0 0 Cut 0 0 Cut + Dalapon 0 0 Cut + Dalapon 0 0 Cut + Dalapon 0 0 Cut + Dalapon 0 0 Cut + Dalapon 0 0 Cut + Dalapon 0 0 137

Appendix: Study plots raw data - Invertebrate numbers per core, Ballykelly 1999

TREATMENT A. tornatilis Aranae Brachycera C. marinus Dalapon0000 Dalapon 0 0 .7 0 Dalapon 0 0 1.3 0 Dalapon 0 0 .3 0 Dalapon 0 0 1.7 0 Dalapon 0 0 .7 0 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 0 .3 0 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 0 .3 0 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 .3 0 0 Experimental Control 0 0 0 0 Experimental Control 0 0 .7 0 Experimental Control 0 0 0 .7 Experimental Control 0 0 0 0 Experimental Control 0 0 0 0 Experimental Control 0 0 1 0 Glyphosate 0 0 0 0 Glyphosate 0 0 .3 0 Glyphosate 0 0 .3 0 Glyphosate 0 0 0 0 Glyphosate 0 0 0 0 Glyphosate .3 0 0 0 Cut + Smother 0 0 .7 0 Cut + Smother 0 0 .3 0 Cut + Smother 0 0 0 0 Cut + Smother 0 0 0 0 Cut + Smother 0 0 0 0 Cut + Smother 0 .3 0 0 Cut 0 0 0 0 Cut 0 0 0 0 Cut 0 0 0 0 Cut 0 0 .3 0 Cut 0 .3 0 .3 Cut 0 0 .3 1 Cut + Dalapon 0 0 1 0 Cut + Dalapon 0 0 0 .3 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 .7 0 Cut + Dalapon 0 0 1.3 0 Cut + Dalapon .3 0 .7 .3 138

Appendix: Study plots raw data - Invertebrate numbers per core, Ballykelly 1999

TREATMENT Coleoptera C. volutator Cyclorrhapha H.diversicolor Dalapon 0 4.6 0 0 Dalapon 0 5.3 0 .3 Dalapon 0 .3 0 0 Dalapon 0 .3 0 0 Dalapon 0 .3 0 0 Dalapon 0 1 0 0 Cut + Glyphosate 0 8 .3 .7 Cut + Glyphosate 0 .3 0 1.7 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate .3 0 0 .3 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 0 0 0 Experimental Control 0 4 0 0 Experimental Control 0 0 0 0 Experimental Control 0 1 0 0 Experimental Control 0 6 0 .3 Experimental Control 0 0 0 0 Experimental Control 0 .7 .3 0 Glyphosate 0 2 0 .3 Glyphosate 0 .7 0 1 Glyphosate 0 0 0 0 Glyphosate 0 0 0 0 Glyphosate 0 0 0 0 Glyphosate 0 .3 .7 0 Cut + Smother 0 .3 0 0 Cut + Smother .3 3.7 0 .7 Cut + Smother 0 2.3 0 .7 Cut + Smother 0 1.7 0 2.7 Cut + Smother 0 1.3 0 0 Cut + Smother 0 0 .3 0 Cut 0 2.3 0 0 Cut 0 10.7 0 1 Cut 0 .7 0 1 Cut 0 0 0 0 Cut 0 8.3 .3 1 Cut 0 .3 .7 0 Cut + Dalapon 0 8.3 0 .3 Cut + Dalapon 0 1.3 0 0 Cut + Dalapon 0 3 0 1 Cut + Dalapon 0 1.7 0 0 Cut + Dalapon 0 1.3 0 0 Cut + Dalapon 0 .7 0 0 139

Appendix: Study plots raw data - Invertebrate numbers per core, Ballykelly 1999

TREATMENT Hirundina H. ulvae L. littorea M. balthica Dalapon 0 59.3 0 0 Dalapon 0 98.7 0 0 Dalapon 0 54.7 0 .3 Dalapon 0 68 0 0 Dalapon .3 10 0 0 Dalapon 0 14 0 0 Cut + Glyphosate 0 33.3 0 0 Cut + Glyphosate 0 28 0 0 Cut + Glyphosate 0 30.7 0 0 Cut + Glyphosate 0 10.6 0 .3 Cut + Glyphosate 0 26.3 .3 0 Cut + Glyphosate 0 16.7 0 0 Experimental Control 0 40 0 0 Experimental Control 0 29.7 0 0 Experimental Control 0 19.6 0 0 Experimental Control 0 14.3 0 0 Experimental Control 0 12.3 0 .3 Experimental Control 0 14.7 0 0 Glyphosate 0 53.7 0 .3 Glyphosate .3 31 0 0 Glyphosate 0 23.3 .3 0 Glyphosate 0 35.3 0 0 Glyphosate 0 49.3 0 .3 Glyphosate .3 9 .3 0 Cut + Smother 0 3.7 0 0 Cut + Smother 0 54.7 0 0 Cut + Smother 0 46 0 0 Cut + Smother 0 5.3 0 .7 Cut + Smother 0 8.7 0 0 Cut + Smother 0 22 0 0 Cut 0 28 0 0 Cut 0 34 0 0 Cut 0 47 0 0 Cut 0 30 .7 0 Cut 0 19.7 0 0 Cut 0 10 0 0 Cut + Dalapon 0 85.3 0 0 Cut + Dalapon 0 53.7 0 0 Cut + Dalapon 0 48.3 0 0 Cut + Dalapon 0 8 0 0 Cut + Dalapon 0 3 0 0 Cut + Dalapon 0 2.3 0 0 140

Appendix: Study plots raw data - Invertebrate numbers per core, Ballykelly 1999

TREATMENT Nematocera R. obtusa Dalapon 0 0 Dalapon .3 0 Dalapon 0 .3 Dalapon 0 0 Dalapon 0 0 Dalapon 0 0 Cut + Glyphosate .3 0 Cut + Glyphosate 0 0 Cut + Glyphosate 0 1.3 Cut + Glyphosate 0 0 Cut + Glyphosate 0 .3 Cut + Glyphosate 0 0 Experimental Control 0 0 Experimental Control 0 0 Experimental Control .3 0 Experimental Control 0 0 Experimental Control 0 .3 Experimental Control 0 .3 Glyphosate 0 0 Glyphosate 0 0 Glyphosate 0 .7 Glyphosate 0 .3 Glyphosate 0 .3 Glyphosate 0 .3 Cut + Smother .3 .3 Cut + Smother 0 0 Cut + Smother 0 0 Cut + Smother 0 0 Cut + Smother .7 0 Cut + Smother .3 .3 Cut 0 0 Cut 0 0 Cut 0 .3 Cut 0 0 Cut 1 0 Cut 0 0 Cut + Dalapon .7 .3 Cut + Dalapon .7 .3 Cut + Dalapon .3 0 Cut + Dalapon 0 0 Cut + Dalapon 1 .7 Cut + Dalapon 0 0 141

Appendix: Study plots raw data - Invertebrate numbers per core, Ballydrain 1998

TREATMENT Brachycera C. marina C. marinus C. volutator Dalapon 0 .3 0 0 Dalapon 0 0 0 2 Dalapon 0 0 0 1.3 Dalapon 0 0 0 0 Dalapon 0 0 0 0 Dalapon .3 0 0 0 Cut + Glyphosate 2.7 0 0 12.7 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 0 .7 0 Cut + Glyphosate .3 0 0 0 Cut + Glyphosate 0 0 0 1.3 Cut + Glyphosate 0 0 0 0 Experimental Control .3 .7 0 3.7 Experimental Control 0 0 1.3 17.7 Experimental Control 0 0 0 0 Experimental Control .3 0 .3 0 Experimental Control .3 0 0 2 Experimental Control 0 0 .3 0 Glyphosate 0 0 0 .3 Glyphosate 0 .3 0 8.7 Glyphosate 0 0 0 0 Glyphosate 0 0 0 .3 Glyphosate 0 0 0 0 Glyphosate 0 0 0 0 Cut + Smother 0 0 0 9.3 Cut + Smother 0 0 0 .3 Cut + Smother 0 0 0 0 Cut + Smother 0 0 0 1.7 Cut + Smother 0 0 0 .7 Cut + Smother .3 0 0 0 Cut .3 0 0 6.3 Cut.3000 Cut .7 0 0 4.6 Cut0000 Cut0010 Cut0000 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 0 0 Cut + Dalapon .3 0 0 0 Cut + Dalapon 0 .3 0 12.3 Cut + Dalapon 0 0 0 10 142

Appendix: Study plots raw data - Invertebrate numbers per core, Ballydrain 1998

TREATMENT Cyclorrhapha H.diversicolor H.ulvae L.littorea Dalapon 0 0 13 0 Dalapon 0 0 3 0 Dalapon 0 0 9.7 0 Dalapon 0 1.3 73.7 .7 Dalapon 0 0 37.7 .7 Dalapon 0 .3 25.7 0 Cut + Glyphosate .3 .3 11.7 0 Cut + Glyphosate 0 0 139 1.3 Cut + Glyphosate 0 .7 185.7 0 Cut + Glyphosate 0 .7 20.3 0 Cut + Glyphosate .3 0 146.3 .7 Cut + Glyphosate 0 0 110 .7 Experimental Control 0 0 3.3 0 Experimental Control 0 .3 .7 0 Experimental Control 0 0 13.3 0 Experimental Control 0 0 88.3 .7 Experimental Control 0 0 14 .3 Experimental Control 0 0 126 1.7 Glyphosate 0 0 12.3 0 Glyphosate 0 .3 6 0 Glyphosate 0 0 179.3 1.7 Glyphosate 0 .7 5.3 0 Glyphosate 0 .7 48 .7 Glyphosate .3 0 68.7 2.3 Cut + Smother 0 0 24 .3 Cut + Smother 0 0 42.3 0 Cut + Smother 0 0 73.3 1 Cut + Smother 0 0 6.3 0 Cut + Smother .3 .7 6.6 0 Cut + Smother .3 0 9.3 0 Cut 0 .7 42.7 0 Cut0090 Cut009.3 Cut 0 0 77.7 0 Cut 0 .3 86 .7 Cut .3 .3 17.3 0 Cut + Dalapon 0 0 114.7 0 Cut + Dalapon 0 .3 20.3 0 Cut + Dalapon 0 0 64 0 Cut + Dalapon 0 0 6 0 Cut + Dalapon 0 .3 18.7 0 Cut + Dalapon 0 .3 12.3 0 143

Appendix: Study plots raw data - Invertebrate numbers per core, Ballydrain 1998

TREATMENT M. balthica M. edulis Nematocera Sphaeromatidae Dalapon 0 0 .3 0 Dalapon 0 .3 0 0 Dalapon0000 Dalapon 0 0 .3 0 Dalapon 0 0 1.3 0 Dalapon .3 0 0 0 Cut + Glyphosate .7 0 5 0 Cut + Glyphosate 0 0 0 0 Cut + Glyphosate 0 0 .7 0 Cut + Glyphosate 0 0 .3 0 Cut + Glyphosate 0 .3 1 0 Cut + Glyphosate 0 0 1.3 0 Experimental Control 0 0 0 0 Experimental Control 0 0 .3 0 Experimental Control 0 .3 2 0 Experimental Control 0 0 .3 0 Experimental Control .3 0 0 0 Experimental Control 0 0 .7 0 Glyphosate 0 0 0 0 Glyphosate 0 0 0 0 Glyphosate 0 0 .3 0 Glyphosate .3 0 1.3 0 Glyphosate 0 0 .3 0 Glyphosate .7 0 0 0 Cut + Smother 0 0 1.3 0 Cut + Smother 0 0 0 0 Cut + Smother 0 0 .7 0 Cut + Smother 0 0 .3 0 Cut + Smother 0 0 1 0 Cut + Smother 0 0 1.3 0 Cut 0 0 5.3 .3 Cut .3 0 0 0 Cut 0 0 0 0 Cut 0 0 0 0 Cut 0 0 0 0 Cut 0 0 1 0 Cut + Dalapon 0 .3 1.3 0 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 1 0 Cut + Dalapon 0 0 1.3 0 Cut + Dalapon .3 1 .7 0 Cut + Dalapon .7 0 .7 0 144

Appendix: Study plots raw data - Invertebrate numbers per core, Ballydrain 1999

TREATMENT Brachycera C. marina C. marinus C.volutator Dalapon 0 0 0 0 Dalapon .3 0 0 9.3 Dalapon 0 0 0 .7 Dalapon 0 0 .3 0 Dalapon 0 0 .7 4.7 Dalapon 0 0 0 1 Cut + Glyphosate .7 0 0 3.3 Cut + Glyphosate 0 0 .7 0 Cut + Glyphosate 0 0 .3 1 Cut + Glyphosate 0 .3 0 5.7 Cut + Glyphosate 0 .3 .3 2 Cut + Glyphosate 0 0 .3 4 Experimental Control .3 .3 0 8.7 Experimental Control 1 0 1.3 20 Experimental Control .3 0 0 1.7 Experimental Control 0 0 .3 0 Experimental Control 0 0 .3 3.3 Experimental Control .3 0 0 0 Glyphosate .7 .3 .3 3.6 Glyphosate 0 .3 .3 11 Glyphosate 0 0 0 0 Glyphosate .3 0 .3 6.3 Glyphosate .3 0 4.3 4.3 Glyphosate 0 0 .3 0 Cut + Smother 2.7 0 0 5.3 Cut + Smother .7 0 0 18.6 Cut + Smother 0 0 0 7.3 Cut + Smother 2.7 0 0 2 Cut + Smother .7 0 0 41.7 Cut + Smother 0 0 .3 0 Cut 0 0 0 14.7 Cut .3 0 .3 10.7 Cut.7000 Cut00029 Cut.7000 Cut 0 0 0 10.7 Cut + Dalapon 0 0 0 0 Cut + Dalapon 0 0 0 0 Cut + Dalapon .3 0 .3 25.7 Cut + Dalapon 0 .3 0 67.3 Cut + Dalapon 0 0 0 52.3 Cut + Dalapon 0 0 0 .7 145

Appendix: Study plots raw data - Invertebrate numbers per core, Ballydrain 1999

TREATMENT Cyclorrhapha H.diversicolor H.ulvae L.littorea Dalapon 0 0 12 0 Dalapon 0 0 4 0 Dalapon 0 0 7.7 0 Dalapon 0 .3 17 0 Dalapon 0 .3 13 0 Dalapon 0 .7 81.3 .3 Cut + Glyphosate 0 0 2.7 0 Cut + Glyphosate 0 0 33.7 0 Cut + Glyphosate .3 1.3 27.7 .3 Cut + Glyphosate 0 0 5.7 0 Cut + Glyphosate 0 .7 18 0 Cut + Glyphosate 0 .7 4 0 Experimental Control 0 0 19 0 Experimental Control 0 .3 9 0 Experimental Control 0 1 5.3 0 Experimental Control 0 0 79.7 0 Experimental Control 0 0 5.3 0 Experimental Control 0 0 90.7 1 Glyphosate 0 0 56.3 0 Glyphosate 0 1.7 12.3 0 Glyphosate 0 0 12.3 .3 Glyphosate .3 .3 3 0 Glyphosate 0 .7 96.7 .7 Glyphosate 0 .3 9.3 0 Cut + Smother 0 0 7 0 Cut + Smother 0 .7 4 0 Cut + Smother 0 0 1.7 0 Cut + Smother .3 .7 4 0 Cut + Smother 0 .3 4.3 0 Cut + Smother 0 0 22.3 0 Cut 0 0 2.3 0 Cut 0 .3 11.7 0 Cut 0 .3 1.7 0 Cut 0 .3 40.7 0 Cut.7.3190 Cut 0 .3 4.3 0 Cut + Dalapon 0 .7 3.3 0 Cut + Dalapon 0 0 9.7 0 Cut + Dalapon 0 .3 2.3 0 Cut + Dalapon 0 .7 2.7 0 Cut + Dalapon 0 .7 5 0 Cut + Dalapon 0 1.7 21 0 146

Appendix: Study plots raw data - Invertebrate numbers per core, Ballydrain 1999

TREATMENT M.balthica M.edulis Nematocera Dalapon 0 0 26.7 Dalapon 0 0 18.7 Dalapon 0 0 15.3 Dalapon 0 0 34 Dalapon 0 0 13.7 Dalapon 0 0 35 Cut + Glyphosate 0 0 22 Cut + Glyphosate 0 0 18.7 Cut + Glyphosate 0 0 4 Cut + Glyphosate 0 0 15 Cut + Glyphosate 0 0 17.3 Cut + Glyphosate .3 0 43 Experimental Control 0 0 6.3 Experimental Control 0 0 5 Experimental Control 0 0 37.7 Experimental Control 0 0 14 Experimental Control .3 0 9.3 Experimental Control .3 0 28.3 Glyphosate 0 0 9.3 Glyphosate 0 0 5.7 Glyphosate 0 0 5.7 Glyphosate 0 0 16.7 Glyphosate 0 0 11 Glyphosate 0 0 29 Cut + Smother 0 0 5.3 Cut + Smother 0 0 1 Cut + Smother 0 0 9 Cut + Smother 0 .3 15.7 Cut + Smother 0 0 6.3 Cut + Smother 0 0 9 Cut 0 0 12.3 Cut 0 0 14.7 Cut 0 0 18.3 Cut 0 0 11 Cut 0 0 30 Cut 0 0 38.7 Cut + Dalapon 0 0 6 Cut + Dalapon 0 0 1.3 Cut + Dalapon 0 0 25 Cut + Dalapon 0 0 6.3 Cut + Dalapon 0 0 60 Cut + Dalapon 0 0. 24 147