The Para Grass City, Where the Grass Is Green and the Snakes Are Pretty: the Ecology of Herpetofauna in an Invaded Wetland
ResearchOnline@JCU
This file is part of the following reference:
Pearcy, Ashley Grace (2007) The para grass city, where the grass is green and the snakes are pretty: the ecology of Herpetofauna in an invaded wetland. Masters (Research) thesis, James Cook University.
Access to this file is available from:
http://eprints.jcu.edu.au/29566/
The author has certified to JCU that they have made a reasonable effort to gain permission and acknowledge the owner of any third party copyright material included in this document. If you believe that this is not the case, please contact [email protected] and quote http://eprints.jcu.edu.au/29566/
The Para Grass City, Where the Grass is Green and the Snakes are Pretty: the Ecology of Herpetofauna in an Invaded Wetland
Thesis submitted by Ashley Grace Pearcy BS (Tampa) in November 2007
for the research degree of Master of Science in Tropical Ecology within the School of Marine and Tropical Biology James Cook University
STATEMENT OF ACCESS
I, the undersigned, the author of this thesis, understand that James Cook University will make this thesis available for use within the University Library and, via the Australian Digital Theses network (unless granted an exemption), for use elsewhere.
I understand that, as an unpublished work, a thesis as significant protection under the Copyright Act and;
I do not wish to place any further restrictions on access to this work.
______
ii STATEMENT ON SOURCES DECLARATION
I declare that this thesis is my own work and has not been submitted in any other form for another degree or diploma at any university or other institution of tertiary education. Information derived from the published and unpublished work of others has been acknowledged in the text and a list of references is given.
______
iii Significant Contributions
Funding Burdekin Dry Tropics
Assitance in obtaining funding Lin Schwarzkopf Christopher Johnson
Project done in conjunction with Queensland Parks and Wildlife Service Australian Commonwealth Scientific and Research Organization (CSIRO)
Research Assistants Samantha Forbes Benedicte Eftevand Lauren Hodgson
Assistance with snake regurgitation and identification data collection John Llewelyn
Statistical support Leonie valentine
Supervision and revisions to thesis Christopher Johnson Ivan Lawler Andi Cairns
iv Acknowledgements
This past year has been demanding, full of unexpected difficulties and inspiring accomplishments. I want to thank everyone that has supported me throughout this time, both those that doubted the possibilities that lay ahead and those that had full faith in what may come.
To my parents: I know the stress I lay on you; thank you for taking it in stride. Thank you for supporting me in my Epicurean lifestyle and my never-ending search for happiness. “You are the wind beneath my wings.” To the rest of my family: Bro, thanks for the support despite my off kilter way of doing things. Thank you to the B’s for understanding my absence for the family gatherings and for the support during the rough times. Nic and Fen, thanks for making me laugh and for being my solid ground on the other side of the world. Pa, thanks for always offering up advice, even if I do not always take it. Neane, thanks for being interested in what I do even though it seems so far-fetched from the world in which I grew up. Thanks to everyone for supporting my adventures.
To the team: Sammy Forbes, you have been an inspiration on and off this project. Your hard work leaves no doubt for the future to which you aspire and the guaranteed success you will achieve. Thank you for your dedication and heart. Thanks for the para grass runs and the para grass city. Thanks for the dancing, the music, the laughing, and the surfing. I would not have made it through this without you. Much love and peace. Benedicte Eftevand- keep racking up that CV! I cannot thank you enough for your commitment in the field. Thanks for making me laugh and continually reminding me to “chill out”. Angus McNab thanks for all the help and for being an all-around good mate.
Adrian McMahon, you are my right hand man. I could not have made it through this without the long hours and you by my side to get me through them. Thanks for the technical and moral support! Emadch Beck, Chris Ryen, and Ashley Field, thanks for being supportive and listening to my crazy stories. Thanks for the much-needed distractions which have kept me sane.
To the supervisors: Chris Johnson and Ivan Lawler, thanks for taking me on and helping me to accomplish what I set forth to do.
Thank you to: Andi Cairns for her last minute revisions, advice, and chats; Leonie Valentine, without you I would have never had any confidence to tackle statistics; Tony Grice, Mike Nichols, and Paul Williams for advice on experimental design and working in the Town Common.
A big thank you to anyone and everyone else who has supported me throughout this project and the past year: Richard Rowe, Richard Pearson, Lin Schwarzkopf, Lauren Hodgson, John Llewelyn, Rob Graham, Jason Schaffer, Brian Ross, Sam Noonan, Miwa Takahashi, Zuzu Wahab, Stephen Kolomoyjec, Curtis Lind, Stefan Walker, Dan Ryen, Gerard Wanecheck, Øyvind Syrrist, Eivind Undeheim, Ray Lloyd, Ben Pearson, Andrea Bird, Spencer Bryde
v Abstract
Wetlands are susceptible to invasion by both flora and fauna due to their availability of differing habitat. This induces the enactment of treatments either for prevention or management of introduced species. Many invasive management practices are done with a focus on single taxa, especially with concern to wetlands. Wetlands offer a dynamic habitat for several species, and being aware of how these habitats are used by all species can offer a beneficial knowledge base for those enacting management techniques to ensure minimal consequences. This thesis investigated the influence of invasive species on herpetofauna in a wetland. The majority of this study focused on habitat use and the effects of the para grass Urochloa mutica.. By trapping along the ecotone gradient, abundance and location were used to determine primary habitat. It is concluded that the majority of herpetofauna species are more abundant in the woodland than in the margin or floodplain. This benefits the management treatments done to the floodplain by alleviating loss through the confirmation of larger populations in adjacent woodlands. A few species, however, associated mostly with the floodplain were identified such as Lampropholis delicata and Litoria fallax. These species could face greater consequences from management techniques.
QPWS/CSIRO are experimenting with management techniques to control the spread of the invasive para grass. The effect of these treatments (burning and grazing) on the keelback snake (Tropidonophis mairii) during the wet season was studied by setting up aquatic traps in the floodplain in paddocks each with a different para grass treatment. The results found that snakes are more abundant in undisturbed, or control sites. This presents questions concerning the consequences of the weed vs. those of the management practice.
A final study was done on the relation between Tropidonophis mairii and cane toads Bufo marinus, finding that despite the availability of cane toads, keelbacks selectively avoid them in their diet. Keelbacks are known to consume the toxic cane toad, but this defined preference for other individuals, shows that they are not a viable biological control method for the quickly expanding cane toad population.
vi
Table of Contents Pg Signed statement of access ii Signed statement of sources iii Significant Contribution iv Acknowledgments v Abstract vi Table of Contents vii List of Tables ix List of Figures x I. Chapter 1: Introduction 1 A. Aims 6 II. Chapter 2: The Good, the Bad and the Ugly: Benefits and Disadvantages to Wetland Weed Management 7 A. Abstract 7 B. Introduction 7 C. Sustained Changes 8 D. Weed Management 9 1. Introduction 9 2. Types of Management 10 a. Chemical 10 b. Physical 10 c. Mechanical 11 d. Biological 11 3. Conclusion 12 E. Yay or Nay to Management: Determining the Benefits and Consequences of Weed Management 12 F. Success 16 G. Conclusion 16 III. Chapter 3: Habitat Use by Herpetofauna in a Wetland: Who Lives Where? 18 A. Abstract 18 B. Introduction 18 C. Aims 21 D. Methods 21 1. Study Site 21 2. Experimental Design 23 3. Analysis 26 a. Trap Efficiency 26 b. Habitat 26 c. Weather 27 d. QPWS/CSIRO Para Grass Treatments 27 E. Results 27 1.Trap Efficiency 29 2. Habitat 31 a. Reptiles 31 b. Amphibians 34 3. Dominant Vegetation 37 4. Effect of Weather 39
vii 5. Pre vs. Post Burn 40
F. Discussion 40 1. Habitat Preference of Reptiles and Amphibians 40 2. Dominant Vegetation 42 3. Trap Efficiency 42 4. Effects of Weather 43 5. The Effect of Para Grass Treatments 43 6. Value for Management of Para Grass 43 7. Concerns Caused by Treatments 44 G. Conclusion 45
IV. Chapter 4: Living in an Invaded Wetland: the Natural History of Tropidonophis mairii in the Townsville Town Common 46 A. Abstract 46 B. Introduction 46 C. Aims 48 D. Methods 48 1. Experimental Design 48 2. Analysis 50 E. Results 50 1. Population Size and Composition 50 2. Habitat Use 51 3. Predator-Prey Dynamics 54 F. Discussion 55 1. Population Dynamics 55 2. Floodplain Use 55 3. Feeding Habits 56 G. Conclusion 57 1. Population and Feeding Habits 57 2. Floodplain Treatment Use 57 V. Chapter 5: General Conclusion 58 VI. References 61
viii
List of Tables Table Pg 2.1 Weed ecology and the effect on habitat: predicting invasiveness and effectiveness of treatment 13 3.1 Timeline of data collection 24 3.2 List of herpetofauna species trapped on the Town Common 28 3.3 ANOVA results for untransformed trap efficiency data 29 3.4 ANOVA results for habitat analysis 31 3.5 Interaction between session and distance for reptile and amphibian captures 40 4.1 Prey available to prey consumed by keelbacks in the Town Common 54
ix
List of Figures Figure Pg 1.1 Map of Queensland 3 1.2 Average monthly temperature for Townsville 4 1.3 Average monthly rainfall for Townsville 4 1.4 Average monthly rainfall for Townsville 2006 4 1.5 Topographic map showing location of Townsville Town Common Conservation Park and the locations of QPWS/CSIRO para grass treatment plots 5 3.1 Location of study site within Townsville Town Common and in comparison to QPWS/CSIRO treatments 22 3.2 Example of woodland trap and surrounding vegetation 23 3.3 Transect with array locations 24 3.4 Land array 25 3.5 Water array 25 3.6 Trap efficiency for herpetofauna by number of reptiles and amphibians captured 30 3.7 Number of reptiles trapped and reptile species richness by habitat 32 3.8 Principal Component Analysis of reptile community composition across three habitats 33 3.9 Eigen vectors for reptile community PCA 34 3.10 Number of amphibians trapped and amphibian species richness by habitat 35 3.11 Principal Component Analysis for amphibian community composition 36 3.12 Eigen vectors for amphibian community PCA 36 3.13 Dominant vegetation for both reptile and amphibian species richness and numbers trapped 37 3.14 Dominant vegetation associated with individual reptile species 37 3.15 Dominant vegetation associated with individual amphibian species 38 3.16 Rainfall in millimeters in the Town Common by day of trapping 39 3.17 Rainfall to number of herpetofauna trapped 39 4.1, Population composition of keelbacks in Town Common 4.2 50 4.3 Number of snakes trapped to treatment 51 4.4 Number of keelbacks trapped in relation to percent para grass coverage 52 4.5 Number of snakes trapped in different ranges of water depth 52 4.6 Species composition and abundance changes over time 53 4.7 C. aloboguttata and keelback comparative abundance by treatment in the floodplain 53 4.8 Diet selectivity of successfully regurgitated keelbacks 54 4.9 Diet selectivity of keelbacks by species 55
x Chapter 1 Introduction
Wetlands are defined as soils formed under saturated conditions that are drained, undrained, or seasonally wet long enough to support hydrophilic plants (Heimlich 1994). Because of the availability of water and the diversity of habitat offered, wetlands are prone to invasion. This invasion can be by plants, animals, or both. The dynamic habitat offered by wetlands allows for a unique assemblage of species. Many species use the terrestrial habitat while some use the aquatic habitat. There are a few species which use both habitats, whose use of water is made possible by the proximity of land and vice versa. Overall, each species uses the available habitats, either aquatic or terrestrial, to suit their unique life history functions (Gibbons 2003). Acknowledging this fact allows for appropriate, more effective management plans.
Besides different habitat use by different species, it is important too to acknowledge change by season, as wetlands tend to exhibit marked seasonal differences; a full floodplain in the peak of the wet season recedes to a wetland savanna in the dry season. This change in available habitat will induce a response by the inhabitants (Martin 1988). Species abundance, community composition and habitat use can change with season (Martin 1988; Sinsch 1988; Brown 2002). Using this knowledge, management practices can be implemented with minimal negative effect on wetland populations. Problems occur when management practices for control of invasive exotics of both plants and animals are implemented without a full understanding of all species inhabiting the wetland (Bower 2006).
The importance of the study of habitat use has been well-documented. Adult amphibians will move to breeding pools, deposit eggs, and then return to terrestrial sites (Duellman 1986). Duellman’s work recommended the preservation of isolated pools for breeding amphibians. Both reptiles and amphibians have been shown to use woodlands adjacent to wetlands up to 290 meters from water line to carry out critical life functions (Semlitsch and Bodie 2003). Burke (1995) found that freshwater turtles (Kinosternon subrubrum, Psuedemys floridana, and Trachmeys scripta) nested outside of their protected areas in South Carolina, suggesting the extension of buffer zones. Another turtle study found that Deirochelys reticulara in Virginia used terrestrial sites for hibernation and aestivation, some spending more time on land than in water (Buhlmann 1995). Buhlmann concluded that implementing conservation measures without full knowledge of terrestrial habitat use would result in
1 failure. Habitat studies can be beneficial in the extension of buffer zones in woodlands adjacent to bodies of water in wetlands, establishment of protected areas, and habitat encroachment policies for parks and wildlife services.
However, those are results from normal behaviours of herpetofauna in natural settings. In areas of anthropogenic disturbance, other responses occur. For example, carpet pythons (Morelia spilota imbricata) tend to fare well in disturbed areas where fire has created new habitat in fallen trees (Pearson 2005). On the other hand, prescribed fires have had negative effects on ant communities in Florida pine savannas (Izhaki 2003). It is important to recognize that the effect of disturbance is both species and site specific (Sawadogo 2005), emphasizing the importance of studies of habitat use by all species in a wetland before enacting management plans. These studies can help eliminate loss due to invasive management techniques, habitat encroachment by suburbia, and agricultural land use.
Not only is it important to determine habitat use but also interactions with other species. This includes both native and exotic flora and fauna. These interactions directly relate to habitat use as many of these interactions are predator-prey associations, where a prey seeks cover from predator or a predator seeks advantage over prey. Large mouth bass (Micropterus salmoides) and northern pike (Esox lucius) move to vegetated areas in the streams in which they live to ambush their prey rather than utilising open water (Savino 1989). Another example of interaction with resulting change in habitat is niche displacement by invasives (Mooney 2001). This can be seen in animals such as the grey squirrels overtaking the native red squirrel habitat in North America (Kenward 1993). Interactions which effect habitat use can occur as a result of either natural behaviour or as a response to exotics both flora and fauna.
Tropical wetlands are prone to invasion due to their climate, availability of water, and diverse species assemblage (Zedler 2004). Some common intruders in Australian wetlands are Urochloa mutica (para grass) and Bufo marinus (cane toad). Para grass is problematic as it forms a large vegetation mat, clogging natural filtration processes and flood zones and may also inhibit growth of native species (Douglas 2004).
2 Cane toads are an invasive species which is negatively impacting the Australian fauna (Covacevich 1975). This toad is toxic and can be consumed by very few animals. This aspect combined with its high reproductive rate allows for cane toad populations to increase unhindered by natural predators (Lampo 1998).
Figure 1.1 Map of Queensland. The blue dot represents Townsville, the red, Brisbane (Meterology 2007)
The Townsville Town Common is a wetland located in tropical North Queensland (Figure 1.1), Australia, which has been overrun by invasive plant and animals species (Grice 2006). In 1985, Queensland Parks and Wildlife Services declared the 3245 ha of land known as the Town Common as a conservation park. This area is part of the Bohle River floodplain and is a coastal wetland, inundated by high tides. This tropical wetland was originally a community grazing ground for cattle from the 1800’s to 1970’s. The cattle were fully removed by 1990.
Average temperatures range from a high of 44ºC in the peak of the wet season (December-March) to a low of nearly 0ºC in the dry season (June-September) (Figure 1.2). Average rainfall in the area can be up to 300 mm during the wet season (Figure 1.3). However, annual rainfall varies just as much as the average (Figure 1.4).
3
Figure 1.2 Average monthly temperature for Townsville (Meterology 2007)
Figure 1.3 Average monthly rainfall for Townsville (Meterology 2007)
Monthly Rainfall for Townsville 2006
500 450 400 350 300 250 200
Rain (mm) 150 100 50 0 Jan 1 Feb2 Ma3r Apr4 May5 Jun 6 Jul 7 Aug8 Sep 9 Oc10t Nov11 De12c Figure 1.4 Average monthly rainfall for Townsville in 2006 (Congdon 2006)
4 The two major problems of the Townsville Town Common are para grass Urochloa mutica, which is undergoing treatment (Williams 2005), and cane toads Bufo marinus, which have not yet undergone management practices. CSIRO/QPWS are experimenting with management techniques for the para grass. The para grass treatments were enacted in the floodplain of Long Swamp in the Town Common to open up the area for nesting waterfowl. CSIRO/QPWS have defined twelve 200m X 300m paddocks subject to the following treatments: three burned, three grazed, three burned and grazed, and three control plots (Figure 1.5). While efforts have been successful for waterfowl by decreasing the density of para grass to allow for nesting (Valentine 2007), there have been consequences for other taxa (Bower 2006).
Bower found that herpetofauna abundance reduces as a result of these treatments on the herpetofauna in the floodplain of this wetland. The project studied the direct effects of weed treatments on herpetofauna found in each of the CSIRO/QPWS paddocks. However, the adjacent woodland was not studied.
Figure 1.5 Topographic map showing location of Townsville Town Common Conservation Park and the locations of CSIRO/QPWS para grass treatment plots. (Bower 2006)
To date there is no published study of general habitat use by animals in the Town Common. As mentioned earlier, habitat use will differ by species and by season.
5 This knowledge could help in the management regime, by determining primary habitat and thus the extent of the impact of treatments on inhabitants of the floodplain. Understanding habitat use would also indicate the extent to which the invasives, both flora and fauna, have entered into a natural state, where native species benefit from the presence of the invasive by interacting with the invasive as if it were native (Richardson 2000; Daehler 2001). If a natural state exists for the invasive species in the wetland, harmful consequences of management techniques can outweigh those caused by the exotics by damaging available and used habitat.
The objective of this study is to determine the primary habitat for herpetofauna in a wetland, i.e. which species prefer the adjacent woodland, the margin between woodland and floodplain, or the floodplain itself?
The study goes further to discuss the influence of management treatments of the para grass and of the invasive cane toad on a single species, Tropidonophis mairii, the keelback snake. Dietary habits of keelback snakes were examined to determine preference for or against cane toads, thereby defining the interaction between a predator and an introduced prey. Keelbacks are known as on of the few species which can consume cane toads and not die from the toxins (Covacevich 1975), suggesting that cane toads are a viable prey for the keelback. Applying information from habitat use and interaction studies to both management and conservation efforts can maximise the positive outcomes by reducing potential negative impacts of these efforts.
Aims of this project: • Discuss benefits and consequences of wetland weed management (Chapter 2) • Determine dominant habitat of herpetofauna in a wetland (Chapter 3) • Determine if woodland populations buffer loss incurred by weed treatments (Chapter 3) • Determine whether if there is an change in herpetofauna numbers in areas outside of CSIRO/QPWS treatment sites following treatments (Chapter 3) • Determine effect of para grass treatments on keelback abundance (Chapter 4) • Determine feeding habits of keelbacks with concern to cane toads (Chapter 4)
6 Chapter 2 The Good, the Bad, and the Ugly: Benefits and Disadvantages of Wetland Weed Management
Abstract Many factors must be considered when planning the management of weeds in wetlands, from the ecology and effect of the weed to the effects of the treatment on other flora and fauna. Many studies have addressed the impacts of invasive weeds, but few have assessed potential negative effects caused by treatments aimed at controlling weeds before embarking on management. Treatments such as burning can have negative impacts on different taxa or make the area more susceptible to invasion by other weeds. The possible side-effects of herbicide use are well documented, especially with respect to groundwater contamination. In short, some weed management treatments can do more harm than good. Because of this, the effects of the weed on biodiversity, as well as the consequences of control treatments for the invaded community, should be considered in developing weed management plans. This review presents questions that should be explored before implementing weed management plans in any wetland.
Introduction Weeds are plants that occur where they are not wanted, and which have effects perceived by people as negative (Monaco 2002). In Australia, 1500 to 2000 plant species have been introduced since European settlement and over 200 of these are now considered noxious (or harmful) weeds (Humphries 1991). Some species were brought in as fodder for cattle. Only five percent of pasture species introduced to northern Australia are now recommended for use, and thirteen percent are listed as weeds (Lonsdale 1994). Other species were brought in as ornamentals for aesthetic reasons. For example, water hyacinth (Eichornia crassipes) has become a worldwide invasive species, spreading from Brazil throughout South America to all continents except Antarctica, and thriving in both tropical and temperate climates (Penfound 1948; Schmitz 1988; Howard 1998; Moreira 1999). Invasive plants such as this are a great threat to native flora and fauna of both tropical and temperate regions (Mashhadi 2004) as they may decrease both animal and plant diversity (D'Antonio 1992; Rea 1999; Zedler 2004).
Weeds are potent agents of environmental change (D'Antonio 1992; Turner 1997), and are a common problem throughout wetlands worldwide. Wetlands are thought to be more susceptible than other habitats to weed invasion because of their greater
7 water and nutrient availability and low native plant species richness (Fox 1986). Melaleuca quinquenervia in North America (Turner 1997), Purple loosestrife (Lythrum salicaria) in the United States (Blossey 2001; Morrison 2002), Mimosa pigra and Urochloa mutica (para grass) in Australia (Lonsdale 1993; Williams 2005) are a few extreme examples of the world’s most noxious weeds that alter the state of the wetlands which they invade.
The control of weeds is becoming increasingly necessary to preserve existing wetlands (van der Valk 2006). In many wetlands, treatments are enacted to help manage, control, and/or possibly eradicate weeds and restore the wetland to its natural state (or something as close as possible to the natural state). These management practices range from burning and grazing, mechanical methods, introduction of biological controls, to mass herbicide use. Many of these practices are implemented to improve the habitat for particular groups of native organisms, e.g. waterfowl (Kushlan 1986), with little attention to broader effects on other taxa. Negative effects can and do occur following weed management. These can include habitat loss for certain species, stress caused by disturbance, and chemical side- effects of herbicides (Charudattan 2001; Bower 2006). It is necessary to achieve a balance between the negative and positive effects of weed management when developing management plans for a wetland.
Sustained Changes The history of wetland invasion by weeds is often difficult to reconstruct as many weeds have been in their non-native habitat for hundreds of years. Reasons for weed introduction are numerous and in most cases are due to human activity. Globalisation of trade has created many opportunities for the accidental movement of plants between continents. Ballast water has been a known culprit for the introduction of aquatic weeds (Padilla 2004), and international trade is responsible, unintentionally, for the introduction of over 80% of harmful species introduced into the United States, according to an Office of Technology Assessment report in 1993 (Mashhadi 2004). This movement and introduction of exotics has led to a worldwide weed problem.
Long term, the impact of invasion is change in biodiversity. Weeds can alter the state of a wetland, out-competing native vegetation and decreasing the wetland’s biodiversity. Invasive weeds typically have a high tolerance for disturbance and can thus become dominant in highly-disturbed settings (Thompson 1987; Hobbs 1992;
8 Wilcox 1995; Mashhadi 2004). Because of this ability to thrive in disturbed areas, weed control techniques may, in fact, be doing more to encourage growth than inhibit weeds if they themselves cause significant disturbance. Many weed management plans are associated with disturbance treatments such as burning and grazing. These techniques may have short-term benefits because the disturbance allows for a successional replacement with more weeds. This disturbance can also cater to the growth of other species that would normally not dominate. For example, cattle may only graze a plant when it is in flower and/or plants burned at different life stages may survive better than others under stress (DiTomaso 1997). It is therefore important to have comprehensive knowledge of the impacts of the treatments on the habitat to be managed.
Fauna may respond in different ways to the presence of weeds: (1) they may be relatively unaffected, (2) the effect can be detrimental to the point of habitat destruction, or (3) the weed can actually improve the habitat for animals. The first case (1) becomes a factor when the invasive species is structurally similar to the native plant species (Douglas 2003). The second scenario (2) is of more concern. Habitat loss due to weeds choking out native vegetation is the driving force behind negative impacts of weeds (Turner 1997). As mentioned previously, this can decrease biodiversity. The loss or decline of a single species can impact on the entire community by altering normal interactions between species (both flora and fauna), either through predator-prey relationships and/or niche/habitat displacement. The final case (3) is the naturalization of the weed for a species where the invasive species benefits native species (Richardson 2000; Daehler 2001). Naturalization occurs after the weed has inhabited an area long enough to be of benefit to native species through habitat use or as a prey item (Richardson 2000). This can be seen in fig trees (Ficus sp.) and wasps (Eupristina sp.) in Florida where the invasive fig is occupied and pollinated by the native wasp (Nadel 1992). This presents a concern as the net benefits of the weed may outweigh its negative effects, and thus removing the weed would have a deleterious impact on the habitat.
Weed Management Precautionary and/or early action is most likely to be successful in completely eradicating weeds, because in the early stages of invasion they are isolated or in small patches. Because of this, most management practices are enacted as a result of speculation on potential consequences of weed invasion to an area (Blossey 2001). Following weed establishment, however, eradication may become
9 impossible, and the aim becomes the control of the weed population and its spread on a feasible scale. Management plans are put in place for several reasons: from the ecological, such as preservation of waterfowl nesting habitat (Kushlan 1986; Ferdinands 2005), to the economical, such as the stabilization of flood zones (Salvasen 1990). Weeds are site and species-specific in their response to management techniques (Sawadogo 2005), so each weed species demands a different treatment to maximize control and reduce the rate of spread.
Types of Management Chemical Chemical control of weeds depends on the use of herbicides to control their spread. Herbicides may be effective and fast-acting but often provide only short-term control (Charudattan 2001). Herbicides tend to wash away with rainfall or flooding. Furthermore, it is also common for weeds to become resistant to herbicide (DiTomaso 1997). Herbicides can be applied to large amounts of area with little effort and maximum effect; however, this means that they also have the potential to destroy native vegetation. Another possible negative effect is that the chemicals can enter the ground or surface water systems and be a concern for native wildlife (DiTomaso 1997). Furthermore, the dead and decaying material left behind from the destroyed weeds can have an ecological impact, especially on water quality (Moreira 1989; Charudattan 2001).
Physical The physical control of weeds is restricted to techniques such as burning and grazing using introduced livestock for management purposes. These are used frequently worldwide for a variety of weeds (Briese 1996; Williams 2005). Other options specifically for aquatic weeds include a dye which inhibits photosynthesis by blocking specific colors in the UV spectrum, or benthic coverings such as sand-gravel mixtures, burlap, polyethylene, and synthetic rubber (Charudattan 2001).
Most physical weed management techniques other than burning and grazing operate in the short term and on small scales. They are ineffective in large wetlands due to sheer amount of materials needed e.g. plastics to cover ground (common for aquatic weeds) and gravel for deposition to decrease the open area available for juvenile establishment. Burning and grazing can be done on large scales, but results tend to be patchy (Jacqmain 1999; Sawadogo 2005). These approaches have also been shown in some cases to have negative impacts on herpetofauna (Bower 2006) and
10 small mammals (Andersen 2005). Short-term negative impacts on invertebrate communities following burning have also been confirmed (Siemann 1997; Izhaki 2003). While the majority of negative effects caused by physical weed management techniques remain short-term, potential long-term impacts are not yet fully understood.
Mechanical The term “mechanical control” refers to the pulling out of the weeds. This can include tilling, mowing, cultivating, and hand pulling. This technique is expensive and often ineffective (Charudattan 2001) unless it is used in combination with other management regimes. An example of mechanical control occurs in Portugal where hydraulic excavators and floating harvesters are used extensively to manage invasive weeds like water hyacinth (Eichornia crassipes) and parrotfeather (Myriophyllum aquaticum) (Moreira 1999). Hand-pulling Mimosa pigra in combination with physical and chemical control techniques has reduced its spread in Kakadu National Park, Australia (Cook 1996).
Mechanical techniques also have a tendency, like the chemical control methods, to leave a mass of dead vegetation, depending on the mechanical approach used. Another common disadvantage of mechanical control is that most machines are non- selective, so native vegetation is removed along with the invasive weeds (Charudattan 2001).
Biological Biological control of weeds depends on the introduction of a natural predator or disease or herbivore of the weed to help control its population. This is generally considered to be the ideal remedy for invasive weeds as the treated habitat usually sustains less damage (McMurty 1995). Biological control is also the most economically sustainable, environmentally safe, and enduring weed management technique because its side effects are few, and it can be effective over large areas (McMurty 1995). Control agents suppress the weed and thus encourage growth of native vegetation (Charudattan 2001).
Introduction of biological controls must be done following an intensive study of the effects of the control organisms on the habitat and on native species, as some species introduced as biological control agents have themselves become pests (e.g. Bufo marinus). While experimentation and modeling can provide good grounds for
11 predicting effectiveness, the outcome is often still unknown until it is attempted. Thus, success of biological control agents is difficult to guarantee (Charudattan 2001).
Most weed management practices are undertaken using a combination of physical, mechanical, chemical, and biological control methods in order to be most effective. A biological control agent in combination with other control methods can often stress the exotic making it vulnerable to other control techniques (Charudattan 2001). Management practices for Melaleuca in Florida use herbicides in combination with mechanical removal, together with prescribed flooding and burning (Turner 1997). Similarly, a combination of practices have been successful in reducing the spread of Mimosa in Kakadu National Park in northern Australia (Cook 1996; Paynter 2004).
Yay or Nay: Determining the Benefits and Consequences of Weed Management Despite the increasing number of invasive species worldwide, quantitative data defining the ecological impacts of weeds are rare (Blossey et al. 2001). With little data on the effects of an invasive species, a cost-benefit analysis of weed management is difficult or impossible to perform. Population models have been derived to predict impact for specific weeds based on rate of spread (Bergelson 1993; Woolcock 2000; Fagan 2002; Buckley 2004), but rate of spread is only one of a multitude of characteristics that make a weed invasive and a habitat prone to invasion. There is currently no general equation that encompasses all ecological factors for determining weed management regimes. Thus factors that characterize a weed and its invaded habitat, in this case a wetland, are useful to link it to a management plan. A series of questions must be explored in order to define management practices, which will benefit the wetland community as a whole (Table 2.1).
12 Table 2.1 Weed ecology and effect on habitat: variables for predicting invasiveness and effectiveness of treatments WEED: FACTOR VALUE MEANING Predicted effectiveness of treatments Seed HIGH Species more likely to - Production dominate and quickly LOW Management + planning more effective Seed Dispersal HIGH Habitat promotes dispersal, plus high - dispersion area LOW Habitat does not + promote dispersal Likelihood of HIGH Plant likely to - Maturation dominate vegetation LOW Management + planning more effective Weed Reaction POSITIVE Species thrives on - to Disturbance disturbance. Problems for control techniques such as burning and grazing NEGATIVE Weed becomes + stressed or dies HABITAT: Habitat HIGH Invasive species can - Compatibility become dominant vegetation quickly and easily LOW More difficult for + invasive to dominate; cannot out-compete native vegetation Kweed in habitat HIGH Quasi-threshold high - enough to out- compete natives LOW Quasi-threshold + allows natives to out- compete invasive Domination FAST Problematic species - (derived from SLOW Manageable + seed ecology) Residency INITIAL Best time to act is at + initial outbreak SUSTAINED Invasive may have - become a quasi- natural state for resident fauna * These factors should be weighed with concern to the invasive weed’s ecology and the invaded habitat’s flora. Then benefits and consequences of weed management can be determined.
13 First, what is the ecology of the weed? Certain characteristics make some plants more invasive than others. Understanding how and why certain biological factors promote invasiveness of a species is an important management tool (Mashhadi 2004). Seed dispersal, time needed for establishment, and the similarity between the native habitat and the invaded habitat must be assessed (Turner 1997). Similarity between native and invaded habitat can increase the rate of vegetation domination. Habitat compatibility can be determined if data on species and ecosystem characteristics exist (Rejmanek 2000).
Second, when did the weed invade and how well established is it? If a weed is well established this needs to be considered when determining management practices. When a weed is long established, the native fauna may have come into equilibrium with it, so that it might be considered part of a new ‘native state’. This is the naturalization of the alien species (Richardson 2000; Daehler 2001). The naturalized state is where the community’s inhabitants have come to depend on exotics rather than native species for habitat or resources, and disturbance to this new state by weed treatment or by introduction of native flora may induce more stress than the weed itself. If native species use the invasive weeds to their benefit then the weed should, perhaps, remain. In many cases the natural state has not even been identified, so the treatments may result in the clearing of an area for a species that is less productive to the ecosystem than the actual weed (DiTomaso 1997).
Next, the native vegetation must be considered from two perspectives. First, what is the level of disturbance suffered by the native vegetation due to the weed invasion, and how long has it taken for this to occur? Because loss of native vegetation is the most important cause of reduction in biodiversity (Turner 1997), plant species which would be lost due to widespread invasion of a weed should weigh heavily on the decision for weed management. Biodiversity and ecosystem processes are interactive, and may result in invasive weed-altered ecosystem processes inhibiting native species while favouring certain alien species (D'Antonio 1992). For example, currently there is some debate over purple loosestrife (Lythrum salicaria) and whether or not it is ultimately responsible for the reduction of native flora such as Phalaris arundinacea in New York as native grasses were more successful colonizers than purple loosestrife (Morrison 2002) Following a disturbance, the native grasses establish themselves before purple loosestrife does. This means that efforts put towards weed removal would be ineffective as the decline of native flora is not a consequence of weed presence. The second perspective to be considered is the
14 negative effects of the weed treatment, not the weed itself, on the native flora. Many management practices are non-selective for native versus exotic vegetation (DiTomaso 1997). Again, treatments could cause more negative effects than positive ones or be detrimental to the overall goal by damaging native flora populations.
The fauna of the invaded wetland must be considered with respect to the effects of both the weed and the weed management treatment. It is well known that some weeds decrease animal diversity (DiTomaso 2000; Zedler 2004). For example, Ferdinands (2005) noted a decrease in waterbird diversity due to the invasive para grass (Urochloa mutica) in the Mary River floodplain in the Northern Territory, Australia. However, very little work has been done on the effect of para grass and its removal on other fauna populations. One study found there are at least short-term negative effects associated with weed treatments by showing the herpetofauna populations decreased following burning and grazing (Bower 2006). If the treatment has a negative effect on more species than does the weed, is it wiser to leave the wetland in its present state? This theory has been proposed in a management plan based on land value and the degree of site disturbance (risk of invasion) (Hobbs 1995). While this is a swift assessment, several other factors are pertinent when determining management controls, but the study does recognize a point at which leaving the current state is the most cost-effective, reasonable plan, due to naturalization achieved by alien species. However, if the negative effect of the treatment is short-lived, and populations recover quickly, management practices may be effective in maintaining goals. Many treatments have a tendency to focus on the preservation of a single group such as water fowl (Kushlan 1986; Ferdinands 2005). While the focus group may benefit, other groups may be negatively affected.
Another question to be considered when developing weed management plans is: what economic loss or ecosystem functions will stem from the weed invasion? Wetlands are not only valued for their biodiversity but also for their economic value. They stabilize flood zones and maintain water filtration processes (Salvasen 1990). These services, along with other ecosystem services, have been estimated to save US$33 trillion a year compared with providing these services by mechanical means (Costanza 1997; Daily 1997). Weeds alter basic ecosystem functions that generate these economic values (Vitousek 1990), so economic loss from weed invasion is worth considering when pursuing wetland weed management.
15 Finally, a cost-benefit analysis should be performed. This should be derived from the above factors on a biologically significant scale and then applied to an economical budget. Cost-benefit analyses are common, but tend to favour economic costs and benefits over environmental ones.
All of these questions must be answered and weighed before implementing an effective weed management plan where the benefits outweigh the consequences (Table 2.1).
Success Most weed management plans have had minor successes, but none have been completely victorious against noxious weeds. With the exception of successful biological control, weed management tends to be a high-maintenance activity, requiring frequent repetition of treatments (Hobbs 1992). However, continued persistence in management practices has led to success stories such as in Kakadu National Park in the Northern Territory, Australia with regard to Mimosa pigra (Cook 1996; Paynter 2004). The weed was treated during its initial outbreak, when its population was small, using several different methods including burning and hand pulling individual bushes. The efforts are continued annually and thus far the weed has not spread or developed any large stands throughout the park (Cook 1996; Paynter 2004). Again, the combination of different management treatments can be effective in controlling the spread of weeds.
Conclusion In a good scenario, weed treatment is cost effective with little disturbance to native flora and fauna. A bad scenario occurs when, either by duration of weed residency or when treatment costs outweigh benefits (in biological and/or economical terms), it would be better to leave the weed alone. Unfortunately, the majority of weed management plans lie in the grey area, trying to balance between two species/groups or between economics and environment. Often the effects of either the weed or the treatment can be difficult to assess. Many effects of invasive weeds to fauna may be confined to initial contact, and populations may recover quickly. The same applies to the effects of management practices. Long-term effects are the most pertinent to deciding on weed control techniques and management plans. Control strategies must include all appropriate and effective methods of remediation with a focus on biological control (Charudattan 2001). What is important to realise about weed management is that there is a cause and effect for every part of a
16 community, so a thorough community analysis must be done prior to any decision making process.
17 Chapter 3 Habitat Use by Herpetofauna in a Wetland: Who Lives Where? Abstract While species may be recorded as present in a habitat it does not necessarily follow that this is their preferred or core habitat. This study looked at the abundance of herpetofauna in different habitats in a wetland, dividing it into distinct parts of a whole: woodland; floodplain, and margin. The results show a higher number of herpetofauna in the woodland and suggest the floodplain population is a subset of the woodland population. Dominant vegetation type was not a factor for habitat preference of reptiles or amphibians, however some trends for individual species were determined. The results of this study can potentially be applied to para grass management to buffer losses to floodplain populations incurred by treatments and minimise consequences of treatments by understanding habitat use of species in a wetland.
Introduction Wetlands are maintained and protected for many reasons: preservation of stable flood zones; maintenance of water filtration processes; habitat restoration; and conservation of the unique and diverse species assemblages that they support (Salvasen 1990; Mitsch 1993). They are extremely valuable for maintaining biodiversity (Semlitsch and Bodie 1998). In addition to their biodiversity benefits, there has been a worldwide move to conserve and maintain wetlands for their economic values, which amount to an estimated US$33 trillion per year (in 1997) worldwide in ecosystem services (Costanza 1997; Daily 1997).
Wetland habitats can be divided into two distinct zones: the aquatic habitat or floodplain, which is regularly inundated, and terrestrial habitat which surrounds the submerged area and is strongly influenced by it. There is also an ecotone or margin, defined as the zone 30 meters to either side of the water line (Braithwaite 1988), which accounts for species movement between the aquatic and terrestrial habitats. The importance of the interconnection between these two habitat zones is well documented (Wetzel 1990; Mitsch 1993). An example of this is amphibian use of the floodplain for breeding, followed by their return to terrestrial habitat for other life stages (Duellman 1986). The appeal of the water in a wetland is that fish, which are predators of amphibian eggs, are found in low numbers in many wetlands because wetlands often are completely dry during some parts of the year (Dodd and Cade 1998). Due to the close proximity of water in a wetland, the chances of finding a
18 mate are increased while exposure to prey is decreased. This interconnection between habitats is pertinent when determining management practices. Where conservation has been the objective, the focus tends to be on the floodplain and its benefits for waterfowl (Ferdinands 2005) even though the surrounding woodland may have a higher species abundance and richness for many animal groups (Findlay and Houlahan 1997). One study in wetland in northern Australia found that species richness of herpetofauna was 11-15 times greater in the margin between the woodland and floodplain than in the floodplain itself in both the wet and dry season (Braithwaite 1988). All species were found in the margin but the floodplain fauna were a subset of margin fauna (Braithwaite 1988). While this may be true, species in the margin and woodland still depend on the floodplain for other life functions (Dodd and Cade 1998; Gibbons 2003). Benefits from proximity to water can come in many forms such as food, water availability and breeding grounds.
Species richness of in a wetland is directly related to diversity of available habitat. Species richness typically increases with vegetation cover in habitats at the wetland’s edge (Findlay and Houlahan 1997). Increased forest cover provides structural complexity, offering more refuge from predators, as well as nesting and food resources (Vallan 2002). Different herpetofauna use the available structural complexity to varying degrees. For example, lizards seem to be significantly influenced by structural complexity of habitat whereas snakes are more influenced by food availability (Sexton 1964). Wetlands typically support a higher proportion of amphibian species than reptiles (Friend 1990). Species richness of amphibians, however, is dependent on the amount of water available in the wetland (Dalrymple 1991) and therefore changes seasonally. However, despite changing species richness with season and available vegetation, wetlands still offer a unique assemblage of species due to the diversity of habitat offered (Salvasen 1990).
Wetlands are defined as soils formed under saturated conditions that are drained, undrained, or seasonally wet long enough to support hydrophilic plants (Heimlich 1994). These conditions when paired with tropical or sub-tropical climate are ideal for a plethora of invasive weeds. Invasive weeds may decrease animal and plant diversity (Rea 1999; DiTomaso 2000; Zedler 2004) and are a main focus in conservation and preservation of wetlands. The decreased biodiversity due to weeds is a cause for concern because biodiversity and ecosystem processes are interactive, one result being that ecosystems that have been altered by invasive weeds may lose native species of both flora and fauna and become vulnerable to
19 invasion by other alien species (D'Antonio 1992). Weeds can drive a vicious circle for wetlands, decreasing chances of restoration and management if ecosystem interaction and long-term effects of weeds are not taken into consideration.
Weeds are potent agents of environmental change on local, regional, and global scales (D'Antonio 1992; Turner 1997). Locally, this applies to the Townsville Town Common. Unfortunately, the introduction of para grass Urochloa mutica over thirty years ago has substantially changed the nature of this tropical wetland. Introduced to northern Australia as a fodder plant for cattle, para grass grows rapidly and is able to dominate many native species. It thrives especially in shallow wetlands, where it may grow well into the water, smothering native vegetation, effectively forming a monocultural blanket. Para grass now dominates the vegetation of the Town Common floodplain, and its control and removal have become a significant management issue. These management tactics are monitored for their effect in the floodplain, but do not consider the woodland. While the management actions are being trialed, there is a paucity of information on the woodland/floodplain relationship for the site.
This project aims to determine the dominant habitat of herpetofauna in a wetland. This information will help in enacting weed management techniques by defining which animals are more at risk from management practices due to location in different habitats of a wetland. The Townsville Town Common has been undergoing experimental treatments on the floodplain. Understanding where, i.e. woodland, margin, or floodplain, the herpetofauna are most abundant will help gauge the effect of weed management treatments on this population.
From the knowledge of herpetofauna abundance in the woodland as compared to the floodplain, this project also aims to determine if the woodland populations buffer loss incurred by weed treatments. It has been previously shown that weed treatments have a negative effect on herpetofauna abundance in the floodplain of the Town Common (Bower 2006). If the woodland population is large enough, than these losses are minimal compared with overall population size.
Finally this project aims to determine the role of the areas adjacent to treatments following burning. If it is shown that herpetofauna in the move out of the treated sites to seek refuge, this also helps to buffer losses incurred by the treatments as the population is moving to safety.
20 Aims: • Determine dominant habitat of herpetofauna in a wetland • Determine if woodland populations buffer loss incurred by weed treatments • Determine whether if there is an change in herpetofauna numbers in areas outside of CSIRO/QPWS treatment sites following treatments
Methods Study Site The Townsville Town Common is a tropical wetland located in northeast Queensland. Like most wetlands the Town Common includes a floodplain and adjacent woodland divided by a waterline (Figure 3.1). The waterline bisects an ecotone, so 30 meters to either side of this line is considered margin habitat.
21
Figure 3.1 Location of study site within Townsville Town Common and in comparison to CSIRO/QPWS treatments (Plus 2007)
Vegetation is substantially different between the floodplain and the woodland. The invasive para grass dominates the floodplain. The floodplain is currently undergoing experimental treatments for para grass control (Williams 2005). Para grass does not grow as effectively in the adjacent woodland, but another invasive grass, Guinea grass (Megathyrsus maximum) dominates in some areas. Outside of the floodplain, open woodland with eucalypts and Melaleuca, is the prevalent habitat (Figure 3.2). The wetland was also assigned five types of dominant vegetation for both woodland and floodplain for the purposes of this study: grassed woodland, leaf litter woodland, Guinea grass, para grass and open water. The grassed woodland consists of short prairie grasses with eucalypts and Melalueca. The leaf litter woodland is the same
22 but with a mat of leaf litter opposed to prairie grasses. The Guinea grass and para grass are both invasive grasses dominating the vegetation surrounding the arrays located in these grasses. Finally, open water is in the floodplain, where no dominant vegetation can be identified due to water level. In the dry season, these areas are usually just dried mud.
Figure 3.2 Example of woodland trap and surrounding vegetation
The weather in the Townsville Town Common is typical of tropical North Queensland. Generally the wet warm season lasts from September to March, and the cool dry season from April to August. In 2006, the effects of the wet season lasted well into October. This study occurred over the transition from wet to dry season.
Experimental Design Herpetofauna of the wetland and adjacent woodland were sampled using a combination of pitfall and funnel traps. Three transects were plotted along 200m lines around 100m apart. In each of these transects, six arrays were erected in a line from the woodlands into the floodplain at 30m apart (Figure 3.3). The margin was defined as the area approximately 30m to either side of the waterline. This was difficult to access because the vegetation varied along a defined waterline edge with para grass dominating the floodplain side. The types of traps used depended on habitat, as pitfall traps are ineffective in water. The land arrays used a T- format with a pitfall at the three points and at the cross (Figure 3.4). Four funnel traps were aligned along each arm of the array. Pitfall traps are 20L buckets dug into the ground. The water
23 traps were made from 10m netting, and four minnow traps, two on each end were attached (Figure 3.5). Minnow traps are funnel traps for water. Funnel traps and minnow traps have funnels on either end of a mesh barrel for trapping. The minnow traps were made to float so that the entrance was level with the surface of the water. Trapping was undertaken in three sessions each lasting 14 days and a fourth session lasting only seven days (Table 3.1). Traps were checked twice daily, once in the morning around 6:30am and once in the evening around 4:30pm. Wet sponges were placed in the pitfall traps to keep animals hydrated.
The Town Common is currently undergoing control practices of the invasive weed U. mutica (Williams 2005). While these treatments were not directly related to this project, they were considered a disturbance to the floodplain and the effect on herpetofauna movement was assessed to determine if the woodland is used as a refuge under habitat stress. Differences in species abundance between woodland and floodplain were also used to determine if losses due to treatments on the floodplain are insignificant because of population densities in the woodland.
Table 3.1 Timeline of data collection Event Date Session 1 September 20-October 3, 2006 Session 2 October 6-October 20, 2006 QPWS/CSIRO burning October 30-November 2, 2006 Session 3 November 3-November 17, 2006 Session 4 January 5-January 12, 2007
Woodland Margin Floodplain
30m 30m
Land Arrays Water Arrays
30m
Figure 3.3 Transect with array locations along 200m line between aquatic and terrestrial habitat showing margin at 30m to either side of the water’s edge.
24
Pit trap
10m
Funnel trap
20m
Figure 3.4 Land array with T- format drift fence including pit trap and funnel locations
X 10m X X X Minnow trap
Figure 3.5 Water array with modified floating minnow traps, two on either end, with 10m drift fence
Sampling Methods Reptiles and amphibians were collected in a mark-recapture study to estimate population densities. A mark-recapture study allows for the estimation of population size once the ratio of marked to total animals in the population stays constant (Cox 2002). All individuals trapped had data taken and were marked unique to this experiment.
Snakes were individually marked with a unique scale-clipping, removing half a ventral scale in numerical order from the vent. Snakes were measured (SVL and tail length), weighed, and sexed before being released at their point of capture (recorded by GPS). The majority of snakes and Lialis burtonis (Burton’s legless lizard) were caught by road-cruising at night, following the evening trapping session. Road- cruising consisted of driving back and forth along the main road at 20-40kmh from dusk until late into the night. When a snake was observed, it was caught at the site where it was found. There were at least two observers each session. Because the road is a linear site cutting through the woodland, individuals caught via this method
25 were used for the purpose of building a marked population and not for the primary habitat study.
Lizards were given an individual number via toe clipping. With ventral side up, toes numbered 1-10 from left to right on forelimbs. Toes were numbered 20-100 by 10 (20,30,40,50, etc.) on hind limbs. Lizards were measured (SVL and tail), weighed, and released within three meters of where they were trapped.
Frogs had a single toe removed (entire 2nd toe, right forelimb). They were measured (SUL), weighed, and then released within three meters of capture. Individual numbers for frogs were not required as general population demographics were all that was required for this study. The few turtles (Chelodonia canni) trapped were marked via shell notching in an alphabetical code starting on the first scute to the right of middle on the shell’s edge. Recapture rates were too low among lizards, frogs and turtles to deduce population sizes for different species.
Analysis Trap Efficiency Due to the differences in habitat, i.e. aquatic and terrestrial, different trap types were used to collect data. Funnel traps (minnow traps in water) were used in both the woodland and the floodplain, while only the woodland also included pitfall traps. Trap efficiency, i.e., animals caught per trap, between funnel traps and pitfall traps is a difference in potential source of bias in these data. Levene’s Test of Equality of Error of Variances was run for all data sets, and results were shown where Levene’s was insignificant. In order to determine the efficiency of each trap type, a Two-way ANOVA was used to compare trap type to trapping session as treatments to assess reptile or amphibian abundance in the woodland where both traps were in use. Results determined which trap types would be used in other analyses. Error bars indicate standard error.
Habitat When comparing herpetofauna abundance and species richness in different habitats, i.e. woodland, margin and floodplain, only individuals trapped in funnel traps were used as these were directly comparable. Raw data was used for all analyses except reptile species richness and abundance by habitat and amphibian abundance by habitat, where the data was transformed by square root to fit to a normal distribution. Two-way ANOVAs were used to determine which habitat contained the higher
26 abundance and species richness of both reptiles and amphibians at transect level. Two-way ANOVAs were again used to test significant variation in abundances determine individual species. Error bars indicate standard error.
A principal component analysis was used to describe variation in community structure among sites and habitats. A correlation matrix rather than a covariance matrix was used due to the large differences in the abundance of each species (Table3.2).
The effect of dominant vegetation type on number of reptiles and amphibians trapped and reptile and amphibian species richness was tested using a One-Way ANOVA. Error bars indicate standard error. Relationships between numbers trapped of individual species to dominant vegetation type was tested using a Chi- Square analysis (Preacher 2001). Degrees of freedom for all Chi-Square tests was four. Because different vegetation types surround individual arrays, these data are presented to determine trends for association of individual species with vegetation type.
Weather Daily weather reports including temperature and rainfall for the Townsville were collected from the Bureau of Meteorology. Regressions were applied to the data comparing capture rate (number of reptiles or amphibians trapped daily divided by the total number of traps) and rainfall.
QPWS/CSIRO Para Grass Treatments Because the transects for this study did not intersect CSIRO/QPWS plots (Figure 3.1), the direct effect of the burning treatment could not be analysed. However, using a One-way ANOVA, the distance of arrays from the burn site was compared to the number of herpetofauna trapped in those arrays prior to and following the burn. Only floodplain arrays were used as the woodland is a different habitat. In this analysis, the prescribed burning is considered a disturbance to the floodplain.
Results Over the course of this study, thirty-four herpetofauna species were trapped (Table 3.2). The most commonly captured species during the dry season were Bufo marinus (cane toad), and Opisthodon ornatus (ornate burrowing frog). The most abundant snake in the Town Common was the keelback, Tropidinophis mairii.
27 Table 3.2 List of herpetofauna species trapped on the Town Common. These are species trapped during this study. Snake numbers include those trapped via road cruising, which were not used in analysis of habitat preference. D= Dry Season, YR= Year Round, W=Wet Season Family Scientific Name Common Name # Season trapped
Snakes Boidae Aspidites Black Headed Python 1 D melanocephalus Liasis mackloti Water Python 18 YR Morelia spilota Carpet Python 10 D Colubridae Boiga irregularis Brown Tree Snake 12 D Dendralaphis Common Tree Snake 2 D punctulata Tropidonophis mairii Keelback 110 YR Elapidae Demansia vesitgiata Black Whip 13 YR Cryptophis nigrostriatus Black-stripe Snake 2 D Furina barnardi Yellow Nape 1 YR Furina ornata Orange Nape 1 YR Pseudonaja textilis Eastern Brown 2 D Lizards Agamidae Diporiphora australis Tommy Round-head 3 D Gekonidae Gehyra dubia Tree Dtella 16 D Heternotia binoei Byone’s Gecko 9 D Pygopodidae Delma tincta 4 D Lialis burtonis Burton’s Legless Lizard 3 D Scincidae Carlia pectoralis 75 D Cryptoblepharus 100 D virgatus Ctenotus robustus 11 D Glaphyromorphus 5 D punctulatus Morethia taeniopleura Fire-tailed skink 24 D Tiliqua scincoides Eastern Blue-tongued Lizard 1 D Frogs Bufonidae Bufo marinus Cane Toad 485 YR Hylidae Cyclorana alboguttata Green-striped Frog 4678 W Litoria bicolor Northern Dwarf Tree Frog 1 D Litoria caerulea Common Green Tree Frog 1 D Litoria fallax Eastern Dwarf Tree Frog 14 D Litoria nasuta Rocket Frog 84 D Litoria rothi Northern Laughing Tree Frog 2 D Litoria rubella Desert Tree Frog 14 D Myobatrachidae Limnodynastes Marbled Frog 67 D convexiusculus Limnodynastes Spotted Marsh Frog 153 YR tasmaniensis Opisthodon ornatus Ornate Burrowing Frog 287 YR Turtles Chelidae Chelodonia canni Cann’s Snake-necked Turtle 5 D *Season trapped: YR= year round, D=dry season. W=wet season
28 Trap Efficiency Pitfall traps proved to be more efficient in catching herpetofauna than funnel traps (ANOVA Number of reptiles trapped: p=0.009, Number of amphibians trapped: p<0.001) (Table 3.3, Figure 3.6). Number of reptiles trapped remained the same between sessions (Figure 3.6) indicating trapping session is not a factor with concern to trapping efficiency. Reptile species richness was the same between trap types and sessions. Pitfall traps again proved more efficient when trapping amphibians (Table 3.3). Trapping efficiency was different between sessions, probably due to season and the drying of the wetland. It is important to note that while pitfall traps are a very effective method, there are limitations in catch rate. As with any trapping method, the level of activity and the size of the population will affect the success of the trap.
As pitfall traps were successful in trapping high numbers of both reptiles and amphibians, this data helped to build a marked population of herpetofauna. No species were trapped in pitfall traps and not in funnel traps. All other analyses only used data from funnel traps for a valid comparison between floodplain and woodland habitats. Despite the lower numbers of amphibians trapped in funnel traps, the numbers compare to the amount of reptiles trapped in funnel traps and remain similar across sessions (Figure 3.6).
Table 3.3 ANOVA results for untransformed trap efficiency data showing a significant difference between trap efficiency ANOVA df Mean Sq. F Sig. Session 2 0.517 0.069 0.934 Trap 1 56.067 7.450 0.009 Number trapped Session x 2 0.617 0.082 0.921 Trap Reptiles Session 2 0.600 0.401 0.671 Species Trap 1 4.817 3.223 0.078 Richness Session x 2 0.267 0.178 0.837 Trap Session 2 1082.830 35.940 <0.001 Trap 1 12041.667 399.662 <0.001 Number trapped Session x 2 1070.617 35.534 <0.001 Trap Amphibians Session 2 3.264 3.877 0.027 Species Trap 1 132.012 156.679 <0.001 Richness Session x 2 0.467 0.554 0.578 Trap
29
Pitfall traps Funnel traps
Pitfall traps Funnel traps
Figure 3.6 Trap efficiency for herpetofauna by number of reptiles and amphibians captured during trapping session in the Town Common.
30 Habitat Table 3.4 ANOVA results for habitat analysis. Reptile abundance, reptile species richness, and amphibian abundance data were square root transformed. Mean ANOVA df F Sig. Sq. Transect 2 2.636 7.760 0.002 Abundance Habitat 2 45.785 134.766 <0.001 Transect x Habitat 4 0.749 2.204 0.095 Reptile Transect 2 0.197 1.250 0.302 Species Richness Habitat 2 15.083 95.714 <0.001 Transect x Habitat 4 0.351 2.228 0.092 Transect 2 0.355 0.176 0.807 Abundance Habitat 2 19.930 9.878 0.267 Transect x Habitat 4 1.340 0.664 0.346 Amphibian Transect 2 0.290 0.484 0.643 Species Richness Habitat 2 6.292 10.486 0.016 Transect x Habitat 4 3.225 5.375 0.047
Reptiles Overall, number of reptiles trapped and reptile species richness were higher in the woodland than in the floodplain (Figure 3.7), but both were lower in the woodland in the third transect (Table 3.4). Transect 3 differed from the other two (Tukey HSD: 1 vs. 3, p= 0.009; 2 vs. 3, p= 0.002), perhaps because of a difference in dominant vegetation as later discussed. Data for number of reptiles trapped and reptile species richness by habitat were transformed by square root to fit a normal distribution.
31 Woodland X Margin Floodplain
Woodland X Margin Floodplain
Figure 3.7 Number of reptiles trapped and reptile species richness by habitat in the Town Common. Both data sets were square root transformed to best fit a line of normality.
32
Reptiles in the Town Common were trapped in higher numbers in the woodland than the floodplain. Numbers and species composition dwindled from woodland to floodplain (Figure 3.7). Species composition changes from woodland to floodplain (Figure3.8). PC1 represents the change in habitat or in this case, the margin, as the majority of woodland plots aggregated to one side while the margin and floodplain sites clustered to the other end of the axis (Figure 3.8). However, there is still some overlap in species composition between habitat (Figure 3.8).
Prinicpal Component Analysis for Reptile Community Composition in Three Habitats
5 4 3 2 Floodplain 1 Margin 0 Woodland -2 -1 -1 0 1 2 3 -2 PC2 (23.123%) -3 -4 PC1 (75.017%)
Figure 3.8 Principal Component Analysis of reptile community composition across three habitats. Each marker represents 1 of 3 transects during each of 4 trapping sessions. At the red circle, the floodplain sites are overlapping due to similar species composition.
Although the majority of reptiles were associated with the woodland, there were a few which influenced the outcome of the reptile community PCA. As seen in Figure 3.9, the majority of abundant species associate with PC1.
33 Eigen Vectors for Reptile Community Composition in Three Habitats
2 1.5 Cryptoblepharus virgatus 1 Carlia pectoralis 0.5 Morethia taeinopleura 0 Demansia vestigiata -1 0 1 2 3 4 Heteronotia binoei -0.5 Tropidinophis mairii PC2 (23.123%) -1 -1.5 PC1 (75.017%)
Figure 3.9 Eigen vectors for reptile community PCA comparing individual species that effect the outcome of the PCA.
Species such as Carlia pectorails and Cryptoblepharus virgatus that were only found in the woodland drove the results of the PCA (Figure 3.8) towards the woodland side of PC1. PC2 shows no obvious association (23.123%). Nevertheless, it could be associated with moisture as C. virgatus were trapped in relatively moist locations while the rest were found in the drier woodland habitats.
Amphibians Overall number of amphibians trapped was similar between habitats and between transects (Table 3.4, Figure 3.10). No significant difference could be found between habitats for amphibians, but individual species responded very differently to the various habitats, making the floodplain and margin community dramatically different from the woodland community (Table 3.4, Figure 3.10, Figure 3.11).
34 Woodland X Margin Floodplain
Woodland X Margin Floodplain
Figure 3.10 Number of amphibians trapped and amphibian species richness by habitat in each transect. Number of amphibians data has been square root transformed to best fit normality.
35
Principal Component Analysis for Amphibian Community Composition in Three Habitats
4
3
2
1
0 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 Floodplain
PC2 (6.759%) -1 Margin -2 Woodland
-3 PC1 (88.229%)
Figure 3.11 Principal Component Analysis for amphibian community composition in the Town Common. Each marker represent 1 transect 3 at each of 4 sessions. The red circle indicates where all the floodplain sites are overlapping due to similar species composition.
PC1 most likely correlates with habitat as the majority of floodplain and margin sites aggregate on one side of the axis (Figure 3.11).
Eigen Vectors for Amphibian Community Composition in Three Habitats
6 5 4 Bufo marinus 3 Opisthodon ornatus 2 Limnodynastes tasmaniensis 1 Litoria fallax 0 Litoria rubella -5 -1 0 5 10 15 20
PC2 (6.759%) -2 Litoria nasuta -3 -4 PC1 (88.229%)
Figure 3.12 Eigen vectors for amphibian community PCA showing species that drive trends in the PCA.
Bufo marinus and Opisthodon ornatus heavily influence the PCA as they are the most abundant amphibians and are associated with the woodland side of PC1 (Figure 3.12). Litoria fallax borders on being a floodplain species, while the others are more associated with the margin when compared to the PCA (Figure 3.11).
36 Dominant Vegetation Preference Each array was dominated by different vegetation. This vegetation difference may also account for significant differences between transects. Guinea grass, leaf litter woodland and grassed woodland were the dominant vegetation types in the woodland, while para grass and open water were the floodplain vegetation types. Leaflitter woodland was significantly different to both open water and para grass (Student’s t-test=0.042 and 0.038, respectively). Number of reptiles trapped was significantly effected by vegetation type (ANOVA: F4,13=3.205, p=0.049). Species richness was higher in woodland sites and low in floodplain vegetation for reptiles (Figure 3.13). Vegetation type did not effect number of amphibians trapped (ANOVA:
F4,13=1.718, p=0.206). Amphibian species richness remained consistent across vegetation types (Figure 3.13).
Dominant Vegetation Preference by Abundance Dominant Vegetation Preference by Species Richness 70 Reptile Abundance 12 Reptile Species 60 Amphibian Abundance Richness 10 Amphibian 50 Species Richness
8 40
6 30 Abundance
4 20 Sepcies Richness
10 2
0 0 Guinea grass Grassed Leaf litter Open water Para grass Guinea grass Grassed Leaf litter Open water Para grass woodland woodland woodland woodland Dominant Vegetation Dominant Vegetation
Figure 3.13 Dominant vegetation for both reptile and amphibian species richness and numbers trapped. Amphibians are more associated with floodplain vegetation types, i.e. para grass, while reptiles remain consistent with woodland habitat types.
Dominant Vegetation Preference for Individual Reptile Species
18
16
14 Carlia pectoralis
12 Cryptoblepharus virgatus 10 Demansia vestigiata
8 Gehyra dubia Abundance
6 Heteronotia binoei
Tropidinophis mairii 4
2
0 Guinea grass Grassed woodland Leaflitter woodland Open water Para grass Dominant Vegetation
Figure 3.14 Dominant vegetation associated with individual reptile species. Tropidonophis mairii has the most varied habitat, while the majority of others are associated with woodland vegetation.
37 Some individual reptile species showed preferences for a certain vegetation type (Figure 3.14). Carlia pectoralis abundance was related to vegetation type (Chi square: x2 = 29.43, p< 0.001). Cryptoblepharus virgatus and Gehyra dubia also associated with vegetation type (Chi square: x2 = 53.667, p< 0.001, x2 = 22.667, p< 0.001, resp.), seeming to prefer leaflitter woodland. Demansia vestigiata abundance did not vary significantly (Chi square: x2 = 9, p= 0.061), but D. vestigiata appeared to prefer grasses associated with the woodland. Abundance of Heteronotia binoei and Tropidonophis mairii were unrelated to vegetation types (Chi square: x2 = 4, p= 0.406, x2 = 2, p= 0.736, resp.). T. mairii was found consistently across all dominant vegetation types.
Dominant Vegetation Preference for Individual Amphibian Species
45
40
35 Bufo marinus 30 Limnodynastes 25 tasmaniensis Litoria fallax
20 Litoria nasuta Abundance
15 Opisthodon ornatus
10
5
0 Guinea grass Grassed Leaf litter Open water Para grass woodland woodland Dominant Vegetatio
Figure 3.15 Dominant vegetation associated with individual amphibian species. Litoria nasuta strongly associates with para grass where as Bufo marinus prefers woodland vegetation.
Some individual amphibian species showed preference for a particular vegetation type (Figure 3.15): Bufo marinus associated primarily with woodland habitats (Chi square: x2= 23.0, p < 0.001); Limnodynates tasmaniensis and Litoria fallax, associated with open water and para grass (Chi square: x2= 28.923, p< 0.001; x2= 13.778, p= 0.008, resp.); Litoria nasuta was the most widespread in its habitat use and does not significantly associate with a single vegetation type (Chi square: x2= 9.0, p= 0.061).
38 Effect of Weather
Rainfall (mm) in Town Common during trapping
6
5
4
3
2 Rainfall (mm)
1
0
5/1/077/1/079/1/07 20/9/0622/9/0624/9/0626/9/0628/9/0630/9/062/10/068/10/06 4/11/066/11/068/11/06 11/1/07 10/10/0612/10/0614/10/0616/10/0618/10/0620/10/06 10/11/0612/11/0614/11/0616/11/06 Date
Figure 3.16 Rainfall in millimeters in the Town Common by day of trapping
Amphibian Capture Rates in Relation to Rainfall
1.6
1.4
1.2
R2 = 0.6364
1
0.8
0.6 Amphibian Capture Rate
0.4
0.2
0 0 1 2 3 4 5 6 Rainfall (mm)
Reptile Capture Rate in Relation to Rainfall
0.14
0.12
0.1
0.08
0.06 Reptile Capture Rate 0.04
R2 = 0.0512 0.02
0 0 1 2 3 4 5 6 Rainfall (mm) Figure 3.17 Rainfall to number of herpetofauna trapped in the Townsville Town Common during this study. Trendline with regression is shown.
Rainfall varied throughout the study (Figure 3.16). With respect to the weather, amphibian capture rates increased with rainfall (R2=0.6346) (Figure 3.17). As seen in Figure 3.17, reptile abundance in unrelated to rainfall (R2=0.0512).
39
Pre v. Post Burn Para grass treatments were ongoing during this study. Of particular note, was the burning to the floodplain. While no transects intersected directly with CSIRO/QPWS treatment plots, the affect of the treatment on the herpetofauna population was assessed in the form of a disturbance to the floodplain. As arrays for this study were near treatment plots, distances of these arrays from treated plots were measured. A comparison between numbers trapped and distance from treated plots would show possible movement of herpetofauna away from treatments. The relationship between number of reptiles trapped pre and post burn to distance was insignificant (ANOVA:
F3,11= 1.894, p= 0.189). The interaction between trapping session and distance from the burn sites was insignificant for reptiles and amphibians (Table 3.5).
Table 3.5 Interaction between session and distance for reptile and amphibian captures Parameter Estimates t- Ratio Significance Interaction: Reptiles -1.52 0.1563 Interaction: Amphibians -0.15 0.8848
Discussion Habitat Preferences of Reptiles and Amphibians Thirty-six species of reptiles and amphibians were trapped in the Town Common (Table 3.2) in this study. These species range from one of the largest skinks, Tiliqua scincoides, to one of the most deadly snakes in the world, Pseudonaja textilis (Broad 1979). Other species are present in the Town Common, having been observed though not captured in the systematic survey. These include sand goannas (Varanus gouldii), frilled lizards (Chlamydosaurus kingii), and different species of frogs (Crinia deserticola) and skinks (Lampropholis delicata). These species were either present during different seasons or trapping methods used were not suitable for them. This study was conducted at the beginning of the dry season, and thus species collected are species that are generally most active or abundant in the dry season. Wet season sampling is different in both species richness and abundance (Sexton 1964; Dalrymple 1991; Vonesh 2001). However, it is probably more appropriate to say that species richness and abundance would change due to conditions as wet and dry season are rarely clearly defined.
Species richness is important to this study because management plans for the invasive para grass must account for all species that would be affected by such management. In the case of the Town Common, para grass management efforts are
40 carried out for the purpose of encouraging the return of waterfowl to the wetland (Grice 2006). Population surveys of other animals are conducted only prior to burning each year- over a one-week period. This does not fully measure the impact of the treatments on wetland biodiversity. These treatments have been shown to have negative effects on herpetofauna (Bower 2006). For this reason, these species should continue to be monitored. This study presented an effective way of monitoring incurred consequences of weed treatments through the study of herpetofauna species richness and abundance in different habitats of the Town Common.
Reptile species richness and abundance was higher in the woodland than in the floodplain or margin. Only Tropidonophis mairii and Demansia vestigiata were trapped near or in the floodplain. Reptiles tend to move out of flooded areas as most are not sufficiently aquatic to remain in water for several months (Roe 2003). Thus the preference of most reptiles for the woodland over the floodplain at least is consistent with their ecology and physiology. The reptiles that were found in the floodplain were more associated with the margin, which suggests they use the floodplain for some activity such as foraging, but likely inhabit the woodland. A single lizard would appear to be an exception. Lamprophilis delicata, was not captured during this study but has previously been trapped in the Town Common in a floodplain study (Bower 2006).
Amphibian species richness and abundance were similar between the woodland and floodplain. This is to be expected since amphibians utilize both habitats (Semlitsch and Bodie 2003). While amphibian abundance and species richness were similar in the two habitats, community composition was very different (Figure 3.11). Burrowing frogs, e.g. Opisthodon ornatus, and Bufo marinus were more abundant in the woodland, which is characteristic of their life functions as they are terrestrial (Williams 1998; Seebacher 1999) . Litoria fallax was primarily found in the floodplain, perhaps responding to the structural complexity and the concealment provided by the para grass. Aside from L. fallax and L. nasuta, most species of amphibians trapped on the floodplain side of the para grass edge were trapped in the margin and were species more associated with the woodland. They, like the reptiles, are probably entering the floodplain edge to forage or to find water.
41 Dominant Vegetation Preference As seen from earlier data, reptiles tended to be restricted to the woodland, while amphibians were more widespread. In general, amphibians dominated the floodplain vegetation, which was predominantly para grass. However, those found the woodland vegetation tended not to be found in Guinea grass.
Reptiles and amphibians used both invasive grasses, but in most cases they were not the preferred habitat. For the most part, this does not support the theory of naturalization of either grass. Naturalization is when an invasive species is used as if it were a native species (Richardson 2000; Daehler 2001).
Where naturalization is a problem is most apparent is with concern to the para grass. Two species, Litoria nasuta and Litoria fallax, are associated with para grass (Figure 3.15). L. nasuta appeared to prefer para grass, although since para grass is the dominant vegetation of the floodplain L. nasuta may be more associated with the floodplain (i.e. water) than para grass per se. However, L. fallax was also associated with para grass. The para grass offers a structurally complex environment that is approximately the same colour as the skin of L. fallax. This may make the para grass a more appealing habitat as it offers camouflage. These two noted trends in para grass usage present serious issues for conservation efforts with concerns for costs and benefits of management efforts. If, indeed, para grass has reached a naturalized state, attempt to remove it could harm biodiversity of frogs.
Trap Efficiency While pitfall traps were far more efficient in this study, they are not the best technique for trapping larger animals. They were helpful in building marked populations for further study. Pitfall traps are inadequate for capturing larger reptiles, which are able to escape from the pits. Funnels, on the other hand, are more useful for trapping larger reptiles, such as snakes (Greenberg 1994; Enge 2001).
When pursuing further studies across both aquatic and terrestrial habitats, trap types should change with season. Thus, during dry periods, pitfalls could be used both in the floodplain and in the woodland, and during the wet season, funnels traps alone would be used. A similar trapping method should be put in place during the dry season in both the flood plain and the woodland, in order to establish a baseline for comparison. It is difficult to define differences between seasons, as there is no set
42 point at which the season changes. Ideally the study would focus more on seasonal change as opposed to habitat.
Effects of Weather As stated earlier, season plays a significant role in species composition and species activity. Overall humidity seems the biggest factor in successful trapping. What changed was the abundance of animals trapped. After a rainfall, numbers peaked for all herpetofauna, especially amphibians (Figure 3.17). During periods of drought, numbers were low. Because amphibians are caught in higher numbers following a heavy rain, it can be concluded that rainfall is directly related to their behaviour (Martof 1953; Galatti 1992; Parris 1998). This change in behaviour can skew data as amphibians may be less inclined to move, depending on species. They may seek out areas of less water, i.e. the woodland over the floodplain, or vice versa. Weather, especially rainfall, is another factor which should be considered in interpretation of data.
The Effects of the Para Grass Treatments Burning and grazing were carried out during the sampling period for this project. Since the sites were not in the treatment paddocks the effect of the para grass treatments was not significant. These treatments are considered a disturbance to the floodplain and analysed in that manner.
Trapping sessions were held one week before the burnings and immediately following them. Herpetofauna numbers were expected to increase outside of the fire suggesting escape from the burning. However, the distance of the arrays from the fire showed no significant change for reptiles and amphibians (Table 3.5). This means that the fire is not causing herpetofauna to move in the direction of the arrays (towards the woodland) to seek refuge from the fire. This could be for two reasons: herpetofauna are escaping the fire in other directions or they are dying in the fires. This is supported by previous data that showed the burning has a negative effect on reptiles and amphibians (Bower 2006).
Value for Management of Para Grass Habitat management of wetlands often focuses on waterfowl, as is the case for the Town Common. The benefits of the para grass treatments for waterfowl are undeniable (Kushlan 1986). Populations have increased due to the increased availability of open water for nesting sites, once almost completely overtaken by para
43 grass (Valentine 2007). However, despite these obvious benefits, the effects of the treatments were not studied for other animal groups. Shifting focus to the herpetofauna has previously shown that the treatments have a negative effect on both reptiles and amphibians in the floodplain (Bower 2006). While the floodplain leaves reptiles and amphibians vulnerable to the para grass treatments, the woodland is free from the direct effects of these treatments. Since the majority of herpetofauna populations are higher in the woodland than in the floodplain, these numbers mitigate the numbers lost as consequence of burning and grazing.
Concerns Caused by Treatments From prior research (Bower 2006), the negative effects of the para grass treatment on herpetofauna treatments are well documented. Burning and grazing are of significant concern with regards to Lamprophilis delicata. This is a ground-dwelling skink, which was observed swimming in the floodplain during the wet season. This skink is found almost solely in the floodplain associated with para grass. It had no increase in numbers in the woodland following the treatments or in general. If an increase of either of these treatments were to be put in place, a notable decline in population would be expected. Despite the immediate population decline, numbers would likely rebound quickly as this skink is not a rare species in North Queensland.
Burning has a more direct negative effect on amphibians (Bower 2006). This study found that a majority of Litoria nasuta, Litoria fallax, and Limnodynastes tasmaniensis inhabit the floodplain. As discussed earlier, their life history may account for their whereabouts under different conditions, but during the time of the burning, the majority of individuals were found in the floodplain. Thus if a mass burning the wetland were to take place it could have a damaging effect on the populations L. nasuta, L. fallax, and L. tasmaniensis. It might be expected that these species would speak refuge during a fire. However, trapping sessions were held the day following the burning with no significant rise in these populations in the woodland. Traps were located along the margin and in the woodland, so individuals moving out of the floodplain would likely have been trapped. Continued trapping showed no increase in herpetofauna in the woodland. This implies that they are not leaving the floodplain despite the impact of fire, and thus they still are “at-risk” species with respect to fire treatment.
None of these species are rare in the Australian herpetofauna. However, in the wetland ecosystem, their population size would be negatively impacted by these
44 treatments at least as an immediate consequence even if not a permanent one. Special note of the aforementioned species should be taken when pursuing para grass treatment in a wetland.
Conclusion This experiment shows the importance of the habitat surrounding a floodplain in a wetland, and many species that contribute to wetland biodiversity actually depend predominantly terrestrial habitats that fringe the wetland and the ecotone between the terrestrial and aquatic habitats. A comprehensive knowledge of a wetland ecosystem is essential with respect to any management, conservation, and/or restoration efforts planned for that particular wetland.
This study, at the Townsville Town Common, provides information by which a management approach could be designed which would help alleviate the negative effects caused by control of para grass overtaking many Australian wetlands. In addition one skink, Lampropholis delicata was identified as a floodplain associated species, highlighting the species as “at-risk” for any management regimes put into practice in the Town Common.
45
Chapter 4 Living in an Invaded Wetland: the Natural History of Tropidonophis mairii in the Townsville Town Common
Abstract Change in an environment can induce a reaction from its inhabitants; each species will react differently to the change happening in their environment. This study investigated the way keelback snakes (Tropidonophis mairii) use an altered environment. These changes come in two forms for this study, the first being invasive species, which can cause a disturbance to the natural environment. The exotics, which dominate the Town Common, are para grass (Urochloa mutica) and cane toads (Bufo marinus). These have become part of the ecosystem in the Town Common, as both populations are large. The second type of disturbance to the area is the treatments imposed by managers in attempting to reduce para grass. The effects of these treatments on keelbacks were studied during the wet season. Keelback abundance increased with the density of para grass, showing that keelbacks are using the invader for their benefit and the treatments might have negative effects on them. With respect to cane toads, despite their high availability and the ability of keelbacks to consume them, cane toads are non-preferred prey for keelbacks.
Introduction Invasive species and disturbance induce responses from natives by altering the environment (Phillips 2004). Invasive species can have negative effects on natives (e.g. by destroying habitat) or positive effects (by providing new resources), or have neutral effects (by exactly replacing functions that were provided by native species displaced by the invasive).
Invasives can alter the way native species (both flora and fauna) function including by competitive exclusion, niche displacement and predation; these effects can be severe enough to lead to extinction (Mooney 2001). These negative effects result in behavioral responses from native species to increase their chances of survival. These adaptive changes may diminish the impact of invaders and promote coexistence between natives and exotics (Strauss 2006). However, if these responses are not made in due time and/or no adaptation to the novel community or species is made, restrictions on the ecological range of natives can occur, resulting in extinction of native species affected by invasion (Vermeij 1996).
46
Disturbance is another community stressor, which can have both negative and positive effects on the community. Common disturbances are fire, grazing, and human interference. While some disturbance can help to increase biodiversity (Ward 1999) or reduce the probability of greater natural disturbances such as prescribed burning to reduce fuel for wildfire (Pollett 2002), it can also make an area more prone to invasion (Hobbs 1992).
Some areas are especially susceptible to environmental change. For example, wetlands attract many species, both native and exotic, due to the availability of water (Fox 1986). The Townsville Town Common is a wetland that is overrun with invasive flora and fauna species. Para grass (Urochloa mutica) dominates the floodplain, while Guinea grass (Megathyrsus maximum) has invaded the surrounding woodland. Cane toads (Bufo marinus) dominate the herpetofauna. Feral pigs (Sus scrofa) and feral cats (Felis domesticus) have also been observed (Grice 2006).
One of these species, para grass, is currently being subjected to treatment practices to manage its spread (Williams 2005). These treatments, while having an effect on the para grass, also create a disturbance. The principal treatments being used are prescribed burning and cattle grazing. These have been shown to have a negative effect on the herpetofauna in the area (Bower 2006). However, these effects during the wet season have yet to be studied in relation to snakes, the most significant native predators on the Town Common.
Predators are important to ecosystems so examining the effect of disturbance (natural or anthropogenic) on them is beneficial to understanding the community. The dominant predator among the herpetofauna in the Town Common is Tropidonophis mairii, (the keelback snake). Snakes can use disturbed areas to their benefit whether it be to prey on food items or avoid predation (Burger 1989; Pearson 2005). This study looks at variation in the abundance of keelbacks in different habitat types caused by the para grass treatments in the floodplain during the wet season.
Keelbacks feed primarily on anurans (Shine 1991). In the Town Common the dominant anuran species is the toxic Bufo marinus. While keelbacks are known to be able to consume cane toads (Covacevich 1975; Shine 1991; Phillips 2003; Phillips 2004), this study examines to what extent keelbacks consume cane toads and whether or not that consumption is in proportion to their abundance.
47
This project aims to determine how keelbacks use the para grass treated areas during the wet season. While herpetofauna abundance in these plots have been studied during the dry season (Bower 2006), no study has been aimed at the use of treated areas during the wet season. Because a change in vegetation density occurs due to the treatments, the area should be studied year round. Knowledge of habitat use can be applied to potential management treatments to ensure the benefits of the treatments outweigh the consequences.
The second aim of this study is to determine general floodplain use by keelbacks during the wet season. This would be indicative of water depth as this is the main change from wet to dry season in the floodplain. This again applies to habitat use as a variable, which can effect potential management regimes. Treatments can potentially open up clogged water areas, creating a deeper pool, knowing the preferred depth for keelback activity can once again impact the balance between benefits and consequences of treatments.
The final aim of the project is to determine feeding habits of keelbacks with concern to cane toads. The ability of keelbacks to consume cane toads is often confused with cane toads being a common part of keelback diet. When prey consumed is compared with prey availability, the preference for or against cane toads can be determined. This information can be applied to management of cane toads to establish whether or not keelbacks offer a viable biological control for cane toads.
Aims: • Determine how keelbacks use the para grass treated areas • Determine floodplain use by keelbacks in wet season • Determine the feeding habits of keelbakcs with concern to cane toads
Methods Experimental Design Keelbacks were trapped using three different methods: road cruising, land trap arrays, and water trap arrays. Road cruising consisted of driving back and forth along the main road of the Town Common at night at 20-40kph with at least two observers. Due to the location of the road, through multiple habitats, snakes trapped in this method were used to build a marked population and assess species present in the Town Common (Table 3.2). Snakes trapped in land arrays from a previous
48 experiment (Chapter 2) are included in the assessment of population dynamics. During the wet season, use of the floodplain by keelbacks was determined using four floating minnow traps attached to 10m drift fences. Two fences were placed in 12 plots allocated by CSIRO/QPWS for an ongoing experiment for control of the invasive para grass. Each plot was 200m by 300m with different para grass treatments. Three of these were burned, three were grazed, three were treated with a combination of burning and grazing, and three control plots were not subjected to either treatment. The last treatment happened at the October 30-November 2 burning and the cattle remained until the wet season around the end of January. Floodplain trapping occurred from February 9-March 2, 2007, during the wet season. Amphibians were trapped using the same methods in order to determine relative abundance of prey for keelbacks (Introduction, Figure I.5).
Each plot had differences in vegetation density and water depth. Para grass density was determined using a 1m2 quadrat. The quadrat was laid randomly at both ends of each array in each CSIRO/QPWS paddock and photos taken. Four pictures were taken at each array and grids applied to the pictures using Adobe Photoshop to determine percentage of para grass coverage. Percentage was determined by counting the total number of grid blocks with para grass present divided by the total number of grid blocks per quadrat. Average percentage from these photographs was used to estimate para grass coverage by array. Water depth was measured and stratified into ranges for ease of analysis.
Snakes were individually marked with a unique scale-clipping, removing half a scale in numerical order from the vent. Snakes were sexed, measured (SVL tail, head width and length), weighed, and then released within 3m of capture. While this experiment was intended to be a capture-mark-recapture study, too few snakes were recaptured to estimate population size. However, information about sex and age was assessed.
If the snakes had gut contents (indicated by an apparent lump in body), the mass was gently palpated towards the throat, which encouraged regurgitation. If regurgitation was successful, gut contents were examined and identified where possible.
49 Analysis A Chi Square test (Preacher 2001) was run to determine if a relationship existed between number of snakes trapped and treatment, percentage para grass cover, and water depth. Three degrees of freedom were measured for these Chi-Square tests. Only 19 snakes were trapped during the wet season, so a nonparametric test was used because of violation of the assumption of normal distribution.
A Chi Square test was also run to determine the ratio of male to female snakes, while a t-test was performed to determine size difference by sex.
Using Ivlev’s (1977) Electivity Index, diet selectivity was established for keelbacks in the Town Common. The equation for the electivity index is: E*= (ri – pi) / (ri + pi) Where ri is the percent of food class i in the diet and pi is the percent of food class i in the field. Basically, ri is the food consumed, while pi is the food available. This was applied to frogs versus toads and to individual species to determine selective feeding habits. The E*value is either greater than 0, meaning the prey item is selected for, or it is less than 0 meaning it is avoided.
Results Population Size and Composition
Normal Distribution for Snake Snout-Vent Length
0.008 565.96 0.007
0.006 534 0.005 Male 0.004 Female 0.003
Normal Distribution 0.002
0.001
0 0 100 200 300 400 500 600 700 800 Snout-Vent Length (mm)
Normal Distribution for Snake Mass
0.025 63.243
0.02
72.81 0.015 Male Female 0.01 Normal Distribution 0.005
0 0 20 40 60 80 100 120 140 160 Mass (g) Figure 4.1, 4.2 Population composition of keelbacks in Town Common. The graphs are of normal distribution for male and female body size showing females to be larger.
50 A total of 192 keelbacks were marked: 24 juveniles and 168 adults. Snakes were considered juveniles at less than 30cm in length (Brown 2004). There were 171 snakes where sex was identified; of these, 101 were male and 70 were female (Chi Square Test x2= 5.62, p = 0.178, df=1), seven of which were determined to be gravid. Females were significantly larger than males in both weight (t-test: t=2.81, p= 0.006) and length (t-test: t= 2.99, p= .003). Only four snakes were recaptured.
Habitat use Very few keelbacks (19 over 21 days) were trapped in the floodplain during the wet season. However, some trends could be determined despite these low numbers.
Number of Snakes trapped by Treatment
14
12
10
8
6
4
2 Number of snakes trapped
0 Burned Control Grazed Burned and Grazed Treatment
Figure 4.3 Number of snakes trapped to treatment. This graph shows how many individual snakes were trapped in a treatment type.
Type of treatment has an effect on the number of snakes trapped (Chi Square Test x2= 17.003, p < 0.001). More keelbacks were found in control and burned plots than grazed and the combination of grazed and burned plots.
51 Number of Snakes trapped to % Para grass Coverage
12
10
8
6
4
2 Number of snakes trapped
0 0-25 26-50 51-75 76-100 % Para grass Coverage
Figure 4.4 Number of keelbacks trapped in relation to percent para grass coverage
The number of snakes trapped was directly related to the density of vegetation (Chi square test: x2= 20.79, p < 0.001). The density of the grass was determined by the para grass treatment for that area.
Number of Snakes Trapped to Water Depth
9 8
7 6
5 4
3 2
Number of snakes trapped 1 0 0-10mm 11-25mm 26-40mm 41-55mm Water level (mm)
Figure 4.5 The figure shows the number of snakes trapped in different ranges of water depth.
Snake activity was unrelated to water level (Chi Square test: x2= 6.89, p= 0.075). While it is statistically unrelated, there seems to be a trend showing that keelback presence at certain water depths, especially on the lower end, is minimal. This could be related to the difficulty of moving through shallow water depths within vegetated areas.
52 Change in species composition and abundance in the floodplain
Bufo marinus Feb
Limnodynates tasmaniensis Jan Opisthodon ornatus
Nov Litoria nasuta Months
Oct Litoria fallax
Sept Cyclorana alboguttata
Species
Figure 4.6 Species composition and abundance changes over time. Using relative abundance, species were graphed by monthly trapping efforts. Each ball represents the abundance of a species during each month.
Species composition in the floodplain changes with season. In the 21 days of field work on the floodplain in February, 4768 C. alboguttata froglets were trapped compared to a total of four individuals from two other species. Due to this massive increase in C. alboguttata numbers in the wet season, the abundance of this species is compared to the abundance of snakes by treatment (Figure 4.7) as C. alboguttata are available prey items.
C. alboguttata and keelback comparative abundance by treatment
70
60
50 % C. alboguttata abundance %Snake Abundance 40
30
% Abundance 20
10
0 Burn Control Grazed Burned and Grazed Para Grass Treatment
Figure 4.7 C. alboguttata and keelback comparative abundance by treatment in the floodplain. Abundance is by number trapped in treatment type to overall numbers trapped to derive a percent.
53 As mentioned earlier para grass treatment had an effect on snake presence, but it also had an effect on C. alboguttata abundance (Chi Square Test x2=3215.5, p < 0.001). In burned and control plots, snake abundance was higher than C. alboguttata, while the opposite was found in the grazed and burned plots.
Predator-Prey Dynamics Table 4.1 Prey available to prey consumed by keelbacks in the Town Common. Frogs consists of all frogs trapped in the Town Common during the dry season, which excludes C. alboguttata as it was trapped during the wet season. Anurans Available (%) Consumed (%) Frogs 56 89 Toads 44 11
Gut content was examined prior to the wet season, when trapping in the woodland was conducted. Despite the high availability of toads to frogs, snakes still consumed more frogs than toads (Table 4.1). All frog species were considered possible prey items. Only 27 snakes successfully regurgitated allowing gut contents to be identified.
0.4
0.2 Frogs
0
-0.2 Non-selective feeding
-0.4
-0.6 Electivity Index (E*)
-0.8
-1 Toads
-1.2
Figure 4.8 Diet selectivity of successfully regurgitated keelbacks using an Electivity Index (Ivlev 1977), where 0.0 indicates non-selective feeding. Prey items with an E* value greater than 0.0 are preferred food items, while those with a value less than 0.0 are avoided.
54 0.6 Litoria nasuta 0.4 Non-selective feeding 0.2
0 Opisthodon Limnodynastes ornatus -0.2
-0.4
-0.6 Elecitivity Index (E*)
-0.8
-1 Bufo marinus
-1.2
Figure 4.9 Diet selectivity of keelbacks by species in the Town Common using an Electivity Index (Ivlev 1977), where 0.0 indicates non-selective feeding. Prey items with an E* value greater than 0.0 are preferred food items, while those with a value less than 0.0 are avoided.
With respect to diet selectivity, frogs were preferred over toads (Figure 4.8). With respect to individual species: Litoria nasuta was a preferred prey item; Opisthodon ornatus and the Limnodynastes genus values were so close to 0.0 that they would be considered non-selected for prey items; Bufo marinus was avoided as a prey item (Figure 4.9).
Discussion Population dynamics The low recapture rates did not allow for an accurate assessment of population size. There could be several reasons for the low recapture rates such as a very large population or the snakes could have learned to avoid pitfall traps. However, it is noted that these snakes are those most commonly found in the Town Common (Chapter 2, Table 2.2). Of those captured, more were male than female, although the sex ration did not differ significantly from parity (101 males to 70 females). The females tended to be larger than the males (Figure 4.1,4.2), but this may be due to the fact that some were gravid (note that they were longer as well as heavier).
Floodplain Use The most significant factor affecting snake presence/absence on the floodplain, in this study, was vegetation density. In the Town Common, the para grass forms a very thick mat across the floodplain. Areas of low para grass cover are due to the QPWS/CSIRO para grass treatments; thus, treatment and para grass coverage are very closely related. Para grass tended to clog an area, whereas other vegetation such as native couch Cynondon dactylum was less dense. Snake abundance is associated with areas of high vegetation density and/or less disturbance (Figure 4.3,
55 4.4). In gray rat snakes (Elaphe obsoleta spiloides), predation success was higher in areas with vegetation oppose to those sites without (Mullin 1998). High-density vegetation is advantageous for snakes, offering protection from predators and making them less visible to their prey. This is a common predatory behaviour for aquatic life. Large mouth bass (Micropterus salmoides) and the northern pike (Esox lucius) moved to vegetated areas to ambush their prey rather than utilising open water (Savino 1989). Despite the aquatic habitat, the available vegetation does impact behaviour. While snakes are able to use disturbed areas when necessary, in this study on the Town Common the undisturbed areas with dense vegetation seemed preferred over the disturbed.
Despite the flood season, water levels did not significantly influence snake activity in this wetland. Preferred water depth appeared to be greater than 10mm (Figure 4.5). However, due to other environmental factors such as vegetation, it was very difficult to accurately assess a preferred water depth.
Snake distribution was also associated with prey distribution. As noted in the results, there was a major species compositional change in the floodplain during the wet season (Figure 4.6). C. alboguttata enters its breeding season with the rainfall (Booth 2006), and thus its population rapidly increases and completely dominates the floodplain. When snake abundance was compared with presence of C. alboguttata, snakes inhabit the same areas (Figure 4.7). It appears that snakes migrated into the dense vegetation to follow their prey, which was likely seeking refuge from predation. However, the C. alboguttata population was quite large and completely dominated the floodplain during the time of the trapping, so the data is driven by that population size and could be skewed.
Feeding Habits Tropidonophis mairii is acknowledged as being one of the few species which can consume Bufo marinus (cane toads) (Covacevich 1975; Shine 1991). Despite its reputation for consuming cane toads, keelbacks do not preferentially prey on cane toads even though cane toads are far more abundant than other prey items. When the regurgitated mass was identifiable, it was usually a frog rather than a toad. This would appear to be in congruence with a previous theory that keelbacks consume cane toads because they share a similar habitat and not because they are an ideal prey (Covacevich 1975). Taking into account cane toad reproductive potential and survivability (Lampo 1998) and the seasonally influenced rate at which keelbacks
56 feed (Brown 2002; Brown 2002), keelbacks are unlikely to have a significant effect on the cane toad population.
Conclusion Population and Feeding Habits While the capture rates were high, the low number of recaptures made it difficult to assess normal population parameters for keelbacks in the Town Common. However, this knowledge can be acquired with continued capture effort. The keelback would be a good biological indicator of the effect of the para grass treatments by assessing numbers present both in the floodplain and in the woodland before and after treatments and over time.
Keelback snakes are amongst the most common predators on the Town Common and their habitat use will be affected by any negative effects incurred to the area (Bower 2006). In addition as prey items feel pressure from disturbance and invasive, predation success may be significantly reduced. The cane toad not only directly effects the keelback when consumed, it will indirectly affect them by out-competing other frogs consumed by keelbacks.
Floodplain Treatment Use Keelbacks were more abundant in high-density vegetation. It would seem that at least the keelback population is well adjusted to the dominant, invasive para grass during the wet season. With the incurred negative effects from para grass treatments (Bower 2006), the length of time the para grass has dominated, and the ability of keelbacks to adapt to the thick vegetation, it would seem that for keelbacks, para grass treatment on the Town Common may offer no advantages. If studies concerning other animals groups were to result in similar findings, management regimes for the para grass should be reviewed, especially if the negative effects were long lasting. On the Town Common, the long-term effects of the para grass treatments on animals other than water fowl have only been briefly investigated (Bower 2006). More studies on the long-term consequences of burning and grazing on herpetofauna, mammals, and invertebrates should be conducted to evaluate the benefits of the para grass management practices despite the obvious positive impact on waterfowl populations in the Town Common.
57 Chapter 5 General Conclusion
Understanding the assemblage of flora and fauna sustained by the diverse habitats found in wetlands is fundamental to the management of these complex ecosystems. The core wetland habitat includes all available habitat whether aquatic or terrestrial (Burke 1995). Life histories of different native species may be affected by habitat modifications created by introduced exotics, either flora or fauna. The response varies based on site, and the exotic to which the native species is reacting (Sawadogo 2005). This modified habitat not only effects the way the inhabitants respond, but also the way management techniques should be implemented and which should be used. This thesis explored the ecology of herpetofauna in an invaded wetland – the Townsville Town Common – and how the animals living there used their habitat.
The Town Common wetland has been undergoing management practices for the invasive para grass Urochloa mutica. Grazing by cattle and burning techniques have been separately employed to control the para grass. However, there is little published information on habitat-use by resident herpetofauna
It has been demonstrated that the adjacent woodland is just as important to a wetland as the actual floodplain (Semlitsch and Bodie 2003) and that use of these habitats differs both by season and by species (Sexton 1964; Martin 1988; Sinsch 1988). Herpetofauna, in this study, were more abundant in the adjacent woodland than in the treated floodplain. The loss of floodplain residents (from burning and grazing) was buffered by a large population in the woodland. Only a few species such as Lamprophilis delicata and Litoria fallax were associated almost solely with the floodplain. These species should be taken into consideration when implementing full-scale management practices. These results confirm that all taxa should be studied before enacting management regimes as some can be at risk from these treatments.
This study has revealed the importance and relevance of all habitats in a wetland by demonstrating that herpetofauna use both aquatic and terrestrial environments. When determining management practices, this interconnection between the woodland and floodplain should not be ignored, and consideration must also be given to buffer zones. A consideration of the frequent use of the woodland adjacent to the
58 floodplain by herpetofauna during certain times of the year will help determine the best time to enact treatment practices with minimal adverse effect. If a disturbance, either natural or anthropogenic, occurs in one habitat type the effects should be evaluated in both aquatic and terrestrial zones. This study showed these habitats to be closely connected and used especially by wetland herpetofauna.
A wetland changes from a dry savanna to a full floodplain, and this dynamic creates a change in the abundance, species composition, and behaviour of wetland herpetofauna (Martin 1988; Sinsch 1988; Brown 2002). While herpetofauna populations have been shown to be more abundant in the woodland, the effects of the management treatments (grazing and/or burning) to the floodplain persist during the wet season. This thesis looked in depth at the response of Tropidonophis mairii (keelback snake) to grazing and burning of para grass habitats and to the abundant, toxic cane toad (Bufo marinus).
While management treatments are put in place to preserve the integrity of a wetland as a flood stabilization zone and a nesting site for waterfowl (Salvasen 1990), it is important to question the net benefit of these treatments. While they may be beneficial for the focus taxon/taxa or for economical reasons, there may be negative effects of the treatments (Bower 2006), and additionally from the loss of the invasive species itself (Daehler 2001; Strauss 2006) in both long and short-term responses. This study showed that keelbacks are more abundant in dense para grass vegetation; the treated paddocks had less keelbacks. This implies that para grass has become naturalized; natives now use the invasive as if it were part of the natural vegetation. For example, invasive fig trees (Ficus spp) in Florida, USA are now occupied and pollinated by native wasps (Eupristina spp), and functional needs are met for both (Nadel 1992). In cases such as this and in relation to para grass, it becomes a balance between the positives of weed management vs. those of the weed. The theory of naturalization of the invasive grass should be taken into account prior to instating any weed management regimes. This too can help incur less damage on the native populations that have become accustomed to the exotic species.
While some invasives can enter a natural state, others remain detrimental to the invaded habitat. Cane toads fit this profile. Their numbers increase relatively unrestrained by natural enemies as very few animals can consume their toxic bodies. This study showed that despite the high availability of cane toads, keelbacks, one of
59 the few species which can withstand the toxin (Covacevich 1975) , choose not to eat them. As cane toads become more of a problem for Australian fauna, it is important to note that although they can be consumed by some species, they are preferentially selected against. This reiterates the need for a control and/or eradication method for cane toads, and it needs to be enacted before the consequences of the presence of this species cannot be countered.
While it is important to manage weed spread for ecological and economical reasons it is important too to realize that each species will react differently to the disturbance to which they are subjected. It is nearly impossible to alter an environment − even for beneficial reasons − without causing a negative response from an inhabitant. It is important to be aware of these responses when enacting management techniques in order to decrease the negative consequences. Future investigation needs to focus on more groups (flora and fauna) and how they use both the invaded and treated areas of a wetland. This needs to occur in several seasons to be aware of long-term effects of either the invasive or the management technique. These investigations will not only benefit invasive treatments but also can also help in terms of habitat encroachment by development and the allotment of buffer zones surrounding wetlands.
60 References:
Bergelson, J., Newman, J. A., and Flouresroux, E. M. (1993). Rates of Weed Spread in Spatially Heterogeneous Environments. Ecology 74(4): 999-1011.
Blossey, B., Skinner, L. C., and Taylor, J. (2001). Impact and management of purple loosestrife (Lythrum salicaria) in North America. Biodiversity and Conservation 10: 1787-1807.
Booth, D. T. (2006). Effect of soil type on burrowing behavior and cocoon formation in the green-striped burrowing frog, Cyclorana alboguttata. Canadian Journal of Zoology 84.
Bower, D. (2006). Impact of burning and grazing on reptiles and amphibians in a para grass dominated wetland. School of Tropical Biology. Townsville, James Cook University. Honours Thesis.
Braithwaite, R. W., Friend, G. R., and Wombey, J. C. (1988). Reptiles and amphibians. Monsoonal Australia: Landscape ecology and management. Ridpath, M. G. Haynes, C.D., and Williams, M. A. J. Rotterdam, A. A. Balkema.
Briese, D. T. (1996). Biological Control of Weeds and Fire Management in Natural Area: Are They Compatible Strategies? Biological Conservation(77): 135-141.
Broad, A. J., Sutherland, S. K., and Coulter, A. R. (1979). The lethality in mice of dangerous Australian and other snake venoms. Toxicon 17: 661-664.
Brown, G. P., Shine, R., and Madsen, T. (2002). Responses of three sympatric snake species to tropical seasonality in northern Australia. Journal of Tropical Ecology 18: 549- 568.
Brown, G. P. and Shine, R. (2002). Influence of weather conditions on activity of tropical snakes. Austral Ecology 27: 596-605.
Brown, G. P. and Shine, R. (2004). Effects of reproduction on the antipredator tactics of snakes (Tropidnophis mairii, Colubridae). Behavioral Ecology and Sociobiology 56(3): 257-262.
Buckley, Y. M., Rees, M., Paynter, Q., and Lonsdale, M. (2004). Modeling integrated weed management of an invasive shrub in tropical Australia. Journal of Applied Ecology 41: 547-560.
Buhlmann, K. A. (1995). Habitat use, terrestrial movements, and conservation of the turtle, Deirochelys reticularia in Virginia. Journal of Herpetology 29: 173-181.
Bureau of Meteorology. (2007). Retrieved October 23, 2007, from http://www.bom.gov.au/climate/averages/tables/CW_032040.shtml.
Bureau of Meteorology (2007). Climate and Temperature for Townsville, QLD. Retrieved October 23, 2007, from http://www.bom.gov.au/cgi-bin/climate/cgi- bin_scripts/map_script_new.cgi?32040
61 Burger, J. and Zappalorti, R. T. (1989). Habitat Use by Pine Snakes (Pituophis melanoleucus) in the New Jersey Pine Barrens: Individual and Sexual Variation. Journal of Herpetology 23(1): 68-73.
Burke, V. J. and Gibbons, J. W. (1995). Terrestrial Buffer Zones and Wetland Conservation: A Case Study of Freshwater Turtles in Carolina Bay. Conservation Biology 9(96): 1365-1369.
Charudattan, R. (2001). Are we on top of aquatic weeds? Weed problems, control options and challenges. The World's Worst Weeds, British Crop Protection Council Symposium, Farnham, Surrey, U.K.
Congdon, R. (2006). Average rainfall for Townsville. A. Pearcy. Townsville.
Cook, G. D., Setterfield, S. A., and Madison, J. P. (1996). Shrub Invasion of a Tropical Wetland: Implications for Weed Management. Ecological Applications 6(2): 531-537.
Costanza, R., d'Arge, R., de Groot, R., Farber, S., Grasso, M., Hannon, B., Limburg K., Naeem, S., O'Neil, R. V., Pareulo, J., Raskin, R. G., Sutton, P., and van den Belt, M. (1997). The value of the world's ecosystem services and natural capital. Nature 387: 253-260.
Covacevich, J. and Archer, M. (1975). The distribution of the cane toad, Bufo marinus, in Australia and its effects on indigenous vertebrates. Memoirs of the Queensland Museum 17: 305-310.
Cox, G. W. (2002). General Ecology Laboratory Manual: Eighth Edition. New York, McGraw-Hill.
Daehler, C. C. (2001). Darwin's Naturalization Hypothesis Revisited. The American Naturalist 158(3): 324-330.
D'Antonio, C. M. and Vitousek, P. M. (1992). Biological invasions by exotic grasses, the grass/fire cycle, and global change. Ann. Rev. Ecol. Syst. 23(63-87).
Daily, G., Ed. (1997). Nature's Services: Societal Dependence on Natural Ecosystems. Washington, D. C., Island Press.
Dalrymple, G. H., Steiner, T. M., Nodell, R. J., and Bernardino, Jr., F. S. (1991). Seasonal Activity of the Snakes of Long Pine Key, Everglades National Park. Copeia 1991(2): 294-302.
DiTomaso, J. M. (1997). Risk Analysis of Various Weed Control Methods. California Exotic Pest Plant Council.
DiTomaso, J. M. (2000). Invasive Weeds in Rangeland: Species, Impacts, and Management. Weed Science 48(2): 225-265.
Dodd, C. K. and Cade, B. S. (1998). Movement Patterns and the Conservation of Amphibians Breeding in Small, Temporary Wetlands. Conservation Biology 12(2): 331-339.
62 Douglas, M. M. and O'Connor, R. A. (2003). Effects of the exotic macrophyte, para grass (Urochloa mutica), on benthic and epiphytic macroinvertebrates of a tropical floodplain. Freshwater Biology 48: 962-971.
Douglas, M. M. and O' Connor, R. A. (2004). Weed invasion changes fuel characteristics: Para grass (Urochola mutica) on a tropical floodplain. Ecological Management and Restoration 5: 143-5.
Duellman, W. E. and Trueb, L. (1986). Biology of Amphibians. New York, McGraw Hill.
Enge, K. M. (2001). The Pitfalls of Pitfall Traps. Journal of Herpetology 35(3): 467- 478.
Fagan, W. F., Lewis, M. A., Neubert, M. G., and van den Dreissche, P. (2002). Invasion theory and biological control. Ecology Letters 5: 148-157.
Ferdinands, K., Beggs, K., and Whitehead, P. (2005). Biodiversity and Invasive Grass Species: Multiple Use or Monoculture. Wildlife Research 48(2): 225- 265.
Findlay, C. S. and Houlahan, J. (1997). Anthropogenic Correlates of Species Richness in Southeastern Ontario Wetlands. Conservation Biology 11(4): 1000-1009.
Fox, M. D. and Fox, B. J. (1986). The susceptibility of natural communities to invasion. Ecology of Biological Invasions: An Australian Perspective. R. H. Groves and J. J. Burdon, Eds. Canberra, Australian Academy of Science: 57- 66.
Friend, G. R. and Cellier, K. M. (1990). Wetland Herpetofauna of Kakadu National Park, Australia: Seasonal Richness Trends, Habitat Preferences and the Effects of Feral Ungulates. Journal of Tropical Ecology 6(2): 131-152.
Galatti, U. (1992). Population Biology of the Frog Letodactylus pentadactylus in Central Amazonian Rainforest. Journal of Herpetology 26(1): 23-31.
Gibbons, J. W. (2003). Terrestrial Habitat: A Vital Component for Herpetofauna of Isolated Wetlands. Wetlands 23: 630-635.
Google Earth Plus (2007). Towsville Town Common. Retrieved October 23, 2007, 2007.
Greenberg, C. H., Neary, D. G., and Harris, L. D. (1994). A Comparison of Herpetofaunal Sampling Effectiveness of Pitfall, Single-ended, and Double- ended Funnel Traps Used with Drift Fences. Journal of Herpetology 28(3): 319-324.
Grice, A. C. (2006). Personal Communication. A. Pearcy. Townsville.
Heimlich, R. E. (1994). Costs of an Agricultural Wetland Reserve. Land Economics 70(2): 234-246.
Hobbs, R. J. and Huenneke, L. F. (1992). Disturbance, Diversity and Invasion: Implications for Conservation. Conservation Biology 6(3): 324-337.
63
Hobbs, R. J. and Humphries, S. E. (1995). An integrated approach to the ecology and management of plant invasions. Conservation Biology 9: 761-770.
Howard, G. W. and Harley, K. L. S. (1998). How do floating aquatic weeds affect wetland conservation and development? How can these effects be minimised? Wetlands Ecology and Management 5: 215-225.
Humphries, S. E., Groves, R. H., and Mitchell, D. S. (1991). Plant invasions and Australian ecosystems: a status review and management direction. Plant Invasions: The Incidence of Environmental Weeds in Australia. Canberra, Australian National Parks and Wildlife Service: 1-127.
Ivlev, V. S. (1977). Eksperimentalnaya ekologiya pikaniya ryb (Experimental ecology of fish feeding). Kiev, Naukova Dunka.
Izhaki, I, Levey, D. J., and Silva, W. R. (2003). Effects of prescribed fire on ant community in Florida pine savanna. Ecological Entmology 28(4): 439-448.
Jacqmain, E. I., Jones, R. H., and Mitchell, R. J. (1999). Influences of Frequent, Cool- season Burning Across a Soil Moisture Gradient on Oak Community Structure in Longleaf Pine Ecosystems. The American Midland Naturalist 14(1): 85-100.
Kenward, R. E. and Holm, J. L. (1993). On the replacement of the red squirrel in Britain: a phytotoxic explanation. Proceedings: Biological Sciences 251(1332): 187-194.
Kushlan, J. A. (1986). The Management of Wetlands for Aquatic Birds. Colonial Waterbirds 9(2): 246-248.
Lampo, M. and De Leo, G. A. (1998). The Invasion Ecology of the Toad Bufo marinus from South America to Australia. Ecological Applications 8(2): 388-396.
Lonsdale, V. M. (1993). Rates of spread of an invading species- Mimosa pigra in northern Australia. Journal of Ecology 81: 513-521.
Lonsdale, V. M. (1994). Inviting trouble: introduced pasture species in northern Australia. Australian Journal of Ecology 19: 345-354.
Martin, K. C. and Freeland, W. J. (1988). Herpetofauna of a northern Australian monsoon rain forest: seasonal changes and relationships to adjacent habitats. Journal of Tropical Ecology 4: 227-238.
Martof, B. S. (1953). Territoriality in the Green Frog, Rana clamitans. Ecology 34(1): 165-174.
Mashhadi, H. R. and Radosevich, S. R. (2004). Invasive Plants. Weed Biology and Management. Inderjit. Dordrecht, Kluwer Academic Publishers: 1-28.
McMurty, J. A., Andres, L. A., Bellows, Jr., T. S., Hoyt, S. C., and Hagen, K. S. (1995). A historical overview of regional research project W-84. Biological Control in the Western United States. J. R. Nechols, L. A. Andres, J. W. Beardsley, R. D. Goeden, and C. G. Jackson, Eds. Oakland, University of California, Division of Agriculture and Natural Resources.
64 Mitsch, W. J. and Gosselink, J. G. (1993). Wetlands. New York, Van Nostrand Reinhold.
Monaco, T. J., Weller, S. C., and Ashton, F. M. (2002). Weed Science: Principles and Practice. Indianapolis, Jon Wiley and Sons.
Mooney, H. A. and Cleland, E. E. (2001). The evolutionary impact of invasive species. PNAS 58(10): 5446-5451.
Moreira, I., Ferreria, T., Montiero, A., Catarino, L., and Vasconcelos, T. (1999). Aquatic weeds and their management in Portugal: insights and the international context. Hydrobiologia 415(0): 229-234.
Moreira, I., Ferreria, T., and Monteiro, A. (1989). Aquatic weed bioecology and control in Portugal. A review. Portuguese- German Cooperation in Applied Agricultural Activities. A. Bianchi. Vila Real, University of Vila Real Publication: 71-106.
Morrison, J. (2002). Wetland Vegetation Before and After Experimental Purple Loosestrife Removal. Wetlands 22(1): 159-169.
Mullin, S. J., Cooper, R.J., and Gutzke, W.H.N. (1998). The foraging ecology of the gray rat snake (Elaphe obsoleta spiloides). III. Searching for different prey types in structurally varied habitats. Canadian Journal of Zoology 76: 548- 555.
Navel, H., Frank, J. H., and Knight, Jr., R. J. (1992). Escapees and Accomplices: the naturalization of exotic ficus and their associated faunas in Florida. The Florida Entomologist 75(1): 29-38.
Padilla, D. K. and Williams, S. L. (2004). Beyond ballast water: aquarium and ornamental trades as sources of invasive species in aquatic ecosystems. Frontiers in Ecology and the Environment 2(3): 131-138.
Parris, M. J. (1998). Terrestrial burrowing ecology of newly metamorphosed frogs (Rana pipines complex). Canadian Journal of Zoology 76: 2124-2129.
Paynter, Q. (2004). Revegetation of a wetland following control of the invasive woody weed, Mimosa pigra, in the Northern Territory, Australia. Ecological Management and Restoration 5(3): 191-198.
Pearson, D., Shine, R., and Williams, A. (2005). Spatial ecology of a threatened python (Morelia spilota imbricata) and the effects of anthropogenic habitat change. Austral Ecology 30(3): 261-274.
Penfound, W. T. and Earle, T. T. (1948). The Biology of the Water Hyacinth. Ecological Monographs 18(4): 447-472.
Phillips, B. L., Brown, G. P., and Shine, R. (2003). Assessing the Potential Impact of Cane Toads on Australian Snakes. Conservation Biology 17(6): 1738-1747.
Phillips, B. L. and Shine, R. (2004). Adapting to an invasive species: Toxic cane toads induce morphological change in Australian snakes. PNAS 101(9): 17150-17155.
65 Pollett, J. and Omi, P. N. (2002). Effect of thinning and prescribed burning on crown fire severity in ponderosa pine forests. International Journal of Wildland Fire 11: 1-10.
Preacher, K. J. (2001). Calculation for the chi-square test: An interactive calculation tool for chi-square tests of goodness of fit and independence.
Rea, N. and Storrs, M. J. (1999). Weed invasions in wetlands of Australia's Top End: reasons and solutions. Wetlands Ecology and Management 7(1-2): 47-62.
Rejmanek, M. (2000). Invasive plants: approaches and predictions. Austral Ecology 25: 497-506.
Richardson, D. M., Pysek, P., Rejmanek, M., Barbour, M. G., Panetta, F. D., West, C. J. (2000). Naturalization and invasion of alien plants: concepts and definitions. Diversity and Distributions 6: 93-107.
Roe, J. H., Kingsbury, B. A., and Herbert, N. R. (2003). Wetland and Upland Use Patterns in Semi-Aquatic Snakes: Implications for Wetland Conservation. Wetlands 23 (4): 1003-1014.
Salvasen, D. (1990). Wetlands: mitigating and regulating development impacts. Washington, D.C., Urban Land Institute.
Savino, J. F. and Stein, R.A. (1989). Behavior of fish predators and their prey: habitat choice between open water and dense vegetation. Environmental Biology of Fishes 24(4): 287-293.
Sawadogo, L., Tiveau, D., and Nygard, R. (2005). Influence of selective tree cutting, livestock and prescribed fire on savannah woodlands of Burkina Faso, West Africa. Agriculture, Ecosystems and Environment 105(1-2): 335-345.
Schmitz, D. C., Nelson, B. V., Nall, L. E., and Schardt, J. D. (1988). Exotic Aquatic Plants in Florida: A Historical Perspective and Review of the Present Aquatic Plant Regulation Program. Symposium on Exotic Pest Plants, University of Miami, Miami, Florida.
Seebacher, F. and Alford, R. A. (1999). Movement and Microhabitat Use of a Terrestrial Amphibian (Bufo marinus) on a Tropical Island: Seasonal Variation and Environmental Correlates. Journal of Herpetology 33(2): 208-214.
Semlitsch, R. D. and Bodie, J. R. (1998). Are Small, Isolated Wetlands Expendable? Conservation Biology 12(5): 1129-1133.
Semlitsch, R. D. and Bodie, J. R. (2003). Biological Criteria for Buffer Zones around Wetlands and Riparian Habitats for Amphibians and Reptiles. Conservation Biology 17(5): 1219-1228.
Sexton, O. J., Heatwole, H., and Knight, D. (1964). Correlation of microdistribution of some Panamanian reptiles and amphibians with structural organization of the habitat. Caribbean Journal of Science 4: 261-295.
Shine, R. (1991). Strangers in a Strange Land: Ecology of Australian Colubrid Snakes. Copeia 1991(1): 120-131.
66 Siemann, E., Harstaad, J. and Tilman, D. (1997). Short-term and Long-term Effects of Burning on Oak Savanna Arthropods. American Midland Naturalist 137(349-361).
Sinsch, U. (1988). Seasonal changes in the migratory behaviour of the toad Bufo bufo: direction and magnitude of movements. Oecologia 76: 390-398.
Strauss, S. Y., Lau, J. A., and Carroll, S.P. (2006). Evolutionary responses of natives to introduced species: what do introductions tell us about natural communities? Ecology Letters 9: 357-374.
Thompson, D. Q., Stuckey, R. L., and Thompson, E. B. (1987). Spread, impact and control of purple loosestrife (Lythrum salicaria) in North American wetlands. U.S. Department of the Interior, Fish and Wildlife Service, Washington, Dc, USA. Fish and Wildlife Research Report 2.
Turner, C. E., Center, T. D., Burrows, D. W., and G.R. Buckingham (1997). Ecology and management of Melaleuca quinquenervia, an invader in wetlands in Florida, U.S.A. Wetlands Ecology and Management 5(3): 165-178.
Valentine, L., Schwarzkopf, L., Johnson, C. N., and Grice, A. C. (2007). Burning season influences the response of bird assemblages to fire in tropical savannas. Biological Conservation 137(1): 90-101.
Vallan, D. (2002). Effects of anthropogenic environmental changes on amphibian diversity in the rain forests of eastern Madagascar. Journal of Tropical Ecology 18: 725-742. van der Valk, A. G. (2006). The Biology of Freshwater Wetlands. Oxford, Oxford University Press.
Vermeij, G. J. (1996). An Agenda for Invasion Biology. Biological Conservation 78: 3- 9.
Vitousek, P. M. (1990). Biological invasions of ecosystem processes: towards an integration of population biology and ecosystem studies. Oikos 57: 7-13.
Vonesh, J. R. (2001). Patterns of Richness and Abundance in a Tropical African Leaf-litter Herpetofauna. Biotropica 33(3): 502-510.
Ward, J. V., Tockner, K. and Schiemer, F. (1999). Biodiversity of floodplain river ecosystems: ecotones and connectivity. Regulated Rivers: Research and Management 15(1-3): 125-139.
Wetzel, R. G. (1990). Land-water interfaces: metabolic and limnological regulators. Verhandlungen Internationale Vereinigung für Theoretische und Angwandle Limnologie 24: 6-24.
Wilcox, D. A. (1995). Wetland and aquatic macrophytes as indicators of anthropogenic hydrologic disturbance. Natural Areas Journal 15: 240-248.
Williams, P. R., Collins, E. M., and Grice, A. C. (2005). Cattle Grazing for Para Grass Management in a Mixed Species Wetland of Northeastern Australia. Ecological Management and Restoration 6(1): 75-76.
67 Williams, S. E., and Hero, J. (1998). Rainforest Frogs of the Australian Wet Tropics: Guild Classification and the Ecological Similarity of Declining Species. The Royal Society Proceedings: Biological Sciences 265(1396): 597-602.
Woolcock, J. L. and Cousens, R. (2000). A mathematical analysis of factors affecting the rate of spread of patches of annual weeds in an arable field. Weed Science 48(1): 27-34.
Zedler, J. B. and Kercher, S. (2004). Causes and Consequences of Invasive Plants in Wetlands: Opportunities, Opportunists and Outcomes. Critical Reviews in Plant Sciences 23(5): 431-452.
68