SEARCH

SEARCH BY CITATION

Keywords:

  • amphibians;
  • climate change;
  • Limnodynastes tasmaniensis;
  • predictive modelling;
  • urbanization

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

Climate change and urbanization are among the most serious threats to amphibians, although little is known about their combined effects. We used a predictive spatial habitat suitability model to explore the potential impacts of climate change and urban development on the spotted marsh frog (Limnodynastes tasmaniensis) on the urban-fringe of Melbourne, Australia. The CSIRO climate-change predictions for the region indicate likely temperature increases of 3°C, and annual rainfall reductions of around 200 mm by the year 2070. Much of the study area overlaps a region that has been identified as one of the city's growth corridors. We used Bayesian logistic regression modelling to estimate current and future habitat suitability of pond sites in the Merri Creek catchment, exploring a range of best- to worst-case scenarios through the use of hydrological and urbanization models. Our predictions for 2070, even under a moderate climate-change scenario, suggest that the majority of ponds in the study area will be dry throughout much of the year. This has obvious implications for L. tasmaniensis, which is an aquatic breeding species. However, in the short term, urbanization is likely to have a more significant effect on the distribution of L. tasmaniensis in the Merri Creek catchment, particularly if development moves beyond the current urban growth boundary. The combined effects of climate change and urbanization could have a profound impact on the species, potentially causing it to disappear from within the study area. We provide recommendations for including such predictive models in urban planning and restoration activities to prepare for future conservation challenges.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

The rate of global extinctions of plant and animal species is accelerating, and vertebrate taxa are disappearing at a disproportionate rate (Baillie et al. 2004). Of these, amphibians are considered to be particularly vulnerable. In recent decades, there has been a significant increase in reports of declines and extirpations of amphibians worldwide (Wake 1991; Houlahan et al. 2000; Gardner 2001), and some 30% of extant species are threatened with extinction (Baillie et al. 2004; Cushman 2006). Climate change is among the many processes postulated to explain these declines (Alford & Richards 1999; Kiesecker et al. 2001). Climate change may be responsible for species range shifts and loss of suitable habitat (Carey & Alexander 2003; Araújo & Pearson 2005), the emergence of catastrophic pathogenic outbreaks (Blaustein & Kiesecker 2002; Pounds et al. 2006) and changes in breeding phenology and altered wetland hydroperiod for larval development (Beebee 1995; Rowe & Dunson 1995; Blaustein et al. 2001; Ryan & Winne 2001).

Habitat loss and fragmentation through urbanization is another major threat to amphibians, leading to reduced population sizes and increased isolation (Cushman 2006; Parris 2006; Curado et al. 2011; Hamer & Parris 2011). Urbanization reduces the number of wetland habitats available to amphibians, and reduces the quality of those that remain through changes to aquatic vegetation, water chemistry and flow regimes, establishment of exotic species, and by acting as a source of various pollutants (Ehrenfeld 2000; Urban et al. 2006; Felton et al. 2009). Roads and other high-density developments act as a barrier to the movement and migration of frogs and roads are a major source of mortality in built-up areas (Vos & Chardon 1998; Carr & Fahrig 2001; Bouchard et al. 2009). Traffic noise may also reduce the distance over which calling frogs can be heard by potential mates, due to acoustic masking (Bee & Swanson 2007; Parris et al. 2009).

There is a risk that research focusing on the impacts of climate change or habitat loss in isolation may underestimate the likelihood of changes in the suitability of aquatic habitats (Travis 2003; Pyke 2005). A review of extinction research and climate change by Brook et al. (2008), suggest that rates of extinction may be accelerated by synergies between threatening processes, rather than one all-pervasive threat. The ability for species to adapt or migrate in response to climate change may be greatly reduced in fragmented landscapes, particularly for species with low colonizing or dispersal ability (Travis 2003).

This study provides a predictive, spatial habitat suitability model of the impacts of climate change and urbanization on the potential distribution of the spotted marsh frog (Limnodynastes tasmaniensis) within an urban growth area on the northern outskirts of Melbourne, Australia. Given the ecology and life-history of L. tasmaniensis, such as its reliance on terrestrial habitat for foraging and migration, and the seasonal availability of water for breeding and juvenile development, we hypothesise that some populations of this species will be adversely affected by changes in climate and landscape structure.

We used Bayesian logistic regression modelling to estimate current and future habitat suitability of wetlands in the Melbourne region, exploring eight scenarios covering a range of best- to worst-case changes in climate and urbanization between 2010 and 2070. The first three scenarios examine how pond habitat suitability varies with different climate-change predictions, the next three scenarios examine the impact of assumptions regarding future urban growth and the last two scenarios combine these two impacts.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

Study area

This study was conducted within the southern part of the Merri Creek catchment (Fig. 1). The Merri Creek is a major tributary of the Yarra River that flows for more than 70 km from Mount Disappointment in the Great Dividing Range to the northern suburbs of Melbourne, Victoria (McMahon & Schulz 1993). Since European settlement, the Merri Creek catchment has experienced extensive land clearing for agriculture, and industrial and urban development. Although severely degraded in parts, there are still important remnants of former vegetation communities characteristic of the basalt plains north of Melbourne, including red gum woodlands, stony knolls and grasslands. These grasslands are now considered critically endangered, with only 0.1% of their pre-European extent remaining (Beardsell 1997; Bush et al. 2003). In biological terms, the Merri Creek is considered a habitat corridor of regional and state significance (McMahon & Schulz 1993), supporting numerous threatened fauna species including birds, amphibians and insects (Bush et al. 2003).

image

Figure 1. Study sites along the Merri Creek corridor, located both inside and outside the urban growth boundary, Melbourne, Victoria, Australia.

Download figure to PowerPoint

The Hume Growth Corridor, which extends along Merri Creek, is one of five designated growth areas proposed for future development in Melbourne (Department of Sustainability and Environment 2004a). An urban growth boundary (UGB) was established to manage growth on the city's fringe and to protect the ‘green wedges’ from further ad-hoc development (Buxton & Tieman 2004). Nonetheless, the UGB has been extended several times in key growth corridors to provide new land for urban and industrial development (Buxton & Tieman 2004).

Species description

Limnodynastes tasmaniensis occurs in many habitats, from wet coastal woodlands to the dry interior regions of Australia. It is one of the most common species within its range, and is often the first to colonize new habitats, such as drainage lines and farm dams (Barker et al. 1995). The frog is usually found in association with water, and in dry periods, will shelter in cracks in the ground, or under logs, stones and debris throughout the surrounding landscape (Barker et al. 1995). Breeding occurs most commonly between August and March; however, it has been known to breed outside of these times (Barker et al. 1995). Males call from the edge of shallow water, partly concealed by vegetation. The species lays floating foam nests of around 1000 eggs in water attached to emergent vegetation and tadpoles take 3–5 months to develop (Anstis 2002). Although the species is not considered to be under immediate threat in most regions, the effects of climate change are likely to impact upon the frog by reducing the availability of water for breeding and juvenile development. For this species, temporary water bodies play an important role for breeding and juvenile development (Cogger 1992). It is also reliant on terrestrial habitat for migration, foraging and overwintering sites (Barker et al. 1995). Even though L. tasmaniensis is thought to be tolerant of and even benefit from certain landscape modifications, it may also be threatened by the impacts of urbanization (Frogs Australia Network 2005; Smallbone et al. 2011). A study by Smallbone et al. (2011) found that the species declined with increasing urban intensity.

Sampling protocols used in previous research

This study combined presence/absence and habitat data from two different surveys to develop the habitat model, and predict the current habitat suitability of ponds in the Merri Creek study area. KMP conducted a study of pond-breeding frogs across Greater Melbourne (2000–2002; Parris 2006), collecting data from nocturnal searches for frogs across 104 sites over spring and summer. She recorded the following habitat variables for each pond site: pond area, the presence or absence of a vertical pond wall, and the proportion of the pond with fringing, emergent and submerged vegetation. Fringing vegetation was calculated as the proportion of pond edge with fringing terrestrial vegetation up to 1 m from the pond edge. Emergent and submerged vegetation was measured as the proportion of the pond surface area covered by aquatic vegetation that did and did not extend above the surface of the water. Road cover was calculated in ArcGIS 9.1 (http://www.esri.com/software/arcgis/index.html) as the proportion of land within a 500 m radius of the centre point of each pond (0.787 km2) that was covered with sealed roads. Road cover was used as a surrogate measure of the level of urbanization and isolation of each pond.

GW Heard et al. (2001–2007, unpubl. data) collected data on frog presence/absence and habitat variables for 139 creek and pond sites in the Merri Creek catchment. This included both daytime and nocturnal searches for all amphibian species. They calculated fringing vegetation as the percentage of pond edge covered by fringing terrestrial vegetation. Emergent vegetation was broken into two categories, one being the percentage of water surface area covered by emergent vegetation, the other being the percentage of water edge covered by emergent vegetation. Submerged vegetation was measured as the percentage of water surface area covered with submerged vegetation. In addition, the study estimated the proportion of surface area covered by floating vegetation.

Habitat model

KMP developed a habitat model in a previous study to assess the applicability of metacommunity theory to urban amphibian assemblages (Parris 2006). The study evaluated three assumptions that underlie metacommunity theory – the effect of patch area, the effect of patch isolation and species–environment relations. The study revealed that pond size (patch area) has a positive effect on amphibian assemblages but that ponds with high levels of surrounding road cover (isolated patches) have a detrimental effect on amphibian assemblages. KMP's habitat model was incorporated into this study and her data were used as informative priors on the likelihood of the species occupying a pond given certain habitat conditions. Data collected by GW Heard et al. (2001–2007, unpubl. data) were incorporated into the model as the primary source of data pertaining to the frog and habitat variables found at the sites in the Merri Creek catchment. These data were used to examine habitat suitability for L. tasmaniensis in the study region. The hydrological model developed for this study was used to estimate changes in pond (patch) area as a result of predicted changes in climate for the region. The urbanization model developed by Parris (2006) was used to assess the amount of road cover currently surrounding the pond sites.

Hydrological model

Here we developed a model to predict hydrological changes to the study sites in the Merri Creek catchment under various climate-change projections. There are many different hydrological factors influencing the inflow and outflow of water in streams. There are also many uncertainties in projections from hydrological models (R Nathen & SK Merz 2007, pers. comm.). The sites chosen for this study were subsequently restricted to lentic wetlands such as swamps, ponds, dams and quarries, of which there were 46 in total. The hydrology of such wetlands (herein called ‘ponds’) can be approximated using a water-balance model (or bucket model), in which rainfall and evaporation are the principal hydrological drivers (Stamm et al. 1993). Only the most important processes are represented in simple ‘bucket’ models (Zhang et al. 2002). This model does not take into account other hydrological factors, such as seepage due to porous soil types, or input from run-off on adjacent lands. However, given that precipitation and evaporation are the main hydrological factors influencing these ponds and that the study only requires estimates of changes in pond depth and area; the water-balance model was considered a reasonable tool for describing pond hydrology in this study.

The hydrological model extrapolates average monthly rainfall and evaporation data, as recorded by the Bureau of Meteorology at various weather stations for the Greater Melbourne region (Torok & Nicholls 1996). The model assumes the mean rainfall and evaporation of a given month linearly changes from the present value to the value predicted at 2070 for each climate-change scenario. Variability was incorporated into the projected time series data by adding a random variability factor in to each time step. The size of the variability factor was set by analysing historic time-series data from the same region. Changes in pond surface area under each of the time steps were calculated by making an assumption that ponds have a spherical depth profile.

The climate-change projections for mean rainfall and evaporation in the study area between 2010 and 2060 were generated using the software package OzClim (CSIRO 1996). The CSIRO climate-change projections (Whetton et al. 2002) provided a range of possible climatic changes for the Greater Melbourne region. However, uncertainties about future greenhouse gas emissions and the shortcomings of current climatic models limit the level of accuracy that can be attained through predicting future climatic conditions. Therefore, this study investigated a range of climate-change projections given by the CSIRO (2005), providing both upper and lower limits of the projections, to capture a range of possibilities.

Given the climate scenarios, the hydrological model then calculates changes in pond depth and surface area over time as a function of the projected rainfall and evaporation time series data (Fig. 3). For example, in the warmer, drier months, rainfall decreases and evaporation increases causing pond levels to drop and surface areas to decrease. In the cooler, wetter months, rainfall increases and evaporation decreases, causing a rise in pond levels and surface area. Under normal climatic conditions, pond levels should undergo a cycle of seasonal loss and gain of water, with some years being dryer than others. However, under CSIRO climate-change projections for the Melbourne region, average rainfall is likely to decrease and average temperature (and thus, evaporation) is likely to increase permanently. This has the potential to alter the hydrology of ponds in this region, and may drive many to desiccation (this is shown in Fig. 3).

Over the last 8 years when our data were collected, drought conditions in Melbourne (and Australia as a whole) have been the worst on record. This has been attributed to climatic events such as El Niño–Southern Oscillation and perhaps the effects of global warming (CSIRO & Australian Bureau of Meteorology 2008). Therefore, it is important to predict the potential effects further water constraints may have on biodiversity, particularly for species dependent on permanent and/or temporary water sources for breeding and juvenile development. For the purpose of this study, we treated the last 8 years of severe drought conditions as a local and separate phenomenon to global climate change. Therefore, our climate-change projections make the possibly optimistic assumption that the region will return to average rainfall conditions before being affected by climate change. The three climate-change scenarios used are described in Table 1 and can be summarized as follows: no climate change, lower-limits of the CSIRO projections and upper-limits of most severe climate-change projections. Given the sensitivity of L. tasmaniensis to water availability (pond area), we were able to calculate its probability of occupying each of the ponds under the climate-change projections.

Table 1. The climate-change and urbanization scenarios explored in the study, with descriptions of factors included in each projection
Scenario numberScenario namePredictionAssumptions
  1. UGB, urban growth boundary.

1Best-case, no climate change or further urbanizationRemains the same as current habitat suitability• Average rates in rainfall and evaporation are continual
2Intermediate climate change, no further urbanizationNegative change in habitat suitability (probability of occupancy by Limnodynastes tasmaniensis)• Decreased rainfall (−5% by 2030 and −15% by 2070) and increased evaporation (+6% by 2030 and +18% by 2070)
• Increasingly warmer and drier climate
3Worst-case climate change, no further urbanizationNegative change in habitat suitability (probability of occupancy by L. tasmaniensis)• Decreased rainfall (−10% by 2030 and −25% by 2070) and increased evaporation (+11% by 2030 and +34.4% by 2070)
• Increasingly warmer and drier climate
4Intermediate urbanization up to UGB, no climate changeNegative change in habitat suitability (probability of occupancy by L. tasmaniensis)• Average increase in road cover within 500-m radius of pond sites (7.5% increase to all sites within UGB, 0% increase to sites outside UGB)
• Sites have a maximum road cover of 25%
• Urban development continues up to the UGB
5Intermediate urbanization beyond UGB, no climate changeNegative change in habitat suitability (probability of occupancy by L. tasmaniensis)• Average increase in road cover within 500 m radius of pond sites up to and beyond UGB (7.5% increase to all sites within the entire study area)
• Sites have a maximum road cover of 25%
• Urban development extends beyond the UGB
6Worst-case urbanization up to UGB, no climate changeNegative change in habitat suitability (probability of occupancy by L. tasmaniensis)• Maximum increase in road cover within 500 m radius of pond sites (25% increase to all sites within UGB, 0% increase outside UGB)
• Sites have a maximum road cover of 25%
• Urban development continues up to the UGB
7Intermediate climate-change and urbanization projectionNegative change in habitat suitability (probability of occupancy by L. tasmaniensis)• Intermediate climate-change and urbanization projection (development beyond the UGB). Average increase in road cover within 500 m radius of pond sites (7.5% increase to all sites within the entire study area)
• Sites have a maximum road cover of 25%
• Urban development extends beyond the UGB
8Worst-case climate-change and urbanization projectionNegative change in habitat suitability (probability of occupancy by L. tasmaniensis)• Worst-case climate-change and urbanization projection (development beyond the UGB). Maximum increase in road cover within 500 m radius of pond sites (25% increase to all sites)
• Sites have a maximum road cover of 25%
• Urban development extends beyond the UGB

To examine the validity of the hydrological model, a test was conducted using recorded rainfall/evaporation data from the Merri Creek catchment between January 2002 and January 2007. The data were used as an input for the model, showing estimated changes in pond levels over that time. The model showed significant decreases in pond depth as a result of the drought conditions seen over the last 5 years. The results of the hydrological model were compared with observed changes in pond depth for each pond in the study area over the 5 years that the study took place. The trends in pond depth predicted by the model were in concordance with those observed by GW Heard et al. (2001–2007, unpubl. data), indicating that a simple input–output model simulates the hydrology of this system reasonably well.

Urbanization model

Current levels of urban development for each study site were quantified following techniques used by Parris (2006). Road cover was used as a quantitative measure of the level of urbanization in the area surrounding a pond and therefore its isolation from other ponds. It can also be used as a measure of pond quality, as roads can have a profound influence on both the physical and chemical compositions of aquatic environments (Trombulak & Frissell 2000). A radius of 500 m was chosen as an appropriate scale for measuring the effect of roads on frogs. This number was chosen based on other studies investigating buffer distances and amphibians (e.g. Semlitsch & Bodie 2003; Lauck 2005; Mazerolle et al. 2005) as there is currently little information on the dispersal distances of Australian frogs.

Information on the UGB was obtained from the Victorian Department of Sustainability and Environment's Corporate Geospatial Data Library (Department of Sustainability and Environment 2004b). To represent a range of future urban development, three urbanization scenarios were developed (Table 1). Climate-change scenarios 1–3 assumed no further urban development, scenario 4 assumes that development will continue up to the UGB, encompassing all sites within the boundary, and will not continue beyond that point. It incorporates an average increase in road density for each site, suggesting that future urban development will be of a similar density to that seen in nearby urban-fringe areas. Scenario 5 assumes that development will extend beyond the UGB encompassing all the sites in the study area. It also incorporates an average increase in road density for each site, suggesting that future urban development will be of a similar density to other fringing urban areas. Scenario 6 assumes that development will continue up to the UGB, encompassing all sites within the boundary and will not continue beyond that point. It incorporates a higher road density for each site, suggesting that future urban development be of a similar density to that of the built-up inner suburbs.

Scenarios 7 and 8 incorporate both climate change and urbanization, and represent an intermediate and an extreme scenario, respectively (Table 1). This allows an examination of the probability of L. tasmaniensis occupying the study sites under the worst-case climate/urbanization scenario, and a more moderate climate-change/urbanization scenario.

Predicting current and future habitat suitability

Bayesian logistic regression modelling in WinBUGS 1.4 (Spiegelhalter et al. 2003) was used to estimate the current probability of occupancy of L. tasmaniensis at the study sites in the Merri Creek catchment. These data were used to estimate the posterior distribution of the variables of interest (McCarthy 2007). The model included three explanatory variables: pond size, road cover in a 500 m radius around the pond, and the proportion of the pond covered by emergent, submerged and fringing vegetation combined. Each of these habitat variables were given a score between 0 and 1 and the three proportions were added together to give a maximum value of 3.

We used WinBUGS to estimate future probabilities of L. tasmaniensis occupying ponds under a range of climatic and urbanization scenarios. Given that we know pond size has a positive effect on the probability of occupancy, climate change is likely to alter the amount of water entering and evaporating from water bodies, thus the amount of habitat available to the species. Similarly, increased road cover associated with further urban development is likely to have a negative effect on the probability of ponds being occupied. By altering the habitat attributes in the habitat model, we gain a new estimate of whether the ponds are likely to be occupied by the species under a chosen scenario.

WinBUGS was used to generate 100 000 samples from the posterior distribution of the model after discarding the initial 5000 samples as a ‘burn-in’. The mean and standard deviation of the model coefficients were calculated, along with the 2.5th and 97.5th percentiles of the distribution. This interval was used to represent a 95% Bayesian confidence interval (95% credible interval). The probability of occupancy was converted to proportional symbols at the locations of the study sites on a map of the study area in ArcGIS 9.1, to provide a visual representation of the suitability of each site as habitat for L. tasmaniensis.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

Current habitat suitability for L. tasmaniensis

The predicted probability that a site would be occupied by the L. tasmaniensis increased with increased pond area and vegetation cover (Table 2, Fig. 2). However, it decreased dramatically as road cover increased within a 500 m radius of the pond. Indeed, of the three habitat variables, road cover had the greatest influence on the probability of occupancy. For example, the probability of occupancy increased from an average of 15% to 70% with increased pond size, but decreased from an average of 80% to 1% as road cover increased from 0 to 25%, when the other explanatory variables were held constant at their mean.

Table 2. Coefficients of the parameters included in the habitat model, including the mean, standard deviation (SD), 2.5th and 97.5th percentiles of the posterior distribution
VariableMeanSD2.5%97.5%
Constant−2.1641.726−5.6271.186
Pond area0.83670.4818−0.089721.807
Vegetation0.85550.5744−0.24642.004
Road cover−32.9311.17−56.89−13.39
image

Figure 2. Habitat variables influencing the probability of occupancy by Limnodynastes tasmaniensis: (a) log pond size (m2); (b) vegetation cover, calculated as the proportion of fringing, emergent and submerged vegetation; and (c) road cover, with the other variables held constant at their mean. Road cover is the proportion of land within a 500-m radius covered with sealed road. Solid lines show the predicted relationships and dotted lines the 95% credible intervals.

Download figure to PowerPoint

Most suitable habitat for the species occurred in the outer fringe of the city or where development has been minimal. The current habitat suitability map for L. tasmaniensis (see scenario 1, Fig. 4) showed that around half of the ponds with the highest probability of being occupied by L. tasmaniensis occur within the UGB.

Impacts of climate change on habitat suitability for L. tasmaniensis

We calculated the impacts of climate-change projections on water levels for each of the ponds that represented the range of pond depths observed throughout the study area. The deepest pond had a depth of 2500 mm and the shallowest pond a depth of 250 mm. The ‘best-case’ no climate-change projection (scenario1, Fig. 3) showed the normal fluctuations of pond levels under an average weather pattern without a change in climate. The ‘intermediate’ climate-change projection (scenario 2, Fig. 3) suggested that weather patterns will reach a threshold by 2020, whereby pond levels will gradually but steadily decrease to the point where all the ponds will become ephemeral by 2050. Thus, there would no longer be any permanent, naturally fed water bodies in the study area. Under the ‘worst-case’ climate-change projection (scenario 3, Fig. 3) all the ponds in the study area will become ephemeral by 2030. Both the intermediate and worst-case climate-change projections suggested that there will be significant decreases in rainfall and higher rates of evaporation.

image

Figure 3. Changes in water levels over time for four ponds of varying depths under scenarios 1–3 (see Table 1).

Download figure to PowerPoint

Scenarios 2 and 3 (Fig. 4) evaluated the probability of L. tasmaniensis occupying ponds under both the intermediate and worst-case climate-change projections. The intermediate climate-change scenario showed an average decrease in the probability of occupancy of 9% across pond sites, and the worst-case climate-change scenario showed an average decrease in the probability of occupancy of 41% across the pond sites (Fig. 5).

image

Figure 4. The probability of occupancy for each of the study sites under scenarios 1–6. Scenarios 1–3 show the probability of occupancy of Limnodynastes tasmaniensis under three climate-change projections. Scenarios 4–6 show the probability of occupancy of L. tasmaniensis under three urbanization projections. Scenario 4 assumes that an intermediate level of urban development extends up to the urban growth boundary (UGB). Scenario 5 assumes that an intermediate level of urban development extends beyond the UGB. Scenario 6 assumes that the maximum level of urban development extends up to the UGB. The probability of occupancy values are depicted by the size and colour of the points at each pond location. To view this figure in colour, see the online version of this paper.

Download figure to PowerPoint

image

Figure 5. The average probability of occupancy of Limnodynastes tasmaniensis across pond sites under three climate-change scenarios (1–3), under three urbanization scenarios (4–6) and under two combined climate-change and urbanization scenarios (7 and 8). The error bars represent standard deviations.

Download figure to PowerPoint

Impacts of urban development on habitat suitability for L. tasmaniensis

Under scenarios with an intermediate level of urban development and no climate change (scenarios 4 and 5, Fig. 4), habitat that we found to be of high suitability (greater than 75% probability of occupancy) was predicted to decline in suitability such that the average probability of occupancy will be less than 30% (Fig. 5). The scenario with a high level of urban development up to the UGB (scenario 6), habitat of high suitability was predicted to be rendered effectively unsuitable (less than 1% probability of occupancy for ponds within this region; Fig. 4). This demonstrated the significant effect of urbanization on the probability that a site will be occupied by L. tasmaniensis.

Habitat suitability for L. tasmaniensis with climate change and future urban development

The probability of L. tasmaniensis occupying ponds in the study area was dramatically reduced under the combined climate-change and urbanization scenarios. Given that the probability of occupancy was so low under these scenarios, they were not displayed in the habitat suitability maps in Figure 4. Instead, each of the scenarios was represented in terms of the average probability of occupancy and the standard deviation in Figure 5. Under the combined intermediate climate-change and urbanization scenario (scenario 7, Fig. 5), habitat that we found to be of high suitability (greater than 75% probability of occupancy) was predicted to decline in suitability, such that the average probability of occupancy will be less thatn 20%. For the ‘worst-case’ climate-change and urbanization projection (scenario 8, Fig. 5), all ponds within the study area showed a less than 1% probability of occupancy. When the impacts of climate change and urbanization on L. tasmaniensis were combined, their effects became compounded to the point where this species is unlikely to persist in the study area.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

To our knowledge, this study represents the first attempt to explore the combined impacts of climate change and urbanization on amphibians in urban-fringe environments. The impact of a range of scenarios was explored using Bayesian habitat modelling techniques, and an extensive habitat-occupancy dataset for L. tasmaniensis within the study area. The study demonstrates that L. tasmaniensis is under significant threat of local extinction on Melbourne's urban-fringe unless steps are taken to avoid habitat degradation resulting from urban growth, and efforts are made to mitigate the potential effects of climate change.

Despite the fact that the CSIRO (2005) projections, show large differences in the upper and lower limits of the predictions (Fig. 3), future climatic conditions for south-eastern Australia and Victoria are likely to include higher temperatures and considerably less rainfall than at present (Whetton et al. 2002). The predictions of the hydrological model under the climate-change scenarios suggest that water is likely to become a limited resource in the Merri Creek catchment in the next few decades, altering the hydrological regimes of wetlands, and reducing their size and depth. Ponds are likely to become increasingly dry, and under some scenarios reaching a critical threshold in terms of their ability to maintain a balance between seasonal fluctuations of rainfall and evaporation. Under these environmental conditions, frogs would be subject to unpredictable seasonality of rainfall and shorter hydroperiods.

Even though L. tasmaniensis is likely to be negatively impacted by drier conditions in the Merri Creek catchment, we cannot be certain that climate change and increasing pond desiccation will extirpate the species in this area. There is a chance that the species may adapt to the new hydrological conditions and therefore persist in the area. Also, small isolated ponds, which represent about one-third of the water bodies incorporated in the study, should not be regarded as ‘non habitat’ or uninhabitable by frogs. A study by Semlitsch and Bodie (1998) found that small, isolated wetlands can be valuable amphibian habitat. They found that these habitats can be species rich and serve as an important source of juvenile recruitment.

Many of the ponds within the Merri Creek study area could currently be considered as high-quality habitat. However, our modelling predicts that continued urban growth along the corridor is likely to have a detrimental effect on L. tasmaniensis populations. In the short term, the effects of the urbanization scenarios we explored are likely to be more significant than those of climate change. These results were not surprising, as urban infrastructure creates formidable barriers to the movement of amphibians (Vos & Chardon 1998) and roads are a major cause of mortality for populations in built-up areas (Carr & Fahrig 2001; Hels & Buchwald 2001). Some water bodies may also be impacted upon by drainage from surrounding impervious surfaces such as roads and other urban infrastructure, which often causes extreme hydrological events, toxic discharge of pollutants and nutrients and potentially brings an influx of exotic predators and competitors (Riley et al. 2005).

If urban development were to continue up to or beyond the UGB with an average density of roads (scenarios 4 and 5, Fig. 4), ponds in the study area would become significantly less suitable for L. tasmaniensis. If urban development were to continue up to or beyond the UGB with a high density of roads (scenario 6, Fig. 4), ponds in the study area would become profoundly less suitable for L. tasmaniensis. Increasing levels of urbanization will result in a deterioration of habitat quality (Hamer & Parris 2011) and will impede the movement of frogs and disrupt local metapopulation dynamics (Parris 2006). Under these scenarios, ponds are encompassed by the high levels of development and there is a less than 1% probability of them being occupied by the species.

Given that the L. tasmaniensis is widely distributed throughout Australia and inhabits relatively dry habitats, climate change alone may not be enough to cause their disappearance from the Merri Creek catchment. A study by Wilbur (1987) found that in response to drought conditions, Bufo americanus tadpoles were observed to metamorphose at a faster rate and escape from pools that had shortened hydroperiods. However, they were unable to do this when faced with the additional stresses of competition and food shortages. Under similar circumstances, L. tasmaniensis may encounter a synergistic effect from climate change and urbanization that could push them beyond their tolerance threshold, and cause local extinction of this species. The results of the combined effects of climate change and urbanization (scenario 7 and 8, Fig. 5) showed the most dramatic effect on the probability of occupancy of L. tasmaniensis in the Merri Creek catchment, with worst-case scenarios predicting a less than 1% probability of occupancy. Travis (2003) suggested that species are far more likely to suffer from climate change in a fragmented landscape, especially those that are habitat specialists and have poor colonizing ability.

Management recommendations

Predicting extinction risks for species under the combined threats of climate change and habitat loss is one of the greatest challenges facing ecologists and conservational biologists today (Travis 2003). In isolation, these factors are known to have negative effects on biodiversity. When combined, the problems become compounded, increasing the risk of extinction (Travis 2003). In the past, species may have responded to climate change by either adapting to change, shifting to more suitable habitats or going extinct (Pearson & Dawson 2003). However, adaptation may be a far more challenging process than in the past, if other threats such as human-induced land-use changes, are having an effect on a species ability to persist in a given habitat (Pearson & Dawson 2003). Species may also be limited in their ability to disperse to new habitats as a response to climate change, as migratory paths may have disappeared through habitat fragmentation and isolation (Thomas et al. 2004).

The results of this study suggest that climate change is likely to cause significant changes in hydrological regimes for ponds in the Merri Creek catchment, and that this may threaten the L. tasmaniensis populations. However, the results of this study also show that urbanization is likely to have a greater effect on the species in the short term. Therefore, it may be necessary to protect existing habitat if we are to maintain a population of this species in the study area. Proactive planning in the Merri Creek catchment could potentially mitigate the effects of urbanization on wetlands and their biotic communities. Given that connecting habitat has an important function in maintaining viable populations and metacommunities (Leibold et al. 2004), adequate protection and enhancement of existing terrestrial and aquatic habitats should be considered. Recognizing that climate change is likely to be an ongoing phenomenon, and that water is going to be in limited supply, we need to plan for a drier landscape. Greater efficiency in the gathering and storage of water, and protection of existing water supplies will be necessary. Run-off from impervious urban surfaces may also prove useful, in terms of supplying local wetlands with water. However, there would need to be robust measures to filter out toxic substances often contained in urban run-off.

Predictive models similar to those presented in this study should be used in planning activities such as future growth scenario planning, designing conservation reserves and planning restoration of terrestrial and aquatic habitats. These modelling exercises highlight the importance of recognizing that multiple factors affect species within a region, and that habitats may not remain static over time. These exercises also provide a focus for management actions through the identification of key factors that are likely to influence future habitat availability for species. It is recommended that similar predictive models are incorporated in urban planning and restoration activities to prepare for future conservation challenges.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

We would like to thank Mick McCarthy, Bill Langford, Georgia Garrard, Rory Nathen (Sinclaire Knights Merz) and Jane Elith for advice and constructive comments. Data pertaining to the study sites were collected in collaboration with Peter Robertson, Michael Scroggie and Brian Malone, and collection of these data was made possible by funding from the Victorian Department of Sustainability and Environment, the Growling Grass Frog Trust Fund (consisting of DSE, Australian Gas Limited and Friends of Merri Creek) and La Trobe University. This work was supported by an Honours scholarship from the Australian Greenhouse Office and Australian Research Council grant (LP0454979) for which there are a number of industry partners: The Department of Environment, Water, Heritage and the Arts; Port Phillip and Westernport Catchment Management Authority; Hume City Council; the City of Whittlesea; and Stockland property developers. This work forms part of the Applied Environmental Decision Analysis (AEDA) CERF collaboration supported by the Commonwealth Department of Environment and Water Resources.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES
  • Alford R. A. & Richards S. J. (1999) Global amphibian declines: a problem in applied ecology. Annu. Rev. Ecol. Syst. 30, 13365.
  • Anstis M. (2002) Tadpoles of South-Eastern Australia, A Guide with Keys. Reed New Holland, Sydney.
  • Araújo M. B. & Pearson R. G. (2005) Equilibrium of species' distributions with climate. Ecography 28, 69395.
  • Baillie J. M., Hilton-Taylor C. & Stuart S. N. (2004) IUCN red list of threatened species. A Global Species Assessment. IUCN Gland, Switzerland and Cambridge, UK.
  • Barker J., Grigg G. C. & Tyler M. J. (1995) A Field Guide to Australian Frogs. Surrey Beatty and Sons, Chipping Norton.
  • Beardsell C. (1997) Sites of faunal and habitat significance in North East Melbourne. Unpublished report for the North East Regional Organisation of Councils (NEROC). (Dunmoochin Biological Surveys, Melbourne.).
  • Bee M. A. & Swanson E. M. (2007) Auditory masking of anuran advertisement calls by road traffic noise. Anim. Behav. 74, 176576.
  • Beebee T. J. C. (1995) Amphibian breeding and climate. Nature 374, 21920.
  • Blaustein A. R., Belden L. K., Olson D. H., Green D. M., Root T. L. & Kiesecker J. M. (2001) Amphibian breeding and climate change. Conserv. Biol. 15, 180409.
  • Blaustein A. R. & Kiesecker J. M. (2002) Complexity in conservation: lessons from the global decline of amphibian populations. Ecol. Lett. 5, 597608.
  • Bouchard J., Ford A. T., Eigenbrod F. E. & Fahrig L. (2009) Behavioural responses of northern leopard (Rana pipiens) to roads and traffic: implications for population persistence. Ecol. Soc. 14, 23.
  • Brook B. W., Sodhi N. S. & Bradshaw C. J. A. (2008) Synergies among extinction drivers under global change. Trends Ecol. Evol. 23, 45360.
  • Bush J., Miles B. & Bainbridge B. (2003) Merri Creek: managing an urban waterway for people and nature. Ecol. Manag. Rest. 4, 17079.
  • Buxton M. & Tieman G. (2004) Urban Consolidation in Melbourne 1988–2003. RMIT, Melbourne.
  • Carey C. & Alexander M. A. (2003) Climate change and amphibian declines: is there a link? Divers. Distrib. 9, 11121.
  • Carr L. W. & Fahrig L. (2001) Effect of road traffic on two amphibian species of differing vagility. Conserv. Biol. 15, 107178.
  • Cogger H. G. (1992) Reptiles and Amphibians of Australia. Reed Books, Sydney.
  • CSIRO (1996) CSIRO, OzClim: A Climate Scenario Generator and Impacts Package for Australia. Commonwealth Scientific and Industrial Research Organisation, Atmospheric Research, Melbourne, Australia. [Cited 21 September 2007.] Available from URL: http://www.dar.csiro.au/res/cm/ozclim.htm
  • CSIRO (2005) Australian climate change projections for impact assessment and policy application. Climate Impact Group, CSIRO Division of Atmospheric Research, Melbourne.
  • CSIRO & Australian Bureau of Meteorology (2008) An assessment of the impact of climate change on the nature and frequency of exceptional climatic events. Exceptional circumstances report.
  • Curado N., Hartel T. & Arntzen J. W. (2011) Amphibian pond loss as a function of landscape change – a case study over three decades in an agricultural area of northern France. Biol. Conserv. 128, 23140.
  • Cushman S. A. (2006) Effects of habitat loss and fragmentation on amphibians: a review and prospectus. Biol. Conserv. 128, 23140.
  • Department of Sustainability and Environment (2004a) Urban Development Program Report.
  • Department of Sustainability and Environment (2004b) Corporate Geospatial Data Library. Land Information Group, Department of Sustainability and Environment. Level 13, 570 Bourke Street, Melbourne,Victoria 3000.
  • Ehrenfeld J. G. (2000) Evaluating wetlands within an urban context. Ecol. Eng. 15, 25365.
  • Felton A., Fischer J., Lindenmayer D. B. et al. (2009) Climate change, conservation and management: an assessment of the peer reviewed scientific fournal literature. Biodivers. Conserv. 18, 224353.
  • Frogs Australia Network (2005) Australian Frog Database. [Cited 10 August 2007.] Available from URL: http://www.frogsaustralia.net.au/frogs/display.cfm?frog_id=43
  • Gardner T. (2001) Declining amphibian populations: a global phenomenon in conservation biology. Anim. Biodiv. Conserv. 24, 120.
  • Hamer A. J. & Parris K. M. (2011) Local and landscape determinants of amphibian communities in urban ponds. Ecol. Applic. 21, 37890.
  • Hels T. & Buchwald E. (2001) The effect of road kills on amphibian populations. Biol. Conserv. 99, 33140.
  • Houlahan J. E., Findlay C. S., Schmidt B. R., Myer A. H. & Kuzmin S. L. (2000) Quantitative evidence for global amphibian population declines. Nature 404, 75255.
  • Kiesecker J. M., Blaustein A. R. & Belden L. K. (2001) Complex causes of amphibian population declines. Nature 410, 68184.
  • Lauck B. (2005) The impact of recent logging and pond isolation on pond colonization by the frog Crinia signifera. Pac. Conserv. Biol. 11, 506.
  • Leibold M. A., Holyoak M., Mouquet N. et al. (2004) The meta-community concept: a framework for multi-scale community ecology. Ecol. Lett. 7, 60113.
  • Mazerolle M. J., Desrochers A. & Rochefort L. (2005) Landscape characteristics influence pond occupancy by frogs after accounting for detectability. Ecol. Applic. 15, 82434.
  • McCarthy M. A. (2007) Bayesian Methods for Ecology. Cambridge University Press, Cambridge.
  • McMahon A. R. G. & Schulz M. (1993) Merri Creek Concept Plan Flora and Fauna. A report prepared for Merri Creek Management Committee and Melbourne Water. Ecological Horticulture, Clifton Hill.
  • Parris K. M. (2006) Urban amphibian assemblages as metacommunities. J. Anim. Ecol. 75, 75764.
  • Parris K. M., Velik-Lord M. & North J. M. A. (2009) Frogs call at a higher pitch in traffic noise. Ecol. Soc. 14, 25.
  • Pearson R. G. & Dawson T. P. (2003) Predicting the impacts of climate change on the distribution of species: are bioclimatic models useful? Glob. Ecol. Biogeogr. 12, 36171.
  • Pounds J. A., Bustamante M. R., Coloma L. A. et al. (2006) Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439, 16167.
  • Pyke C. R. (2005) Interactions between habitat loss and climatic change: implications for fairy shrimp in the central valley ecoregion of California, USA. Clim. Change 68, 199218.
  • Riley S. P. D., Busteed G. T., Kats L. B. et al. (2005) Effects of urbanization on the distribution and abundance of amphibians and invasive species in southern California streams. Conserv. Biol. 19, 1894907.
  • Rowe C. L. & Dunson W. A. (1995) Impacts of hydroperiod on growth and survival of larval amphibians in temporary ponds of Central Pennsylvania, USA. Oceol. 102, 397403.
  • Ryan T. J. & Winne C. T. (2001) Effects of hydroperiod on metamorphosis in Rana sphenocephala. Am. Midl. Nat. 145, 4653.
  • Semlitsch R. D. & Bodie J. R. (1998) Are small, isolated wetlands expendable. Conserv. Biol. 12, 112933.
  • Semlitsch R. D. & Bodie J. R. (2003) Biological criteria for buffer zones around wetlands and riparian habitats for amphibians and reptiles. Conserv. Biol. 17, 121928.
  • Smallbone L. T., Luck G. W. & Wassens S. (2011) Anuran species in urban landscapes: relationships with biophysical, built environment and socio-economic factors. Landsc. Urban Plan. 101, 4351.
  • Spiegelhalter D. J., Thomas A., Best N. G. & Lunn D. (2003) WinBUGS Version 1.4 User Manual. Medical Research Council Biostatistics Unit, London.
  • Stamm J. F., Wood E. F. & Lettenmaier D. P. (1993) Sensitivity of a GCM of global climate to the representation of land-surface hydrology. J. Clim. 7, 121840.
  • Thomas C. D., Cameron A., Green R. E. et al. (2004) Extinction risk from climate change. Nature 427, 1457.
  • Torok S. & Nicholls N. (1996) A historical annual temperature dataset for Australia. Aust Meteorol. Mag. 45, 25160.
  • Travis J. M. J. (2003) Climate change and habitat destruction: a deadly anthropogenic cocktail. Proc. R. Soc. Lond. B Biol. Sci. 270, 46773.
  • Trombulak S. C. & Frissell C. A. (2000) Review of ecological effects of roads on terrestrial and aquatic communities. Conserv. Biol. 14, 1830.
  • Urban M. C., Skelly D. K., Burchsted D., Price W. & Lowry S. (2006) Stream communities across a rural-urban landscape gradient. Divers. Distrib. 12, 33750.
  • Vos C. C. & Chardon J. P. (1998) Effects of habitat fragmentation and road density on the distribution pattern of the moor frog Rana arvalis. J. Appl. Ecol. 35, 4456.
    Direct Link:
  • Wake D. B. (1991) Declining amphibian populations. Science 253, 860.
  • Whetton P. H., Suppiah R., McInnes K. L., Hennessy K. J. & Jones R. N. (2002) Climate Change in Victoria: High Resolution Regional Assessment of Climate Change Impacts [Internet document]. Report undertaken for the Victorian Department of Natural Resources and Environment, Melbourne, and Department of Natural Resources and Environment. CSIRO Division of Atmospheric Research, Aspendale, Victoria. [Cited 14 July 2007.] Available from URL: http://www.greenhouse.vic.gov.au/climatechange.pdf
  • Wilbur H. M. (1987) Regulation of structure in complex systems: experimental temporary pond communities. Ecology 68, 143752.
  • Zhang H., Henderson-Sellers A., Irannejad P., Sharmeen S., Phillips T. & McGuffie K. (2002) Land-surface modelling and climate simulations: results over the Australian region from sixteen AMIP II models. Bur. Met. Aust. Rep. 89, 18.