Will Climate Change Reduce the Efficacy of Protected Areas for Amphibian Conservation in Italy?

Worldwide amphibian declines have been attributed to various interacting factors that include habitat loss and degradation, UV radiation, disease, and climate change. Recently, amphibian population disappearances within Italy have been associated with increasing temperatures, placing these species at great risk to climate change induced-extinction. Therefore, in situconservation efforts such as integrating potential climate change impacts with the selection of protected areas within Italy are imperative to protecting amphibians. D’Amen et al. (2011) used a niche modeling estimate of potential amphibian range shift under two IPCC climate change scenarios and two dispersal assumptions to analyze amphibian extinction risk within Italy and to analyze the efficacy of the current Italian reserve network for protecting amphibian diversity. These predicted distributions were additionally used to perform gap and irreplaceability analyses to identify unprotected areas that could contribute to amphibian species conservation in the future. The results indicate that range alterations should be expected for all amphibian species under future climatic scenarios and that these range modifications will determine the degree to which species are represented within the Italian protected area network.  Additionally, the results show that the existing protected area network incompletely represents total amphibian diversity and its geographic pattern, which ultimately will cause a decrease in species representation over the entire protected area system with climate change. Inclusion of the authors’ suggested irreplaceable regions within the country could improve the future efficiency of the Italian protected area system.—Megan Smith

D’Amen, M., Bombi, P., Pearman, P.B., Schmatz, D.R., Zimmermann, N.E., Bologna, M.A., 2011. Will Climate Change Reduce the Efficacy of Protected Areas for Amphibian Conservation in Italy? Biological Conservation, 144: 989-997.

The authors collected amphibian presence data from the CKmap 5.3.8 database, which reports species occurrence within the Universal Transverse Mercator (UTM, 10 X 10) grid that intersects the Italian region. Newly recognized amphibian species’ distribution information was updated using maps from the IUCN Red List. To avoid small sample bias, species were excluded from the study if there were fewer than 20 records for these amphibian species. Alien species, cave species, and species only marginally present in Italy were also excluded. In total, the authors analyzed 22 species, 20 of which were ranked by the European Red List of Amphibians in the Low Concern Category.
D’Amen et al. calibrated niche-based models using climatic, land use, and topographical predictors. Of 20 predictor variables, only those that were correlated with a value of 0.70 or lower were used. The seven bioclimatic predictors used were annual mean temperature, mean diurnal temperature range, isothermality, temperature annual range, mean temperature of wettest quarter, precipitation of warmest quarter, and precipitation of coldest quarter. Their values were derived from the WorldClim database, which are climate grids for 1950–2000. Potential future climate values during 2041–2070 were obtained from climate grids of a previous study for IPCC scenarios A1F1 and B1 and from the HadCM3 circulation model. This model is commonly used for predicting climate change effects on fauna distribution in Europe. The authors selected A1F1 and B1 climate scenarios because they are based on contrasting story lines that cover a range of possible demographic, socio-economic and technological changes that are believed to affect green-house-gas emissions. The future climate scenarios were expressed as anomalies of past climate scenarios, interpolated, and then recombined with the grids from the WorldClim data set to produce high-resolution climate scenarios. The effects of existing land use on habitat availability, as well as topographic information, was included in the models.
Information on the location of existing protected areas in Italy was obtained from the National Ministry for the Environment. Protected areas within Italy consisted of Nationally Designed Protected areas (NPAs) and sites included in the European Natura 2000 network. The NPAs were composed of 774 parks founded by national or local administrations before 2004. They covered a surface area of 29,400 km2. The Natura 2000 network was composed of 2885 sites that largely overlap with NPAs. Combined, they increased the protected area in Italy by 34,700 km2. Natura 2000 areas that didn’t overlap with NPAs were defined as EPAs. The Overall Protected Areas (OPAs) covered a total surface area of 64,100 km2.
The authors experienced problems matching reserve boundaries with species distribution. Therefore, they used a threshold to determine whether reserves were considered present or absent in a grid cell. Numerous threshold values were tested (from 0% to 100%) and the value that resulted in the selection of a number of cells with a total surface equal to the total surface of Italian protected areas was used. Ultimately, the threshold was defined as any cell with a proportion of park coverage larger than 40%.
Amphibian species distributions were modeled using 8 different techniques within the R-based BIOMOD package. These models included Generalized Linear Models, Generalized Additive Models, Classification Tree Algorithms, Artificial Neural Networks, Mixture Discriminant Analysis, Multivariate Adaptive Regression Splines, Generalized Boosted Regression models, and Random Forest. The models were evaluated for their species distribution performance and consensus and those with the highest validation scores were included in the study. Since individual dispersal capability could restrict the ability of a species to geographically track suitable climatic conditions, species-specific dispersal limitations were considered. Additionally, the authors included a no-dispersal scenario since species range shifts are limited by extrinsic factors such as a highly fragmented landscape. These dispersal scenarios were used to estimate the proportion of current amphibian habitat that remained suitable under future climate conditions. Potential distributional shifts were calculated as the difference in the total number of grid cells currently occupied and those occupied under each of the future climate change scenarios.
Conservation targets for species conservation were used in both the gap and irreplaceability analyses. They were calculated by examining the species-specific extent of occurrence. These targets were later adjusted based on the modifications of the species range sizes under the future climatic scenarios. 1000 km2 (10 cells) was defined as the minimum area needed for a species’ viability. If species occupied area less than 10 cells, the conservation target was set to 100%. The conservation target of widespread species (those that occupied more than 1000 cells) was set at 10%. Conservation targets for species with intermediate sized ranges were determined by interpolating the extreme range size targets within a linear regression on the log transformed number of initially occupied cells.
A gap analysis was used to determine the extent of a species representation within the Italian reserve network by comparing amphibian species’ distributions to the distributions of conservation areas. Species that were not represented in any protected area were defined as gap species and species that met a portion of their conservation target were considered to be partial gap species.
An irreplaceability analysis was used to measure the degree to which a cell was required in a reserve network to obtain the species-specific defined conservation targets. Irreplaceability was calculated as the number of combinations of sites that included the focal site and met conservation targets, but when the focal site was removed, the conservation targets would not be met. Current species occurrences and their future potential distributions under the different climate and dispersal scenarios were used to predict the irreplaceability of each cell. Three alternative conservation systems were considered (no reserves, NPAs, and OPAs) to assess the contribution of the existing reserve network.
Finally, the authors analyzed the efficacy of the existing Italian national park system (OPAs) and its components (NPAs and EPAs). They compared the mean irreplaceability value of cells in conservation networks to the mean value expected in cells randomly chosen regardless of their conservation status. The authors also tested whether OPAs and NPAs have higher irreplaceability than the remaining map cells, excluding the protected areas. The irreplaceability values of the new Natural 2000 sites were also calculated and incorporated into the analysis. Irreplaceability values were additionally recalculated considering the existence of NPAs for testing EPAs selection for current and future conservation.
A table displaying the current range extent and percentage of predicted change in the future climatic conditions under different dispersal assumptions was constructed, as was a second table displaying the percentages of target met by each species in NPAs and OPAs under present conditions and alternative dispersal and climate change scenarios. A figure showing the number of gap and partially gap species in the current conditions and in the future scenarios considering only NPAs and OPAs was constructed, as was a figure demonstrating the future irreplaceability patterns within Italy. Finally, a third table presenting the results of the non-randomness test of differences between the mean irreplaceability values in protected cells and in random cell selections was constructed.
The authors found that under a no-dispersal assumption, all but two species were projected to lose suitable habitat in the future climate scenarios. Although there was no statistically significant result, species range losses were generally higher under the A1F1 scenario. With dispersal ability, the models demonstrated that 50% of the species’ range sizes would reduce by 60% under both climate scenarios. Eight species experienced a range reduction regardless of the climate change and dispersal scenarios. One species lost all suitable habitat under the A1F1 and no dispersal scenarios while four species experienced an increase in range size under the dispersal assumption.
Using the 40% threshold value, NPAs and OPAs occupied 282 cells and 617 cells total. All species were present in both the NPAs and OPAs. Nine species’ conservation targets were met in the OPAs, but none were met when the analysis was restricted to NPAs. More than half of the partial gap species met less than 50% of their conservation targets and the predicted number of species that met conservation targets decreased in the future, irrespective of dispersal or climate change scenario. With NPAs, some species were predicted to become gap species and therefore disappear entirely from the currently protected cells. With OPAs, the number of species that met their conservation targets was projected to decrease. One species is projected disappear completely from the OPAs under the no-dispersal assumption.
Under the assumption of no protected cells, the areas with the highest values of irreplaceability for amphibian conservation were the island of Sardinia and the lowlands of Northeastern continental Italy. Secondary regions included Sicily and the Tyrrhenian side of Southern Italy. Included NPAs within the gap analysis reduced the irreplaceability scores for the grid cells within the Italian peninsula and on the Tyrrhenian side of Southern Italy.
Sardinia and Sicily had high irreplaceability values under both future climate scenarios and with no cells considered protected. The cells on the Tyrrhenian side of Southern Italy, the cells in the mountainous areas of central and eastern Alps, and the cells in the central Apennine increased in irreplaceability as well. If all protected areas were considered present, the irreplaceability maps calculated for future conditions signified the focal areas that should be designated as new reserves for the long-term conservation of amphibians. These areas included Sardinia, the lowlands of Northeastern continental Italy, the central Alpine foothills, Sicily, and the Tyrrhenian side of Southern Italy.
Additionally, the comparison of mean irreplaceability values of protected map cells with values calculated from 5000 sets of randomly selected map cells demonstrate that the Italian network reserve (OPAs) and its components (NPAs and EPAs) protected sites with greater irreplaceability than that expected by chance. These results were constant under the future climate scenarios and under the no-dispersal assumption. With dispersal and under both climate scenarios, cells within NPAs were not more irreplaceable in the future than randomly selected cells, and cells within EPAs were only slightly more irreplaceable than the randomly chosen, unprotected cells. However, the reserve system as a whole had a larger irreplaceability value than the unprotected cells under future climatic scenarios.
The results show that the extent of species’ range modifications under future climatic scenarios is dependent on dispersal ability. With dispersal, the models demonstrate that many species will be able to shift their ranges eastward and northward. Unfortunately, many amphibians may be unable to do so due to the highly fragmented landscape within Italy. Therefore, the no-dispersal assumption is the most realistic and under this assumption, and 70% of amphibians’ predicted habitat is likely to be lost under future climatic scenarios.
The species most sensitive to climate chare are P. fuscus, Salamandrina terdigitata, Salamandra atra, and Triturus carnifex. These species are either endemic or subendemic, but are classified as “Least Concern” in the most recent Red List of European Amphibians. Since a loss of distributional area is a good predictor of extinction risk, the results suggest that these species may become extinct within the middle of the current century. One subspecies in particular, P. fuscus insubricus, is endemic to the Po river plain of northern Italy and is recognized as a highly threatened taxon. By the mid 21st century, this subspecies may lose all suitable habitat under the assumption of no-dispersal. It is already difficult for this toad to navigate through its environmental matrix because of anthropogenic impact in the river plain. Therefore, the creation of corridors could allow this toad to disperse to areas with appropriate climate, assisting in its conservation.

Overall, the study indicates that all amphibian species will experience range modifications under the future climatic scenarios and that these modifications will determine the extent to which these species will be represented in the Italian protected area network. The current reserve network does not represent the entirety of amphibian diversity or its geographic pattern, decreasing the species’ future representations within the reserve system. The reserve system would be improved as a whole if the study’s suggested irreplaceable areas (Sicily, Sardinia, and Northeastern Italy) were included within the network. 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s