Modern amphibians are the oldest extant group of terrestrial vertebrates, having existed for hundreds of millions of years¹. Some have undergone very little evolutionary change over this period. But amphibians are far from being primitive animals of low diversity. They are, in fact, more diverse than mammals or reptiles — 6,894 species are presently known², and the best estimate³ for the total number of species is 8,000–12,000. They occupy every terrestrial habitat except Antarctica and the high Arctic, and have an important role in nutrient dynamics, in the cycling of energy flows between terrestrial and freshwater systems, and in controlling populations of pest insects¹.

Amphibians are also the terrestrial vertebrates most at risk of extinction: at least 37% are classed as vulnerable, threatened or endangered⁴ (Fig. 1). This figure is certainly an underestimate of the risk, because the conservation status of about 25% of species remains unknown⁴. Sadly, things are likely to be much worse by 2080, as Hof et al.⁵ report on page 516 of this issue.

The armoured mistfrog Litoria lorica was thought to have been extinct for 17 years, but was rediscovered¹⁴ in 2008. Many other amphibian species are in decline, a situation that Hof et al.⁵ predict will accelerate during the twenty-first century.

R. A. ALFORD

• Full size image (133 KB)

Ever since amphibian declines were recognized as a global problem⁶, it has been suggested that the phenomenon has several causes. Early explanations focused on pollution, habitat loss and climate change⁷, but many declines have occurred in relatively pristine areas not strongly affected by habitat modification, and far from sources of environmental contaminants. In these cases, analyses have indicated that climate change alone is unlikely to be the reason⁷.

A major cause of many of these enigmatic declines was discovered⁸ in 1998. The pathogenic fungus Batrachochytrium dendrobatidis causes chytridiomycosis, a disease that infects amphibians across a wide range of taxa^{8, 9}. This disease has emerged in many areas, and has caused entire regional faunas to decline dramatically, to the point of local extinction. According to the International Union for Conservation of Nature, many amphibian species became critically endangered between 1980 and 2007, in most cases probably because of the global chytridiomycosis pandemic⁷. Habitat loss and climate change have received less attention as causes of amphibian decline, but threaten many more species in the long run⁷. Climate change may already be affecting the course of the chytridiomycosis pandemic, and its influence is likely to continue during this century, because vulnerability to the disease is strongly affected by weather and climate^{7, 10}.

Given the current threats to amphibians, it is crucial to develop plans to protect their diversity. This requires a good understanding of present and future dangers. Hof et al.⁵ have undertaken the complex task of examining the spatial extent of three of the major threats to the global diversity of amphibians — disease, climate and land-use change. By comparing their findings with the spatial distribution of amphibian diversity, they were able to say how the threats have changed and will probably change in the period from 1980 to 2080. The authors did this for all three orders of modern amphibians, which differ in their levels of ecological and taxonomic diversity: roughly 90% of modern amphibians are frogs, 8% are salamanders and 2% are the less well-known caecilians¹. Frogs are far more widely distributed than the other groups.

Hof et al. used a complex, wide range of modelling approaches in their study. Briefly, they used data on the distributions of 5,527 amphibian species in bioclimatic models to predict the global distribution of the species on a latitude–longtitude grid consisting of cells 2° × 2° in size. This analysis took into consideration a broad range of future climate scenarios proposed by the fourth Intergovernmental Panel on Climate Change. To forecast the spread of chytridiomycosis, they used a previously published model¹¹ that predicted the distribution of the causative fungus B. dendrobatidis. Their data on land use and land-use changes came from the Millennium Ecosystem Assessment, a report on the current state and the future of Earth's ecosystems.

According to Hof et al.⁵, the outlook for amphibians is not good. For frogs — the most diverse group — the areas most affected by climate change coincide with regions of greatest species richness. The authors' models indicate that, in some of the regions with the greatest diversity of frogs, more than half of the species will probably be negatively affected by climate change by 2080. Strong climate-change impacts are also likely for some salamanders, particularly tropical faunas.

The models also suggest that land-use changes, especially in tropical regions, are likely to have strong negative effects on amphibians in some of the areas that have high levels of amphibian diversity. Finally, they predict that the distribution of B. dendrobatidis, and thus possibly of chytridiomycosis, will be focused in temperate and mountainous areas. This is better news for frogs, which reach their peak diversity in the lowland tropics, but may be bad news for salamanders, whose centre of diversity is in northern temperate regions.

Possibly the worst news is that, on the whole, the areas most affected by each category of threat do not coincide geographically: less than half of the grid cells in the 25% of land most threatened by any one factor are also in the 25% most threatened by any other factor. Because the threats are spread out, more than half of the total geographic distribution of each major amphibian taxon is in areas that Hof et al. predict will be highly affected by at least one of the three threat factors by 2080. The picture becomes worse when only the most diverse faunas are considered — roughly two-thirds of the areas that have the highest diversities of frogs and salamanders are likely to be highly threatened in some way.

The effects of major changes in land use will probably be as strong as, or even stronger than, Hof and colleagues assume, because the complex life histories of amphibians may render them particularly vulnerable to the disruptive effects of habitat modification¹². But in other respects, the exceedingly gloomy picture presented by the authors might turn out to be too pessimistic. For example, the exact effects of climate change and chytridiomycosis on amphibians are not known, and so their overall impact may be less than is predicted⁵. The somewhat coarse grid used in Hof and colleagues' bioclimatic modelling might also obscure small-scale variations that could allow species to avoid the negative effects of climate change by shifting their habitat ranges relatively short distances, or simply by changing how they use their existing ranges (for example, by choosing less exposed retreat sites)^{5, 7}. Moreover, the authors' analysis equates the presence of B. dendrobatidis to negative conservation effects of chytridiomycosis, but the impact of the disease varies strongly among regional faunas, ranging from disastrous population collapses in some areas to little or no effect in others^{9, 13}.

On the other hand, some of Hof and co-workers' results may be overly optimistic. For example, they did not model possible non-additive impacts of threats, such as the strong possibility that the threat of epidemic outbreaks of chytridiomycosis may worsen with changing climate^{7, 10}, or that habitat modification may restrict amphibians' ability to resist climate change by altering their habitat preferences^{5, 7}. Nevertheless, their work is a valuable step towards a true understanding of overall threat levels to an iconic group of animals. It is also a sobering reminder of how much critical information is needed before we can truly understand the extent of anthropogenic threats to global biodiversity, or be fully prepared to rationally manage them.