The Heat Is Online

Climate Change Drives Amphibian Extinctions

U.S. Global Change Research Program Seminar Series

Amphibian Declines in the Cloud Forests of Costa Rica: Responses to Climate Change?

Is there evidence of climate change in the highland cloud forests of Costa Rica? What is the primary cause of amphibian declines and species shifts in Costa Rica's cloud forests - fungus or climate change, or both? What happens in highland, tropical cloud forests like Costa Rica under a simulated warmer climate? Are the early climate model results consistent with the notion that a climate warming can result in amphibian declines and species shifts in such settings?


Dr. J. Alan Pounds, Resident Scientist, Monteverde Cloud Forest Preserve,and Head, Golden Toad Laboratory for Conservation (GTLC), Costa Rica

Dr. Stephen H. Schneider, Department of Biological Science, Stanford University, Stanford, CA

Present and Future Consequences of Global Warming for Highland Tropical

Forest Ecosystems: The Case of Costa Rica

Highland tropical forests are an important focus of climate-change research. They are rich in endemic species and crucial in maintainingfreshwater resources in many regions. Moreover, they are highly vulnerableto global warming. Much of their remarkable diversity is due to the steepclimatic gradients found on tropical mountains. Biological communities

vary sharply along these gradients because different species are adapted for different subsets of the total range of conditions. Many species will be unable to cope if the gradients change rapidly, and long-term monitoring programs in highland tropical forests are rare. Nevertheless, data gathered by Dr.Pounds and colleagues in Costa Rica's Monteverde Cloud Forest show that dramatic biological changes are already underway. These changes are associated with climatic patterns that indicate an upward shiftin the average position of the cloud belt since the mid-1970s. Computer simulations conducted by Dr. Schneider and colleagues show that such changes in cloud formation heights can be expected to continue as greenhouse gases accumulate in the atmosphere. Thus, tropical cloud forests are endangered ecosystems, not only because of deforestation, but also because of global warming. When tracking future climate change and gauging its biological impacts, such forests should be closely monitored as early indicators of changes.

Effects of Climate Change on Birds, Reptiles, and Amphibians in the Costa Rica Highlands

The golden toad (Bufo periglenes), known only from Costa Rica's Monteverde Cloud Forest, vanished in the wake of a mysterious population crash in 1987. The disappearance of this species from seemingly undisturbed habitats caused widespread concern among scientists and others. In fact, Monteverde's entire amphibian fauna had collapsed, while reports from other mountain areas around the world told of strikingly similar accounts. Subsequent reports indicated that the unknown culprit responsible for the death of frogs in many places was a fungus, suggesting that the "amphibian crisis" was a problem peculiar to amphibians rather than a sign of more far-reaching environmental changes. The patterns at Monteverde however, do not support such a view. Long-term data strongly imply that the amphibian declines are not an isolated phenomenon. Instead, they appear to belong to a broader suite of population changes involving birds, reptiles, amphibians and others. For example, small forest lizards called "anoles" have also suffered population extinctions. Because it is highly unlikely that a fungus which attacks the moist skin of frogs would also attack reptiles, a search for a common denominator was begun.

Evidence now points to climate change, particularly a change in moisture availability. In a could forest moisture is ordinarily plentiful. Even during the dry season, which at Monteverde lasts from January through April, clouds and mist normally keep the forest wet. Trade winds, blowing in from the Caribbean, carry moisture up the mountain slopes, where it condenses to form a large cloud deck that bathes the forest. Changes in precipitation patterns are consistent with the hypothesis that a climate warming, particularly since the mid-1970s, has raised the average altitude at which cloud formation begins thereby reducing the clouds' effectiveness in delivering moisture to the forest.

Observational evidence and instrumental records indicate that the incidence of days without mist during the dry season has quadrupled over recent decades. Although El Niño warm episodes exert a drying of their own, there is evidence for a longer-term drying trend which appears to be operating in the background. The combination of El Niño and this underlying long-term drying trend resulted in major climatic extremes (drying) in 1983, 1987, 1994, and 1998.

Observed biological changes are also consistent with evidence of an elevated cloud base, and attest to the importance of these extremely dry years. The principal response by birds is an upslope movement of cloud-forest-intolerant species. Prior to the recent climate change, these species nested only in premontane habitats farther downslope. At 1,540 meters elevation, the number of these premontane species present has increased at a rate of about 19 species per decade, while 15 species have established breeding populations. The rate of colonization has fluctuated in virtual lockstep with climate.

Biological events characterized by upslope movements have followed periods of reduced mist frequency. These dry periods have also affected high-elevation anole (lizard) species. At 1,540 meters, the two previously most common species had begun to decline by the late 1980s, and had disappeared by 1996. Changes in their abundance are again correlated with variation in mist frequency. The amphibian declines, although more episodic than the anole declines, are also associated with these mist-frequency patterns. The 1987 population crash, which led to the disappearance of the golden toad, began during the driest period on record - the same event that stimulated the first major upslope movement of premontane birds. After the crash, surviving frog populations underwent synchronous downturns in 1994 and 1998. Thus, three demographic events in 1987, 1994, and 1998, correspond to the three largest climatic extremes on record in this region. It is highly unlikely that all three events would, by chance alone, correspond to these extremes.

Observed changes in populations of birds, lizards, and frogs are consistent with one another, yet differ in important ways. All are associated statistically, with the same climatic changes, and occurred simultaneously, implying that all are components of a single phenomenon. Hence, the golden toad, which has been missing for a decade, may become known as the first species whose extinction was attributed to global warming. The diversity and complexity of population changes suggest that climate has orchestrated them through several different chains of events, many of which remain poorly understood. In the case of amphibians, climate-linked epidemics are one likely mechanism of population declines. It is well known that climate variability influences host-parasite and disease-vector relationships. The diversity of responses also underscores how difficult it is to predict the biological consequences of climate change. Thus, one can only make simple predictions about changes in species distribution and abundance. However, once such changes begin to take place and ecological interactions change as a result, the outcome is likely to be as unpredictable as it is profound.

Simulated Cloud Forest Responses to a Warmer World

In the past several years there has been growing evidence that both climatic changes and the impacts of climatic changes may have abrupt, "non-linear" characteristics. Examples of the former include "flip-flops" in North Atlantic Ocean currents or rapid disintegration of the West Antarctic Ice Sheet, whereas examples of the latter could include a rapid disturbance to existing forests from increased fire frequency or a change in the altitude of cloud formation in cloud forests. Although greater uncertainties typically accompany the anticipation of such events, the consequences of such non-linear behaviors are much more problematic since they may be either irreversible or difficult to adapt to because of their abrupt nature. The case of tropical mountain cloud forests is examined below.

Tropical mountain cloud forests occur where mountains are frequently enveloped by trade-wind-derive rographic (mountain) clouds and mist in combination with convective rainfall. Many features of these forests are directly or indirectly related to cloud formation (i.e., vegetation morphology, nutrient budgets, solar insolation). One of the most direct impacts of frequent cloud cover is the deposition of cloud droplets in the form of horizontal precipitation (HP). In systems such as cloud forests, total horizontal precipitation is greater than that from vertical rainfall events during the dry season when such forests can experience water stress.Thus, such forests function as important local and regional watersheds.

Model Results

Theory and decades of modeling suggest that the dynamics of an enhanced hydrological cycle due to a global warming could influence the height at which orographic clouds form in tropical mountain forests. In addition,evidence derived from the analysis of pollen strongly suggests a downslope shift in the range of some current cloud forest species during the last glacial period (suggesting conversely, that a climate warming might shift species ranges upwards). General Circulation Model results applied to the location of four cloud forest sites around the tropics show that the location of the relative humidity surface (RH), a proxy for cloud height, is consistently shifted upwards in the northern hemisphere winter (dry) season (December, January, and February) at all of the sites in a simulated, warmer climate characterized by a doubling of the concentration of atmospheric CO-2. In addition, a downward shift in the relative humidity surface is obtained for the northern hemisphere summer season (June, July, and August) at these same sites. The location of the relative humidity surface in the case of the Monteverde locale, for example, does suggest a rise in winter cloud height of over 200 meters, in the case of a warmer, doubled CO-2 atmosphere. This timeframe represents part of the dry season when the Monteverde cloud forest relies most heavily on the horizontal precipitation from cloud mists. Such a rise would likely be of biological and hydrological significance to the cloud forest's structure and function. As an alternative to the RH surface height proxy, which is an indirect method to predict cloud heights to assess the impact of climate change on cloud forests, biogeography models were also employed to predict the location of cloud forests in simulated climates. These models predict ecosystem locations using specified ecosystem tolerance cutoffs for temperature and moisture. An important temperature variable used in such analyses is the warmth index (WI) - the sum of all monthly mean temperatures exceeding 5 degrees C. The WI is found to correlate broadly with forest type. Model results employing biogeography show that all fourexisting , but widely separated cloud forests, experience an increase in the warmth index and absolute humidity when surface temperatures are warmer, as well as a decrease in these variables when it is coolere.g., ice-age simulations).

Warmer surface temperatures associated with model simulations involving a doubled CO-2 atmosphere provoke an increased exchange of moisture from the surfaces of plants and leaves as well as from warmer oceans. As a consequence of this outcome, the altitude at which some absolute humidity surface would occur in the experiment was expected to rise (model results show a rise of some 300 meters) in the simulation involving a doubling of atmospheric CO-2 (a warmer world). Likewise, in the case of a model experiment involving the peak of the last ice age, the results indicated a descent in the absolute humidity surface of roughly 200-500 meters. These model results suggest that both the temperature and the moisture conditions of the present cloud forest at Monteverde will be shifted upward in altitude in a warmer world. The observational data of Dr. Pounds and his colleagues, pointing to invasions of sub-montane species into cloud forest habitat near Monteverde, Costa Rica, provide further evidence of the climate sensitivity of these ecosystems. Most notably, these researchers have rejected habitat destruction pressures at lower elevations as a cause of this upslope migration of species. Consequently, this and other cloud forests may be experiencing the dual stresses of changing microclimates and invading species from lower elevations driven in part by changes in the height of orographic cloud bank formation in the dry season and/or increased evapotranspiration.

In light of these combined field observations and preliminary modeling results, further research into climate change impacts on cloud forests is essential. However, the implications of even these crude model results suggest that climate change will likely affect the distribution of the potential locations for cloud forests. And if the above analysis proves to be even modestly robust in the face of additional testing, it may indicate that those species situated near mountain tops are likely to be forced out of existence by a climate warming. It is precisely such non-linear relationships evident in the Monteverde cloud forests, where one can anticipate the greatest impacts of climatic changes, whether they are smoothly varying or abrupt.


Dr. J. Alan Pounds is Resident Scientist at the Monteverde Cloud Forest Preserve in Costa Rica. The Preserve is owned and operated by the Tropical Science Center, a nonprofit scientific and educational organization based in San José. Dr. Pounds also heads the Golden Toad Laboratory for Conservation (GTLC) located at the Preserve, and manages the John H.Campbell Weather Station. He is also an Adjunct Professor in the Department of Biology, University of Miami.

Dr. Pounds' research interests focus on tropical ecology and conservation, and particularly the biological consequences of climate change. His goal is to gauge the extent to which global warming is an immediate threat to highland biological communities in the tropics and to develop a mechanistic understanding of the observed impacts on populations. Dr. Pounds completed his graduate studies at the University of Florida in 1987. He received his Ph. D. in population and community ecology, focusing on the communityecology of anoline lizards at Monteverde.

Dr. Stephen H. Schneider is a professor in the Department of Biological Sciences, a Senior Fellow at the Institute for International Studies, and Professor by Courtesy in the Department of Civil Engineering at Stanford University. At Stanford University he teaches classes and courses in departments such as Earth Systems, Civil Engineering, Biological Sciences, and economics; he also teaches a Senior Honors Seminar in Environmental Science, Technology and Policy.