As the Earth warms, many species are likely to disappear, often because of changing disease dynamics. Here we show that a recent mass extinction associated with pathogen outbreaks is tied to global warming. Seventeen years ago, in the mountains of Costa Rica, the Monteverde harlequin frog (Atelopus sp.) vanished along with the golden toad (Bufo periglenes). An estimated 67% of the 110 or so species of Atelopus, which are endemic to the American tropics, have met the same fate, and a pathogenic chytrid fungus (Batrachochytrium dendrobatidis) is implicated. Analysing the timing of losses in relation to changes in sea surface and air temperatures, we conclude with 'very high confidence' (> 99%, following the Intergovernmental Panel on Climate Change, IPCC) that large-scale warming is a key factor in the disappearances. We propose that temperatures at many highland localities are shifting towards the growth optimum of Batrachochytrium, thus encouraging outbreaks. With climate change promoting infectious disease and eroding biodiversity, the urgency of reducing greenhouse-gas concentrations is now undeniable.

Recent warming has caused changes in species distribution and abundance 1±3 , but the extent of the effects is unclear. Here we investigate whether such changes in highland forests at Monteverde, Costa Rica, are related to the increase in air temperatures that followed a step-like warming of tropical oceans in 1976 (refs 4, 5). Twenty of 50 species of anurans (frogs and toads) in a 30-km 2 study area, including the locally endemic golden toad (Bufo periglenes), disappeared following synchronous population crashes in 1987 (refs 6±8). Our results indicate that these crashes probably belong to a constellation of demographic changes that have altered communities of birds, reptiles and amphibians in the area and are linked to recent warming. The changes are all associated with patterns of dry-season mist frequency, which is negatively correlated with sea surface temperatures in the equatorial Paci®c and has declined dramatically since the mid-1970s. The biological and climatic patterns suggest that atmospheric warming has raised the average altitude at the base of the orographic cloud bank, as predicted by the lifting-cloud-base hypothesis 9,10 .This hypothesis builds on evidence that rising sea surface temperatures (SSTs) have altered the climates of tropical mountains. Enhanced evaporation from warm ocean surfaces has generated large amounts of water vapour, and latent heat released as this moisture condenses has accelerated atmospheric warming 5 . Because vertical thermal pro®les have tended towards a moist adiabatic lapse rate, the decline in temperature with increasing elevation has diminished, amplifying the warming in the highlands relative to the lowlands 11±13 . Freezing heights have shifted upwards 11 , and glaciers on high tropical mountains are rapidly melting 14 . If temperature-dependent relative humidity surfaces, and thus cloudformation heights, have likewise shifted upwards 10 , organisms may be affected in various ways. Monteverde's stratus±stratocumulus bank, which forms as the trade winds meet the Caribbean slope of the Cordillera de Tilara Ân,¯ow upwards and cool adiabatically, in¯u-ences several key ecological processes. A lifting cloud base should alter regional hydrology by reducing critical dry-season inputs of mist (low-intensity windblown precipitation) and cloud water (non-precipitating droplets deposited onto vegetation) 15,16 .To examine climate trends, we have analysed patterns of precipitation, stream¯ow, air temperatures and SSTs. The rainfall and air-temperature data (collected by J.H.C.) are from leeward cloud forest (1,540 m; ,1 km west of the Monteverde Cloud Forest Preserve headquarters and ,3 km west of the continental divide). The weather station lies on the western boundary of our 30-ha study plot for anoline lizards, which overlaps a 40-ha plot for birds. Both plots lie within the 30-km 2 anuran study area. The stream-¯ow data (from the Costa Rican Electrical Institute) are for the Rõ Âo Can Äas at Lõ Âbano (300 m; ,23 km northwest of Monteverde). The SST data (from NOAA) are ...

Many of the recent, widespread declines and disappearances of amphibian populations have taken place in seemingly undisturbed, montane habitats. The question of whether the observed patterns differ from those expected from natural population dynamics is the subject of an ongoing controversy with important implications for conservation. We examined this issue for the Monteverde region of Costa Rica’s Cordillera de Tilarán, where a multi‐species population crash in 1987 led to the disappearance of the endemic golden toad (  Bufo periglenes) and many other species. Focusing on long‐term studies of other amphibian assemblages, we developed probabilistic null models for the number of disappearances. Tests of these models at Monteverde suggest that the patterns observed there are highly improbable in the context of normal demographic variability. Twenty species of frogs and toads (40% of the anuran fauna) were missing throughout our 1990–1994 surveys of a 30‐km2 area. Not all organisms in this area had declined accordingly: the relative frequency of absences was much greater for anurans than for breeding birds. Nevertheless, anuran habitats, most of which are protected within the Monteverde Cloud Forest Preserve, seemed unchanged, and none of the breeding‐bird species known to be sensitive to deforestation was missing. Thus, only factors other than direct, obvious human impacts can explain the amphibian declines. Consistent with our tests of null models, analyses of recent population trends do not support the hypothesis that the 1987 crash was an extreme fluctuation from which populations are recovering. Surviving species for which baseline data are available—stream‐breeding glass frogs (  Hyalinobatrachium fleischmanni and Centrolenella prosoblepon) and a pond‐breeding treefrog (  Hyla pseudopuma)—remained far less abundant than they were before the crash and showed no increase during 1990‐1994. We documented an increase only for one terrestrial‐breeding rain frog ( Eleutherodactylus diastema).Pruebas de Modelos Nulos para Disminuciones de Anfibios en una Montaña TropicalMuchas de las disminuciones y desapariciones recientes de poblaciones de anfibios en varias partes del mundo se han producido en hábitats que aparentemente no han sido alterados. La pregunta de si los patrones observados difieren de lo que es predicho por la dinámica natural de poblaciones es el tema de una controversia actual que tiene consecuencias importantes para la conservación. Se examinó esta pregunta para la región de Monteverde en la Cordillera de Tilarán, Costa Rica, donde un colapso de poblaciones en 1987 produjo la desaparición del endémico sapo dorado (  Bufo periglenes) y muchas otras especies. Centrándose en estudios a largo plazo sobre anfibios de otras regiones, se desarrollaron modelos nulos probabilísticos con respecto al número de desapariciones. Al probarse estos modelos para Monteverde se sugiere que los patrones observados son poco probables dentro de los parámetros normales de la variabilidad demográfica. Durante 1990‐1994 en un área de 30‐km2, veinte especies de ranas y sapos (el 40% de los anuros de la región) estuvieron ausentes. No todos los organismos del área disminuyeron de la misma forma: por ejemplo, la frecuencia relativa de especies ausentes fue mucho mayor para los anuros que para las aves que se reproducen en el área de estudio. Sin embargo, la mayoría de los hábitats de anuros están protegidos dentro de la Reserva Biológica Bosque Nuboso de Monteverde, y no parecía que habían cambiado. Además, ninguna de las especies de aves que son afectadas de manera negativa por la deforestación estuvo ausente. Por lo tanto, las disminuciones de anfibios sólo pueden ser explicadas por factores que no son los impactos obvios y directos causados por los seres humanos. De acuerdo con nuestras pruebas de modelos nulos, los análisis de tendencias recientes de abundancia no apoyan la hipótesis de que el colapso de 1987 fuera una fluctuación extrema de la cual las poblaciones están recuperándose. Las especies sobrevivientes de las que existen datos demográficos anteriores al colapso—ranas de vidrio ( Hyalinobatrachium fleischmanni y Centrolenella prosoblepon), que se reproducen en quebradas, y una rana arborícola ( Hyla pseudopuma) que pone huevos en lagunas y pozos—eran mucho menos abundantes durante 1990–1994 de lo que fueron antes de este evento y no se encontró evidencias de aumento. Se documentó un incremento solo en las poblaciones de Eleutherodactylus diastema, que se reproduce en hábitats terrestres.

No abstract