OOS 43-8
Rainforests in the clouds: Decadal trends in rain and fog inputs to relict forests in semiarid Chile

Thursday, August 14, 2014: 4:20 PM
304/305, Sacramento Convention Center
Juan J. Armesto, Ecology, Universidad Católica de Chile, Institute of Ecology and Biodiversity, Santiago, Chile
Aurora Gaxiola, Ecologia, Universidad Católica de Chile, Instituto de Ecologia y Biodiversidad, Santiago, Chile
Olga Barbosa, Institute of Environmental Sciences & Evolution, Universidad Austral de Chile, Valdivia, Chile
Fernanda Vio, Ecology, Universidad Católica de Chile, Institute of Ecology and Biodiversity, Santiago, Chile
Fernando D. Alfaro, Ecología, Pontificia Universidad Catolica de Chile, Instituto de Ecología y Biodiversidad, Santiago, Chile
Mariela C. Nuñez-Avila, Institute of Ecology and Biodiversity, Santiago, Chile
Pablo A. Marquet, Santa Fe Institute, Santa Fe, NM
Background/Question/Methods

Cloud cover is a major and nearly permanent feature (80% persistence) of the eastern Pacific Ocean in the arid and semiarid continental margin of South America.  The cloud field is maintained at a fixed elevation (above 600 m) by a thermal inversion, providing the main or the only source of humidity for a chain of highly vegetated hotspots found along coastal hills from northern Peru to the semiarid margin of the Atacama Desert in Chile. In semiarid Chile (30 S, Fray Jorge Forest), a chain of isolated rainforest fragments occurs on coastal mountaintops inundated by oceanic fog, receiving only sporadic rainfall (annual mean=14 cm). Because fog is essential for the persistence of forest patches and their rich floristic assemblage, closely related to austral temperate rainforests, we initiated a quantitative study of decadal trends in cloud and rainwater inputs.  We were interested in the constancy of fog inputs and their temporal synchrony with rainfall, in the face of variable and declining rainfall. For a period of 10 years, we kept monthly records of water inputs originated from passive cloud interception, stemflow, throughfall, and direct precipitation outside and inside six patches varying in size and at location in the semiarid landscape. 

Results/Conclusions

Rainfall varied greatly among years, influenced by El Niño Southern Oscillation (ENSO), with unusually dry extremes of 10 mm, and unusually wet years reaching >200 mm. Rainfall declined by 12% in decade of study, despite a wet ENSO event in 2002. Peak rainfall is concentrated in the midst of the austral winter, July-August, followed by nearly 8 months of spring-summer drought. In contrast, fog inputs peaked in late spring and summer (September-February) providing moisture to the forest during the rainless period. We estimate that canopy interception and stemflow contributed an additional equivalent to 300 mm. The persistent cloud cover and low air temperatures (10-15 degrees less than the surrounding semiarid shrubland), severely limit losses from evaporation and transpiration, improving water economy at tree and ecosystem levels. Only small differences in water capture were recorded between small (<1 ha) and large (20-30 ha) patches, suggesting that all fragments are capable of similar fog water capture, but small patches have a less efficient water economy due to strong edge effects, increasing their fragility to drying trends. Fog interception showed no significant trend for the decadal period studied, which suggests that fog may compensate current declines in rainfall.