Amazon forest response to droughts, fire, and land use: a multi-scale approach to forest dieback

Carnegie Institution Of Washington, Washington DC

Investigators

Abstract

Amazon forests make up a large fraction of all tropical forests on Earth. These forests influence the global carbon and water cycles and, therefore, the global climate system. In past decades, Amazonian forests have been severely altered or degraded by droughts, logging activities, forest fires, and deforestation. In the near future, forest fires may increase in frequency as the climate changes in tropical regions. However, there is very limited capacity to predict the potential trajectories of tropical forests given current and projected trends in land use and climate change. In this study, airborne and satellite remote sensing will be combined with statistical models to predict Amazon tree mortality, an important ecological process that influences total forest carbon accumulation and storage, species diversity, microclimate characteristics, and water flow. The study will account for disturbances including fires, logging and droughts, as well as soil fertility and plant characteristics in the high diversity region of Madre de Dios, Peru. While previous studies have attempted to develop tree mortality models for tropical forests, most of them either have been limited by small field-based sample sizes, or they lack geographic coverage and spatial detail. The novelty of this study, therefore, is to map mortality of individual trees across the landscape using airborne measurements, then to scale up to the larger regional level using a combination of satellite and modeling approaches. This study will contribute to the US Sustainable Landscape program of the U.S. Agency for International Development that is supporting programs in tropical nations designed to reduce emissions of carbon from deforestation and forest degradation. The project will also train two American doctoral students, three Latin American graduate students, and six undergraduates.