This project is advancing our ability to anticipate the sustainability impacts of development in the Amazon Basin. Building on past and ongoing LBA efforts, it is extending basin-scale empirical modeling of deforestation at the pixel-level, and integrating predicted land cover changes with a regional climate model to ascertain climate impacts associated with development scenarios for the Amazon Basin. The pixel-level land cover model uses remote sensing products of deforestation to probabilistically describe Amazonian landscapes as a function of the spatial distribution and times paths of observed deforestation drivers. These landscapes, in turn, are input to the regional climate model (the Regional Atmospheric Modeling System [RAMS]). First, the land cover change model produces a set of deforestation probabilities, associated with the specific development scenarios, covering the entire basin. Each pixel, with scenario-specific probability, is treated as a Bernoulli trial, and probability functions of GIS software produce hundreds of realizations of the basin landscape for each scenario. The multiple probabilistic landscapes are input to RAMS, which is executed to produce probability density functions of key variables (e.g., total yearly rainfall), with associated estimates of distributional parameters such as mean value (?) and variance (?2). This explicitly addresses uncertainty by developing measures of central tendency (?) and dispersion (?2) for the estimated climate impacts of the development scenarios. The figure shows spatially-explicit output describing differences at basin scale in rainfall associate with business as usual and restraints on development scenarios. The project is also developing models of forest fragmentation with a Brazilian institutional collaborator, IMAZON. This activity uses maps of logging roads developed by analysis of remotely sensed images, and digital elevation maps based on the shuttle radar topography mission (SRTM). The goal is to refine spatially explicit models of human behavior, in order to accurately reflect processes of forest fragmentation in Amazonia.