Reducing the uncertainty of surface albedo in climate models for better predictions of the future Amazonia climate
Silvia
Monteiro
Santos, Universidade Federal de Viçosa, smonteiro@vicosa.ufv.br
(Presenting)
Marcos
Heil
Costa, Universidade Federal de Viçosa, mhcosta@ufv.br
Surface albedo has been shown to significantly affect tropical atmospheric circulation and climate. However, previous studies that considered the effect of changes in surface albedo on the regional climate, including former tropical deforestation studies, usually used a land surface parameterization that represented albedo with an error of the same magnitude of the albedo signal that would drive the climate change. Two decades ago, climate models estimated annual-mean albedo with + – 0.05 absolute accuracy, while ten years ago the uncertainties improved to 0.02. Today, although an error <0.01 for the annual-mean albedo estimate is usual, it can reach 0.03 on a monthly basis. As uncertainties in the representation of land surface albedo can translate to uncertainties in the simulated climate, it is important to estimate the albedo with high accuracy (<0.01), even at the seasonal scale. Our strategy is to reduce this uncertainty through the incorporation of new processes that affect the land surface albedo at a seasonal scale, and through a better calibration of the models against field data collected at different field sites and against remote sensing products. Initial results indicate that some processes that currently are not (well) represented in land surface models, like leaf wetness and the correct partition of the incoming solar radiation in direct and diffuse fields, introduce biases in the surface albedo that account for most of uncertainties at the seasonal scale. The incorporation of these processes in an off-line version of the land-surface model IBIS successfully reduced the seasonal error at tropical rainforest sites. Future tests will be performed with IBIS coupled to a climate model.
Submetido por Silvia N. Monteiro Yanagi em 05-MAR-2004