Automatic calibration of energy-water and carbon exchange processes in the SiB2 model for tropical forest before and after selective logging.
Rafael
Rosolem, Department of Hydrology and Water Resources, University of Arizona, Tucson, AZ - USA / Departamento de Ciências Atmosféricas, IAG, Universidade de São Paulo, São Paulo, SP - B, rafael@hwr.arizona.edu
W.
James
Shuttleworth, Department of Hydrology and Water Resources, University of Arizona, Tucson, AZ - USA, shuttle@hwr.arizona.edu
Eleanor
J.
Burke, Met Office, Hadley Centre for Climate Prediction and Research, Exeter - UK / Department of Hydrology and Water Resources, University of Arizona, Tucson, AZ - USA, eleanor.burke@metoffice.com
Luis
A.
Bastidas, Department of Civil and Environmental Engineering, Utah State University, Logan, UT - USA, luis.bastidas@usu.edu
(Presenting)
Humberto
Ribeiro da
Rocha, Departamento de Ciências Atmosféricas, IAG, Universidade de São Paulo, São Paulo, SP - Brasil, humberto@model.iag.usp.br
Luis
G.
Gonçalves, Department of Hydrology and Water Resources, University of Arizona, Tucson, AZ - USA, gustavo@hwr.arizona.edu
Scott
Dennis
Miller, Department of Earth System Science, University of California Irvine, Irvine, CA - USA, sdmiller@uci.edu
Michael
L.
Goulden, Department of Earth System Science, University of California Irvine, Irvine, CA - USA, mgoulden@uci.edu
Attempts to model surface-atmosphere interactions with greater physical realism have resulted in complex land surface schemes (LSS) with many parameters. While the parameterizations of the LSS models have been designed with the idea that it should be possible to estimate reasonable values for parameters from measurable characteristics of the land surface, recent studies have demonstrated that (even manually) adjustment of a few model parameters can result in significant improvements in model performance. Arguably, understanding of difference between preferred parameters before and after selective logging can provide background information on changes in some of the (physiological, soil physical, morphological) properties of the ecosystem. In the second-generation of the Simple Biosphere Model (SiB2), representation of carbon exchange process was introduced. This work explores specification of parameter sets for SiB2 for a tropical rainforest biome using optimization techniques (automatic calibration) in two cases, before and after selective logging. The data used were from the km 83 Cuiabá-Santarém Highway site (Floresta Nacional do Tapajós) LBA site. Optimization of the parameters in SiB2 was made by minimizing the Root Mean Square Error (RMSE) between time series for three objective functions, latent heat flux, and sensible heat flux and carbon exchange.
Submetido por Rafael Rosolem em 16-MAR-2004
Tema Científico do LBA: CD (Armazenamento e Trocas de Carbono)