A number of articles on this blog highlight the critical role land-surface representation plays in the simulation of climate scenarios. Better land-surface representation is often achieved by enhanced representation of the land-surface processes and land-atmosphere interactions, but despite that, one significant limitation that often remains is the lack of global soil moisture datasets that can be used for developing the model scenarios. Soil moisture representation is considered important since errors in the initializing soil moisture can persist for a significant period (i.e., soil moisture errors have high memory), and lead to higher uncertainty in the model results. Soil moisture availability often controls surface albedo, and the surface Bowen ratio and thus affects the surface temperature but can also contribute to the regional moisture recycling (evaporation and precipitation). Recently new results from the Global Soil Wetness Project (GSWP) were reported by Dirmeyer et al. (GSWP-2: Multimodel Analysis and Implications for Our Perception of the Land Surface, Paul A. Dirmeyer, Xiang Gao, Mei Zhao, Zhichang Guo, Taikan Oki, and Naota Hanasaki, Bulletin of American Meteorological Society, vol. 87 (10), p. 1381-1397 (Available from www.ametsoc.org). The GSWP-2 analysis combines “the simulations of more than a dozen different global land surface models, an unprecedented analysis of terrestrial water and energy budgets is realized.” The GSWP-2 data are available from http://www.iges.org/gswp/ as well as http://haneda.tkl.iis.u-tokyo.ac.jp/gswp2/ and http://www.monsoondata.org:9090/dods/gswp2mma/
Excerpts from the paper read “…the first global multimodel analysis of land surface state variables and fluxes for potential use by meteorologists, hydrologists, engineers, biogeochemists, agronomists, botanists, ecologists, geographers, climatologists, and educators. This is, in many respects, a land surface analog to the atmospheric reanalyses ….but one that encompasses an ensemble of different LSSs (i.e. land surface schemes). Using the results of multiple LSSs provides a result that is not dependent on a single model, is generally superior to the results of any individual model, and is typically as good as or better than the best model at each point and time.” It will be of interest to see how the climate community utilizes this new product and the potential changes that would result in the model scenarios.