Authors: F Chen
Publish Date: 2005/05/13
Volume: 24, Issue: 7-8, Pages: 667-684
Abstract
The International Satellite LandSurface Climatology Project ISLSCP InitiativeI 1degree 1987–1988 data were used to drive a land surface model LSM to simulate global surface energy budgets Simulated surface heat fluxes show remarkable spatial variability and seem to capture well their annual and interannual variability A shift of maximum evaporation across the equator is more closely related to the seasonal shifting of precipitation pattern than to surface radiation changes The NCEP/NCAR reanalysis did not reflect this shift presumably due to its dominant rainfall maximum in the Southern Hemisphere To assess the “reliability” of these fields both Global Soil Wetness Project GSWP and reanalysis were verified against observations at two sites Monthly mean ISLSCP forcing conditions agree fairly well with observations but its precipitation is usually lower during spring and summer Low summer GSWP evaporation may be due to low precipitation and incorrect specification of vegetation and soil conditions The reanalysis had larger seasonal variability than GSWP and observations and overestimated summer heat fluxes because of its large rainfall and surface radiation Despite uncertainty in ISLSCP data an LSM with a modest treatment of vegetation was able to capture reasonably well the seasonal variations in surface heat fluxes at global scales With some caution these types of simulations can be used as “pseudoobservations” to evaluate climatemodel simulations and to investigate global energy budgets For the next phase of ISLSCP data development higher resolution data which can reflect local heterogeneity of vegetation and soil characteristics include more rain gauge data are highly desirable to improve model simulationsThis study was supported by the NCAR Water Cycle Initiatives We express our appreciation to Dr Thomas Warner at NCAR for his review and suggestions We also appreciate the support of Dr Kenneth Mitchell at NCEP where the LSM simulations were done We thank Drs Alan Betts Anton Beljaars and Fred Bosveld for kindly providing us with the FIFE and Cabauw datasets
Keywords: