Infrared and Microwave Satellite Analyses for Land Surface Model Diagnosis
Abstract
Land surface models (LSMs) are essential components of general circulation models for representing exchanges of heat, water, and trace gases between the atmosphere and the surface/biosphere system. Current LSMs have substantial uncertainties and discrepancies among them. Thermal infrared and microwave satellite data can provide information very useful for analyzing LSM performance, complementing information from visible and near-infrared data. For example, our analyses of global data from July 2003 showed NCEP/GDAS land surface temperatures (LSTs) were strongly biased (7 K) in clear areas near midday, relative to LST from MODIS. There are also, however, major discrepancies among satellite products. We have compared LST from MODIS, AIRS, and ISCCP and found, for example, that 21% of clear locations had daytime LST differences > 10 K between ISCCP and MODIS products from July 2003. There were broad arid areas with > 10 K discrepancies in the amplitude of the diurnal temperature range. Our analysis in conjunction with AMSR microwave data found MODIS LST to be much more consistent with constraints from physics and other data than ISCCP was, and much more stable from day to day. Since ISCCP LST production includes geostationary data, ISCCP has the advantage of being able to track the full diurnal cycle, but our results imply that product improvement or scaling will be necessary if the diurnal sampling is to be reliably applied to LSM evaluation. We have used a combination of AMSR and MODIS data to produce a database of microwave land surface emissivity, derived from collocated clear-sky measurements, which we are using as a background information source for retrieving LST in cloudy and clear conditions from AMSR data. These AMSR-derived LST data have a major advantage, with respect to model evaluation: they overcome the clear-sky bias of infrared- derived LST. The derivation of the emissivity database produces metrics of microwave radiative penetration of through the subsurface. These metrics are related to soil and vegetative moisture and have sensitivity beyond the usual microwave soil moisture products. These LST and moisture-related products, analyzed locally and temporally, are a valuable resource for identification of model errors, and for detailed studies to attribute the errors to the components of LSMs that control LST and moisture, and to provide guidance for model improvement.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2008
- Bibcode:
- 2008AGUFM.A43C0312M
- Keywords:
-
- 0315 Biosphere/atmosphere interactions (0426;
- 1610);
- 0325 Evolution of the atmosphere (1610;
- 8125)