ESTIMATING CLEAR-SKY REGIONAL SURFACE FLUXES IN THE SOUTHERN GREAT-PLAINS ATMOSPHERIC RADIATION MEASUREMENT SITE WITH GROUND MEASUREMENTS AND SATELLITE-OBSERVATIONS

The authors compared methods for estimating surface fluxes under clear -sky conditions over a large heterogeneous area from a limited number of ground measurements and from satellite observations using data obta ined from the southern Great Plains Cloud and Radiation Testbed (CART) site, an area of approximately 350 km x 400 km located in Kansas and Oklahoma. In situ measurements from 10 energy balance Bowen ratio (EBB R) stations showed large spatial variations in latent and sensible hea t fluxes across the site because of differences in vegetation and soil conditions. This variation was reproduced by a model for parameteriza tion of subgrid-scale (PASS) surface fluxes that was developed previou sly and extended in the present study to include a distribution of soi l moisture inferred from combined visible and thermal infrared remote sensing data. In the framework of the PASS model, the satellite-derive d normalized difference vegetation index and surface temperature were used to derive essential surface parameters including surface albedo, surface conductance, soil heat flux ratio, surface roughness length, a nd soil moisture, which were then used to calculate a surface energy b udget at satellite-pixel scales with pixel-specific surface meteorolog ical conditions appropriately distributed from their mean values using a distribution algorithm. Although the derived soil moisture may be i nfluenced by various uncertainty factors involved in the satellite dat a and the model, spatial variations of soil moisture derived from the multichannel data from the Advanced Very High Resolution Radiometers o n the NOAA-14 satellite appeared to have some correlation (the correla tion coefficient is as large as 0.6) with the amount of accumulated pr evious rainfall measured at the 58 Oklahoma Mesonet stations located w ithin the CART area. Surface net radiation, soil heat flux, and latent and sensible heat fluxes calculated at a spatial resolution of 1 km ( the size of a satellite pixel) Were evaluated directly by comparing wi th flux measurements from the EBBR stations and indirectly by comparin g air temperature and humidity inferred from calculated sensible and l atent heat fluxes with corresponding values measured at 1.5 m above th e 58 meteorological stations. In calculating regional fluxes, biases c aused by the sampling uncertainty associated with point measurements m ay be corrected by application of the satellite data.