An outstanding problem in hydrology and meteorology spanning all spatial and temporal scales is estimating how the available net radiation at the land surface is distributed into the various component fluxes of the land surface energy balance, most importantly evapotranspiration and the sensible heat flux. The importance of this distribution of water and energy has been demonstrated by many numerical studies using General Circulation Models (GCMs) to explore the effects of varying soil moisture and resulting variations in the balance of sensible and latent heating on the global circulation and precipitation. Other modelling studies have noted the importance of regional variations in soil moisture and evapotranspiration on the character of mesoscale circulations, severe weather and cloud development, precipitation and other aspects of regional and smaller-scale atmospheric phenomena. A better understanding of the evapotranspiration and sensible heating at the land surface would also be valuable to agriculture, hydrology and many related fields.
Currently it is possible only at relatively small spatial scales to measure the sensible heating and evapotranspiration from the land surface or the inter-annual variations which may result from changes in rainfall and resulting soil moisture and vegetative response. The unavailability of such measurements for regional and larger spatial scales has been an impediment to the development of suitable land surface parameterizations and closure of the land surface energy/water balance, even though certain field programs such as the First ISLSCP Field Experiment (FIFE) and others have been designed to bridge the gaps between scales. With inadequate verification of surface quantities at regional and larger scales, investigators have instead mainly relied on indirect methods of verification, such as trying to generalize measurements made over smaller scales, analyzing patterns of land use and vegetation and the use of precipitation indices or remotely-sensed vegetation indices (such as the Normalized Difference Vegetation Index, NDVI) as proxies for evapotranspiration and soil moisture.
Current land surface energy/water budget studies at the CIMSS are designed to exploit combinations of remote sensed and in-situ data types to better understand land surface exchanges at regional and larger scales. The long-term goal is to combine these data types in a land surface "assimilation" scheme based on the principles of optimal estimation (the same sorts of schemes which are currently used for assimilating atmospheric data into numerical prediction models). This will enable the land surface energy and water balance and potentially other land characteristics to be evaluated at regional and larger scales. Ideally, this "optimal" system will account for the relative information content of the various data sources, the relative spatial and temporal distributions of these data and the relative accuracy of the "inversion" methods which are used to translate raw data into the estimates of land surface quantities.
Recent work, conducted by CIMSS scientist George Diak and NOAA scientist Robert Rabin, investigates the relationships between several in-situ and remote sensed signals of the land surface energy balance and soil moisture, as well as NDVI. Some of the surface characteristic signals analyzed are: 1) two precipitation indices, one derived from surface rainfall measurements and the other from satellite microwave data, 2) the temporal changes of surface "skin" (radiometric) temperatures from the GOES satellites, and 3) measurements of the temporal changes of the height of the planetary boundary layer from radiosonde reports, a very sensitive indicator of the surface sensible heating.
The relationship of the surface 12-hour Bowen ratio and NDVI for 5 June 1988 shows agreement with the pattern of developing drought conditions in the area. Recall that the Midwest and Northern Plains experienced drought conditions and abnormally warm temperatures during the summer of 1988. Increasing "greeness" in the NDVI image indicates increasing NDVI and vegetation cover. The contour values represent Bowen ratio (the ratio of sensible to latent heating) values multiplied by 10. The Bowen ratio was estimated from skin temperature data from the GOES and planetary boundary layer height measurements derived from radiosondes. The yellow contours represent the temporal change of surface skin temperature, which is evaluated from GOES data by subtracting the measured skin temperature at three hours after local sunrise from the maximum skin temperature measured during the day. The correlation between the temporal changes of skin temperature and NDVI over the entire region was poor. However, in the eastern section of the region, which was already experiencing drought conditions, the correlation between skin temperature and NDVI was greatly improved.