Submitted to: Remote Sensing of Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: May 1, 2003
Publication Date: July 1, 2003
Citation: Bindlish, R., Jackson, T.J., Wood, E., Gao, H., Bosch, D.D., Lakshmi, V. 2003. Soil moisture estimates from TMI observations over the southern United States. Remote Sensing of Environment. 85:507-515.
Interpretive Summary: This paper describes the first attempt to map at continental scales daily soil moisture from space over an extended period of time. Previously, such attempts have been either limited to small areas or over limited period of time. The methodology used for soil moisture estimation uses data that can be easily observed over all areas. The methodology is dynamic and can be transported to any region. Data used are from the Tropical Rainfall Measuring Mission Microwave Imager observations. The estimated soil moisture during specific field experiments compared well with the observations. This provides a reasonable ground validation at catchment scale. Large scale validation of soil moisture was conducted using the meteorological observations over the southern U.S. This methodology will be extended to the entire temporal domain of data (December 1997-current). Further analysis and validation will be conducted for extended periods. It is anticipated, these estimates of soil moisture will provide a valuable dataset for the hydrologic community. These results will contribute to experiment design and the potential operational implementation of this technique in hydrologic, climate and agricultural applications.
The lack of continuous soil moisture fields at large spatial scales, based on observations, has hampered hydrologists from understanding its role in weather and climate. The most readily available observations from which a surface wetness state could be derived is the TRMM (Tropical Rainfall Measuring Mission) Microwave Imager (TMI) observations at 10.65 GHz. This paper describes the first attempt to map daily soil moisture from space over an extended period of time. Methods to adjust for diurnal changes associated with this temporal variability and how to mosaic these orbits are presented. The algorithm for deriving soil moisture and temperature from TMI observations is based on a physical model of microwave emission from a layered soil-vegetation-atmosphere medium. An iterative, least-squares-minimization method, which uses dual polarization observations at 10.65 GHz, is employed in the retrieval algorithm. Soil moisture estimates were compared with ground measurements over the U.S. Southern Great Plains (SGP) in Oklahoma and the Little River Watershed, Georgia. The soil moisture experiment in Oklahoma was conducted in July 1999 and Little River in June 2000. During both the experiments, the region was dry at the onset of the experiment, and experienced moderate rainfall during the course of the experiment. The regions experienced a quick dry-down before the end of the experiment. The estimated soil moisture compared well with the ground observations for these experiments (standard error of 2.5%). The TMI estimated soil moisture during 6-22 July over southern U.S. were analyzed and found to be consistent with the observed meteorological conditions.