Surface soil moistureand the nature of the overlying vegetation both influence microwave emissionfrom land surfaces significantly. One widely discussed butunderused method for allowing for the effect of vegetation on soil-moistureretrievals from microwave observations is to use remotelysensed vegetation indices. This paper explores the potential for using theNormalised Difference Vegetation Index (NDVI) in soil-moistureretrievals from L-band (1.4 GHz) aircraft data gathered during the SouthernGreat Plains '97 (SGP97) experiment. A simplified versionof MICRO-SWEAT, a soil vegetation atmosphere transfer (SVAT) scheme coupled witha microwave emission model, was used as the retrievalalgorithm. Estimates of the optical depth of the vegetation, the parameter thatdescribes the effect of the vegetation on microwave emission,were obtained by calibrating this retrieval algorithm against measurements ofsoil moisture at 15 field sites. A significant relationship wasfound between the optical depth so obtained and the observed NDVI at thesesites, although this relationship changed with the resolution ofthe microwave brightness temperature observations used. Soil-moisture estimatesmade with the retrieval algorithm using the empirical relationshipbetween optical depth and NDVI applied at two additional sites not used in thecalibration show good agreement with field measurements. style="line-height: 20px;">Keywords: NDVI, soil moisture, passive microwave, SGP97
展开▼
