Development of a New Wetness Index Based on RADARSAT-1 ScanSAR Data

Abstract

A new wetness index (WI) was developed based on the temporal patterns in radar brightness (β⁰) in a timeseries of RADARSAT-1 ScanSAR images collected over the July–September period of 2005. The WI proposed here provided an indirect measure of soil water content (SWC), as β⁰ was documented to vary with land-surface water content. Hydrological factors affecting SWC, such as soil texture, topography, evapotranspiration, etc., were not considered in the current determination of WI. WI-values generated with the proposed method were subsequently compared against field measurements of SWC collected from three separate areas, including densely- and sparsely-forested and non-forested areas (i.e., bare fields), all located in southcentral New Brunswick (NB), Canada. The comparison revealed adequate agreement between WI and SWC for all three areas, including dense forests, yielding coefficients of determination (r²’s) of 74–99 %. Reasonable agreement for dense forests (r² = 74 %) indicated the potential of the method in determining SWC under heavily-vegetated conditions. This correlation would arise because of the equilibrium established between foliage water content (picked up by the radar signal) and SWC under normal, non-stressed conditions. A second evaluation of the method was conducted by comparing WI-values with spatial calculations of SWC obtained with the Soil Water Assessment Tool (SWAT) for bare-field conditions common to the potato-growing area of northwestern NB. Again, suitable agreement was obtained, yielding r²-values ranging from 65 % to 81 %. However, further research is needed to evaluate the usefulness of the method for other forested and non-forested regions of the world. In principle, because the method relies mostly on β⁰, it is highly likely the method can be used to assess SWC in many different types of natural environments.