An empirical soil moisture estimation algorithm using imaging radar

An empirical algorithm was developed for the retrieval of soil moisture content and RMS height from remotely sensed data. It was developed using scatterometer data acquired with the RASAM and POLARSCAT systems. The algorithm requires two co-polarized channels at L-band. Omitting the usually weaker hv-polarized returns makes the algorithm less sensitive to system cross-talks and system noise. Furthermore, since the co-polarized terms are less sensitive to the presence of vegetation than the cross-polarized terms, this algorithm is more robust in the presence of vegetation than one relying on the hv-polarized returns. Even so, the authors results indicate that significant amounts of vegetation causes the algorithm to underestimate soil moisture. In this paper, the algorithm is described and its accuracy tested against the scatterometer data set. This algorithm is then applied to multipolarization SAR data acquired with the NASA/JPL multi-frequency multi-polarization AIRSAR system. In order to assess the accuracy of the studied algorithms, the estimated soil moisture values are compared with the actual “in situ” measurements. Accuracy of 5% or better are typically found for bare surfaces