Soil Moisture Estimation Using Atmospherically Corrected C-Band InSAR Data

A methodology to generate calibrated maps of soil moisture from C-band synthetic aperture radar (SAR) images processed by SAR interferometry (InSAR) technique is presented. The proposed methodology uses atmospheric phase delay (APD) maps obtained from a time series of Sentinel-1 interferograms, to disentangle the APD and soil moisture contributions to Sentinel-1 interferograms. We show how the high spatial resolution and short temporal baseline of Sentinel-1 image can help to estimate soil moisture using a daisy chain InSAR processing. The estimated soil moisture maps are compared with in situ data collected by five soil moisture sensors installed in an experimental field, characterized by bare soil, located close to Lisbon, Portugal. Results show that after removing the APD effects in SAR interferogram, there is a correction of the bias in the soil moisture estimation and an improvement in the correlation coefficient with the soil moisture measurements, from 0.38 to 0.78. Soil moisture changes were measured during a sequence of rain events in the winter season. A root-mean-square (rms) error less than 0.04 m(3)/m(3) was found over a variety of meteorological conditions.

Automated summary

A methodology using C-band SAR images processed by InSAR technique is proposed to generate calibrated maps of soil moisture. By utilizing APD maps from Sentinel-1 interferograms, the methodology separates APD and soil moisture contributions, resulting in improved soil moisture estimation accuracy. The study demonstrates the effectiveness of the approach through comparison with in situ soil moisture data collected in an experimental field, showing a significant increase in correlation coefficient.