Peatland Water Table Depth Monitoring Using Quad-Pol L-Band Sar

With remote sensing increasingly used to support peatland management, this study evaluates the potential of quad-polarimetric L-band SAR for water table depth (WTD) monitoring. We produced Random Forest models predicting WTD, and temporal WTD variation, across multiple locations and dates at a Scottish raised bog, based on observables derived from quad-pol L-band SAR data. Like other studies, our models showed limited ability (R
²
= 0.05) to predict WTD across multiple locations. However, significant validated performance (R
²
= 0.66) was found when predicting temporal WTD variation, exceeding the performance found by comparable studies using C-band SAR. Therefore, further work using quad-pol L-band SAR is recommended, particularly to understand and correct for the effects of non-WTD characteristics on SAR signal which limit model performance spatially.