Evaluation of the OPTRAM Using Sentinel-2 Imagery to Estimate Soil Moisture in Urban Environments.

Abstract

The determination of soil moisture is a crucial issue for various purposes, including hydrological, climatological, and agricultural studies. Over the past few decades, several distinct remote sensing approaches have been developed. One recent development is the Optical TRApezoil Model (OPTRAM). This approach is similar to the traditional TOTRAM, but it replaces the LST index (thermal band) with the STR index, which is calculated using the SWIR band. Numerous studies have demonstrated the effectiveness of OPTRAM in predicting soil moisture. However, the capability of OPTRAM to estimate soil moisture in urbanized areas has not yet been fully recognized. To address this gap, we conducted tests in the Rio Claro municipality, where land use and occupation vary significantly. By utilizing Sentinel-2 multispectral images, we constructed the NDVI-STR space, estimated soil moisture, and compared it with field measurements. The values of R2, MAE, and RMSE for the OPTRAM-derived soil moisture at urbanized of Rio Claro were 0.92, 0.0196, and 0.1413, respectively. These results demonstrate a high level of representativeness for the soil moisture estimates Furthermore, the freely distributed Sentinel-2 satellite images has a spatial resolution that is well-suited to the dimensions of the target areas in the evaluated scene.