Spatiotemporal Analysis of Soil Moisture and Optimal Sampling Design for Regional‐Scale Soil Moisture Estimation in a Tropical Watershed of India

Abstract

AbstractRegional‐scale soil moisture estimates are essential for several hydrological applications and for validating remote sensing‐based soil moisture products. Characterization of the regional‐scale soil moisture variability requires a robust in situ monitoring strategy at point scale to balance between representativeness and minimization of monitoring cost. In this study an optimal sampling design was determined to capture the spatiotemporal variability of soil moisture at watershed scale. The study was conducted for the typical Indian conditions of extreme seasonal variability that leads to very wet (during monsoon) to dry (during hot summer) in the eastern India. Soil moisture monitoring was done at 83 locations in an agricultural watershed of 500 km2 for 56 days of field campaigns across a year. Based on the analyses of 41,832 measurements collected during field campaigns, it was found that maximum numbers of required locations necessary to estimate watershed‐mean soil moisture within ±2% accuracy are 30. Moreover, five randomly selected locations were found to be sufficient for capturing the temporal variability of watershed‐mean soil moisture with an accuracy of ±3%. In addition, five most representative locations identified through time stability analysis can provide robust estimate of watershed‐mean soil moisture with accuracy of ±2%. Further, soil properties and topography are identified as significant physical parameters that jointly control the spatiotemporal persistence and variability of soil moisture in the Indian watershed. These findings will be quite useful to provide guidelines for optimizing short‐term soil moisture campaigns by sampling at few selected points representative of the watershed‐mean behavior.