Development of a Tool to Predict Soil Moisture and Soil Temperature Regimes

Physico-biochemical processes occurring in soil are difficult to predict because the knowledge of local soil properties such as soil moisture and temperature is often limited. Therefore, soil moisture and temperature regime classes are necessary for US soil taxonomy and other classification systems. The goal of this study is to develop a modeling tool to predict soil moisture and temperature at multiple soil horizons and to code the Keys to Soil Taxonomy in a Soil Moisture and Temperature Regime Classification (SMTRC) module for automatic identification of soil moisture and temperature regimes. The Environmental Policy Integrated Climate (EPIC) model was extended as the EPIC–SMTRC tool for this purpose. Field data from the Soil Climate Analysis Network (SCAN) sites and the Wye farm site in Maryland for validation of soil moisture and temperature predictions. Results indicate that predicted daily soil temperatures are in close agreement with observed values with R2 values ranging from 0.73 to 0.98 and Nash–Sutcliffe efficiency (NSE) from 0.50 to 0.96. Predicted soil moisture by EPIC–SMTRC captured observed trends reasonably well. Testing results at the Wye farm indicate that the predicted daily values of water content are satisfactory, with R2 values ranges from 0.66 to 0.88 and NSE from 0.51 to 0.84. The EPIC–SMTRC tool demonstrated the ability to automate the identification of the soil moisture and temperature regimes as currently defined in the US soil taxonomy and can be used to classify soils and to be utilized for other sites.