Development and Evaluation of Soil Moisture‐Based Indices for Agricultural Drought Monitoring

Core Ideas In situ soil moisture data were used to develop agricultural‐drought indices. Promising indices were directly linked to drought impacts (i.e., lower crop yield). Preferred indices, formulated as anomalies, were comparable across time and space. These can be derived from in situ soil moisture data common to networks worldwide. Our methodology is transferrable to other regions with in situ soil moisture data. ABSTRACT Agricultural drought is characterized by low soil moisture levels that negatively affect agricultural production, but in situ soil moisture measurements are largely absent from indices commonly used to describe agricultural drought. Instead, many indices incorporate weather‐derived soil moisture estimates, which is necessary, in part, because the relationships between in situ soil moisture and agricultural‐drought impacts are not well quantified. Our objective was to use in situ soil moisture data from monitoring networks in Oklahoma and West Texas to identify a soil moisture‐based agricultural drought index that is (i) strongly related to crop‐yield anomaly across networks, (ii) comparable across time and space, and (iii) readily understandable. Candidate indices included soil matric potential (MP), soil water storage (SWS), and fraction of available water capacity (FAW), with indices assessed in their raw form and after climatological (i.e., anomalies) or statistical standardization. At the county level, indices related similarly to crop‐yield anomaly, with soil moisture‐yield anomaly correlation coefficients averaging 0.63, 0.76, and 0.76 for winter wheat, hay, and cotton, respectively. However, standardization was essential to maximize temporal and spatial comparability, and at the regional level, standardized indices were more highly correlated with crop‐yield anomaly than non‐standardized indices. Our findings show that existing in situ soil moisture datasets can underpin regional drought‐monitoring systems. The SWS‐anomaly may be the preferred index because it is comparable across space and time, has units that are readily understandable (e.g., mm or inches), and can be broadly applied using data from the many in situ soil‐moisture monitoring networks across the world.