Optimizing soil moisture sensor installation depths and number of sensors for in situ monitoring networks

Abstract Previously, developers of soil moisture monitoring networks have determined how many sensors to use and their installation depths with little objective guidance, which led to heavy reliance on suboptimal past precedents when deploying new networks. One such network, the Oklahoma Hydronet, is being developed to monitor water stored in the state's soils, aquifers, and reservoirs. To monitor soil water storage, an evidenced‐based approach should be used to determine the optimal sensor number and depths. Therefore, our objectives were to identify optimal depths for installing sensors within the profile and the number of sensors required to accurately quantify soil water storage or profile average water content. We evaluated profile depths ranging from 30 to 120 cm. We compared known soil moisture in 34 field‐measured profiles and 6576 modeled profiles with estimates from hypothetical sensors placed at various depths. Based on those profiles, we determined that (1) for a 100‐cm profile, the shallowest sensor should be placed at 10 cm, (2) the deepest sensor should be shallower than the lower boundary of the profile, (3) sensors should be approximately evenly spaced, and (4) four sensors are sufficient. The most widely used sensor depth combination in the United States has sensors at 5, 10, 20, 50, and 100 cm, but this depth combination is not recommended because it resulted in median absolute errors 1.8–2.8 times greater than the best four‐sensor depth combinations. Instead, to quantify soil water storage in a 100‐cm profile, we recommend placing sensors at approximately 10, 30, 60, and 90 cm.