Study of accuracy and sensing volume of a wide range of established and low-cost soil water content probes

Continuous soil moisture content monitoring is key to understand the soil and water flow and transport processes and their impact on a wide range of environmental change and quality processes. Nowadays, there are a wide variety of electromagnetic sensors for estimating soil volumetric content. These include those from established manufacturers (>50€) and low-cost (<50€) amateur electronics enthusiasts with open-source projects. For each sensor, the manufacturer typically provides both a calibration function and an estimation of the sensing volume of the device. The objective of this work is to evaluate the performance –regarding the accuracy and effective sensing volume– of a wide variety of soil water sensors and to compare these results with those provided originally by the respective manufacturer. Twenty-five different electromagnetic sensors representative of the current best-known commercial (19) and the low-cost brands (6) were tested in the laboratory using 3 soils of contrasting texture. Benchmark values for comparison were obtained by the gravimetric method. The sensing volume for each probe was characterized by recording readings while the probes approached a water surface. The Root Mean Square Errors of the sensors ranged from 0.02 to 0.10 cm³/cm³, and the sensing volume of the different probes exhibited a large variability, ranging from 0.5 to 1500 cm³. Importantly, the probes evaluated in laboratory conditions showed different errors for each soil type. The loamy soil readings presented the smallest errors, followed by sandy and clayey soils. No statistically significant differences were found in measurement accuracy between low-cost and higher-priced probes. From the study of the sensing volume explored, with the exception of one case, it can be concluded that the low-cost probes generally explore a smaller volume than the established probes. The selection of the appropriate probe based on its sensing value could be important for different types of risk analysis and management applications.