Sensor-to-Sensor Variability of the ECH2O EC-5, TE, and 5TE Sensors in Dielectric Liquids

Low-budget sensors used in wireless soil water content sensor networks typically show considerable variation. Because of the large number of sensors in sensor network applications, it is not feasible to account for this variability using a calibration between sensor response and soil water content. An alternative approach is to split the calibration into two parts: (i) determination of sensor response-permittivity relationships using standard liquids with a defined reference permittivity, and (ii) site-specific calibration between permittivity and soil water content using a subset of sensors. In this study, we determined sensor response-permittivity relationships for several ECH2O, EC-5, TE, and 5TE sensors by Decagon Devices (Pullman, WA). The objectives of this study were to determine (i) the sensor-to-sensor variability and precision of these sensor types, and ( ii) the increase in accuracy when a sensor-specific calibration is used instead of a single calibration. The results showed that the sensor-to-sensor variability was significantly larger than the measurement noise for each sensor type. When a sensor-specific calibration was used, the RMSE expressed in (equivalent) soil water content ranged from 0.008 cm(3) cm(-3) for the TE sensor to 0.014 cm(3) cm(-3) for the EC-5 sensor in a permittivity range between (similar to)2 and 35. When a single calibration was used, the RMSE was higher and ranged from 0.01 cm(3) cm(-3) for the 5TE sensor to 0.02 cm(3) cm(-3) for the TE sensor. An improvement in accuracy of nearly 0.01 cm(3) cm(-3) can be reached in the high-permittivity range for each sensor type by calibrating each sensor individually.