Calibrating soil water potential sensors integrated into a wireless monitoring network

Abstract

Monitoring of the soil water status is a proper method for optimizing agricultural irrigation. Wireless sensor networks enhance data availability, thus, they can be used as decision support systems. In this study two types of soil water potential sensors were tested. One was the well-established Watermark sensor by Irrometer Co., the other was the relatively new MPS-1 by Decagon Devices, Inc. The goal was to (1) integrate the sensors into a wireless monitoring network, (2) determine and evaluate calibration functions for the integrated sensors, and (3) compare the measuring range and the reaction time of both sensor types in a soil layer during drying. The integration of the sensors into the telemetry network worked well. Data were transmitted over several kilometers and made available via Internet access. Calibration was done for several sensors in a pressure pot. A common calibration function was found for the combination of Watermark sensors with the required interface. Sensor specific calibrations became essential for the MPS-1 due to the very large sensor-to-sensor variation. Four approaches were applied and evaluated: Fitting of a standard power function, fitting of a retention function, using so-called one-point calibrations, and using the factory calibration. The latter was not useful at all. The first two methods performed best. The one-point calibrations turned out to be a sound alternative, because the method is less time consuming. A set of sensors was installed in a thin soil layer in the laboratory in order to compare the water potential measurements during drying. Both sensor types delivered water potential measurements in a range from −10kPa to −600kPa. For values <−130kPa the Watermark sensors reacted significantly more slowly than the MPS-1.