Calibrating Soil Moisture Sensors during Evaporation Experiments on Sphagnum Moss and Peat.

Since the water content of peatlands influences the fluxes of water, solutes, energy, and gases in peatlands, the spatiotemporal dynamics of water content must be determined accurately. This enables the quantification of matter and energy fluxes within and over the system boundaries. To measure local soil moisture, electromagnetic moisture sensors are widely used. These sensor types make use of the effect that soil water, soil air, and soil grains reveal very different di-electric properties. However, there is no unifying principle which makes predicting a soil moisture content from a sensor output possible. Moreover, the sensor output will always be specific to its geometry and electromagnetic properties. For this reason, the constitutive relationship has to be established by calibration. Vaz et al. (2013) carried out an extensive comparison of electromagnetic soil moisture sensors. Their study was conducted on a wide range of mineral soils with varying organic carbon content, salinity, and grain size distribution. They conclude that even in the presence of calibration functions provided by the manufacturer, sensor and soil type specific calibration are necessary. This is a time consuming process. Therefore, the objective of our study was to develop a continuous calibration procedure for organic soils which is based on a transient evaporation experiment under laboratory conditions and evaluation by inverse modelling. The organics soils had very low salinity and mineral content, and high porosities with water contents close to saturation > 80%. Peat samples were instrumented with tensiometers and soil water content sensors at various depths. Open to the atmosphere, the setup was placed on a balance to determine the water loss by evaporation. We identified the soil water retention curve of the sample by inverse modelling with the Richards equation. After computing the spatially-averaged soil water contents representative for the measurement volume of the water content sensors, we derived calibration curves for the sensors by relating the water content to the sensor signal. The novel approach provides calibration points in high temporal resolution by an automated experiment which is an easy-to-implement technique. A comparison with classic techniques reveals the advantages of our novel procedure.