Precision of Neutron Scattering and Capacitance Type Soil Water Content Gauges from Field Calibration

Abstract

AbstractSoil water content gauges based on neutron scattering (NS) have been a valuable tool for soil water investigations for some 40 yr. However, licensing, training, and safety regulations pertaining to the radioactive source in these gauges makes their use expensive and prevents use in some situations such as unattended monitoring. A capacitance probe (CP) gauge has characteristics that would seem to make it an ideal replacement for NS gauges. We determined the relative precision of two brands of NS gauges (three gauges of each) and a brand of CP gauge (four gauges) in a field calibration exercise. Both brands of NS gauges were calibrated vs. volumetric soil water content with coefficients of determination (r2) ranging from 0.97 to 0.99 and root mean squared errors (RMSE) <0.012 m3 m−3 water content. Calibrations for the CP gauges resulted in r2 ranging from 0.68 to 0.71 and RMSE of 0.036 m3 m−3 water content. Average 95% confidence intervals on predictions were three to five times higher for the CP gauges than for the NS gauges, ranging from 0.153 to 0.161 and 0.032 to 0.052 m3 m−3, respectively. Although poorly correlated with soil water content, readings were reproducible among the four CP gauges. The poor correlation for CP gauges may be due to small‐scale soil water content variations within the measurement volume of the gauge. The NS gauges provide acceptable precision but the CP gauge has poor precision and is unacceptable for routine soil water content measurements.