Performance analysis of capacitance and electrical resistance-type soil moisture sensors in a silt loam soil

The performance of the Irrometer 200SS Watermark granular matrix sensor (WM), John Deere Field Connect (JD-v2) probe, and Delta-T PR1-capacitance (PR1-C) probe were evaluated against a Troxler 4302 neutron gauge (NG) for in-season field volumetric water content (θ _v) measurements at two soil depths in a Hastings silt loam soil at the University of Nebraska-Lincoln/Institute of Agriculture and Natural Resources South Central Agricultural Laboratory (SCAL) near Clay Center, Nebraska. The performances of the sensors were investigated over three years (2011-2013) under various water, nutrient, and crop management practices. The WM sensors performed best when using a field-calibrated soil water retention curve (SWRC) [root mean square difference (RMSD) = 0.024 m³ m^-3] as compared to a SWRC developed from a pedotransfer function (RMSD = 0.070 m³ m^-3). The WM sensors using a previously developed SWRC for the experimental field resulted in RMSD values less than 0.05 m³ m^-3 when compared to the NG-measured θ_v at all depths and years. The JD-v2 probes underestimated θ_v in the dry range and overestimated θ_v in the wet range, which resulted in regression slopes and intercepts for the 0.30 and 1.0 m soil depths that were significantly different from unity (i.e., 1.0) and zero (p_0.05 < 0.05), respectively. However, the performance of the JD-v2 probes improved in 2013 from RMSD values of 0.066, 0.068, and 0.067 m³ m^-3 when using the manufacturer's calibration to 0.043, 0.033, and 0.038 m³ m^-3 for the 0.30, 1.0 m, and pooled soil depths, respectively, after calibrating the JD-v2 probes using the 2012 regression responses. The PR1-C probe-estimated θ_v regression coefficients were also significantly different from unity and zero (p0.05 < 0.05), respectively, and the resulting RMSD values ranged from 0.077 to 1.133 m³ m^-3 . In general, the WM and JD-v2 sensors followed the NG total water (TW) trends at all irrigation levels; however, the PR1-C probe was inconsistent and at times estimated unrealistic TW amounts. While there are limitations and concerns for each sensor technology, the results suggest that the WM and JD-v2 sensors can be used to monitor soil moisture to enhance irrigation management following field calibrations. Furthermore, development of soil-specific field calibration curves for individual soil depths is strongly recommended to enhance the performance of each sensor type and improve the accuracy of in-season field θ_v measurements.