Time-domain and frequency-domain reflectometry type soil moisture sensor performance and soil temperature effects in fine- and coarse-textured soils

Abstract

The performances of six time-domain reflectometry (TDR) and frequency-domain reflectometry (FDR) type soil moisture sensors were investigated for measuring volumetric soil-water content (θv) in two different soil types. Soil-specific calibration equations were developed for each sensor using calibrated neutron probe-measured θv. Sensors were also investigated for their performance response in measuring θv to changes in soil temperature. The performance of all sensors was significantly different (P<0.05) than the neutron probe-measured θv, with the same sensor also exhibiting variation between soils. In the silt loam soil, the 5TE sensor had the lowest root mean squared error (RMSE) of 0.041 m3/m3, indicating the best performance among all sensors investigated. The performance ranking of the other sensors from high performance to low was: TDR300 (High Clay Mode), CS616 (H) and 10HS, SM150, TDR300 (Standard Mode), and CS616 (V) (H: horizontal installation and V: vertical installation). In the loamy sand, the CS616 (H) performed best with an RMSE of 0.014 m3/m3 and the performance ranking of other sensors was: 5TE, CS616 (V), TDR300 (S), SM150, and 10HS. When θv was near or above field capacity, the performance error of most sensors increased. Most sensors exhibited a linear response to increase in soil temperature. Most sensors exhibited substantial sensitivity to changes in soil temperature and the θv response of the same sensor to high vs. normal soil temperatures differed significantly between the soils. All sensors underestimated θv in high temperature range in both soils. The ranking order of the magnitude of change in θv in response to 1°C increase in soil temperature (from the lowest to the greatest impact of soil temperature on sensor performance) in silt loam soil was: SM150, 5TE, TDR300 (S), 10HS, CS620, CS616 (H), and CS616 (V). The ranking order from lower to higher sensitivity to soil temperature changes in loamy sand was: 10HS, CS616 (H), 5TE, CS616 (V), SM150, and TDR300 (S). When the data from all sensors and soils are pooled, the overall average of change in θv for a 1°C increase in soil temperature was 0.21 m3/m3 in silt loam soil and -0.052 m3/m3 in loamy sand. When all TDR- and FDR-type sensors were pooled separately for both soils, the average change in θv for a 1°C increase in soil temperature for the TDR- and FDR-type sensors was 0.1918 and -0.0273 m3/m3, respectively, indicating that overall TDR-type sensors are more sensitive to soil temperature changes than FDR-type sensors when measuring θv.