The diurnal course of soil moisture as measured by various dielectric sensors: Effects of soil temperature and the implications for evaporation estimates

Abstract

Soil moisture content, θ, of a bare and vegetated UK gravelly sandy loam soil (in situ and repacked in small lysimeters) was measured using various dielectric instruments (single-sensor ThetaProbes, multi-sensor Profile Probes, and Aquaflex Sensors), at depths ranging between 0.03 and 1 m, during the summers of 2001 (in situ soil) and 2002 (mini-lysimeters). Half-hourly values of evaporation, E, were calculated from diurnal changes in total soil profile water content, using the soil water balance equation. For the bare soil field, Profile Probes and ML2x ThetaProbes indicated a diurnal course of θ that did not concur with typical soil physical observations: surface layer soil moisture content increased from early morning until about midday, after which θ declined, generally until the early evening. The unexpected course of θ was positively correlated to soil temperature, Ts, also at deeper depths. Aquaflex and ML1 ThetaProbe (older models) outputs, however, reflected common observations: θ increased slightly during the night (capillary rise) and decreased from the morning until late afternoon (as a result of evaporation). For the vegetated plot, the spurious diurnal θ fluctuations were less obvious, because canopy shading resulted in lower amplitudes of Ts. The unrealistic θ profiles measured for the bare and vegetated field sites caused diurnal estimates of E to attain downward daytime and upward night-time values. In the mini-lysimeters, at medium to high moisture contents, θ values measured by (ML2x) ThetaProbes followed a relatively realistic course, and predictions of E from diurnal changes in vertically integrated θ generally compared well with lysimeter estimates of E. However, time courses of θ and E became comparable to those observed for the field plots when the soil in the lysimeters reached relatively low values of θ. Attempts to correct measured θ for fluctuations in Ts revealed that no generally applicable formula could be derived.