Influences of thermal dispersion on soil water flux estimates using heat pulse technique in saturated soils

The heat-pulse technique (HPT) shows good potentials for in situ determinations of soil water flux (J), but its applications are limited because of J underestimates. This study aims to experimentally investigate the influences of thermal dispersion on J estimates using HPT in packed sand, silt loam, and sandy clay loam with onedimensional saturated water flow, hypothesizing that J underestimates are caused by neglecting heat dispersion in heat transport model that arises from the heterogeneity of water velocities within and between water-filled soil pores. The results indicated that J estimates exhibited good linearity with the measurements (R-2 > 0.95). However, they were biased toward underestimates. The thermal dispersion's dependence on J was described by a power function. When heat dispersion was considered in the traditional heat conduction-advection model, the traditional ratio method yielded a higher accuracy with relative errors reduced by 45.0-74.3%, the bias reduced by 47.1-74.0% and the root mean square error reduced by 48.0-68.8%. In terms of the range of J herein, the maximum Keith Jirka Jan (KJJ) numbers obtained were 6.8%, 10.4%, and 13.4% for sand, silt loam, and sandy clay loam, respectively. Moreover, their corresponding J thresholds were 41.7, 21.1, and 12.3 mu m s(-1), respectively