The value of soil temperature data versus soil moisture data for state, parameter, and flux estimation in unsaturated flow model

Abstract

This synthetic study explores the value of near-surface soil moisture and soil temperature measurements for the estimation of soil moisture and soil temperature profiles, soil hydraulic and thermal parameters, and latent heat and sensible heat fluxes using data assimilation (ensemble Kalman filter) in combination with unsaturated zone flow modeling (HYDRUS-1D), for 12 United States Department of Agriculture soil textures in a homogeneous and bare soil scenario. The soil moisture profile is estimated with a root mean square error (RMSE) of 0.04 cm³/cm³ for univariate soil temperature assimilation and 0.01 cm³/cm³ for univariate soil moisture assimilation. Soil temperature assimilation performs better for soils with higher clay content compared to soils with higher sand content. The latent and sensible heat fluxes are estimated with smaller RMSE for univariate soil temperature assimilation compared to univariate soil moisture assimilation for 8 out of 12 soil types. As the climate condition changes from hot semi-arid to sub-humid climate, the soil moisture assimilation performs better for high permeable soil but worse for low permeable soil. In summary, the findings suggest that for most soil texture classes, assimilating soil temperature in vadose zone models is skillful to improve latent heat flux, soil moisture profile, and soil hydraulic parameters. Joint assimilation with soil moisture can further enhance the accuracy of the model outputs for all range of soil texture and climate conditions.