Multiscale soil hydraulic parameterization for improved hydrologic simulations using SMAP

Abstract:

Scale-specific estimation of the “effective” soil hydraulic properties (SHP_eff) at remote-sensing (RS) footprint scale can help reduce uncertainties in water balance partitioning leading to improved streamflow simulations, groundwater (GW) recharge, evapotranspiration (ET) and rootzone soil moisture (SM) estimates. The SHP_eff capture the scale-specific relationship the soil hydraulic processes and the influence of various bio-geo-physical controls (topography, vegetation and climate, etc.) on SM dynamics. In this study, we i) parameterize SHP_eff at 3- and 36-KM for the rootzone soil profile for two watersheds in contrasting (semi-arid and arid) hydroclimates and ii) evaluate the influence of the “effective” parameterization on the watershed-scale water balance and accuracy of simulated hydrological variables. The study uses surface SM observations from SMAP and SMAP/Sentinel as target variables to parameterize the scale-specific rootzone SM retention curves using a stochastic framework using based on Markov Chain Monte Carlo (MCMC). Soil-texture based subpixel SHPs are used as the a-priori estimates for MCMC to simulate 1-D flow transport for the SMAP and SMAP/Sentinel footprint and estimate SHP_ef at the respective scales. The novel SHP_eff are then used for watershed-scale hydrologic simulations in a land-surface model at 3- and 36-KM resolution. The study evaluates the influence of scale, seasonality and watershed characteristics on the effective parameterization for the two watersheds. The accuracy of the simulated hydrologic variables using SHP_eff is compared with that of models with aggregated parameters based on soil texture. The challenges, opportunities and avenues of application of the proposed approach to land-surface modeling from regional-to-continental scale are discussed.