Evaluation of the hydro-thermodynamic soil-vegetation scheme on the basis of observations and a Galerkin type finite element numerical scheme

Abstract

The Hydro-Thermodynamic Soil- vegetation Scheme (HTSYS), coupled in a two-way mode with the PennState/National Center for Atmospheric Research (NCAR) Mesoscale Meteorological Model generation 5 (MM5), has been evaluated for a 5 day typical snow-melt period using the Baltic Sea experiment meteorological data center's soil temperature, snow depth, and precipitation datasets. The HTSVS-MM5 evaluation investigates the coupled system's sensitivity to two cloud models and two radiation models, with their cross effects presented along with skill scores for snow depth changes. The coupled model satisfactorily predicts the soil temperature diurnal course cycles, changes in the snow depths, and accumulated precipitation. HTSVS's soil model has been further tested and evaluated in an offline mode for the advanced numerical treatment for the Partial Differential Equations (PDEs) using soil temperature datasets from three sites at Council, Alaska. A Galerkin Weak Finite Element (GWFE) method was tested and evaluated for the numerical treatment of PDEs and the predictions were compared against the existing Crank-Nicholson finite differences scheme (CNFD). GWFE solutions exhibit a remarkable soil temperature predictability, better capture the temperature peaks, and yield non-diffuse and non-oscillatory solutions for relatively high convection dominated regimes, while CNFD performs comparably well in diffusion dominated regimes with a lower computational burden.