Evaluation of 14 frozen soil thermal conductivity models with observations and SHAW model simulations

• 14 frozen soil thermal conductivity (FSTC) algorithms in land surface models were reviewed. These FSTC evaluated with measurements at temperature ≤−4 °C. The Becker 1992 algorithm performed the based on measured FSTC. incorporated into SHAW model for evaluation. No perform well, average bias ranging from −2.2 to 2.8 °C and −0.08 0.1 cm 3 −3 . Frozen (FSTC, λ eff ) is a critical thermophysical property that required variety of science engineering applications. Measurements soils are prone errors because currently available methods result phase change (i.e., thawing refreezing ice) affects estimated , especially near freezing point water (e.g., −4 0 °C) where rapid occurs. In addition, data few compared other physical properties. Therefore, many have been developed them numerical simulation programs calculating However, large discrepancies between simulated observed regimes reported. Previous studies either performance limited or simply programs. study has systematically assess included both observations simulations. this study, various compiled dataset consisting 331 27 seven made temperatures below provided best estimates measurements, but accuracy was not good RMSE = 0.46 W m −1 Bias -0.04 NSE 0.51). also Simultaneous Heat Water (SHAW) compare their effects simulating content two field sites contrasting textures USA. results showed all depths ranged differences liquid Generally no combined satisfactorily dynamic regime. Perspectives future discussed.