Extension Of Thermal Conductivity Parameterization Schemes To Simulate Soil Hydrothermal Properties Using HYDRUS-1D Freeze-thaw Model

Soil internal hydrothermal movement is concerned with surface plant growth,ground-air energy balance,and stability of hydraulic buildings.As key soil hydrothermal parameters,soil thermal conductivity（λ）and moisture characteristic curve（SWRC）are difficult and costly to determine,therefore,various prediction models（schemes）are often used to assist in obtaining or improving their accuracy,and the related parameterization schemes have become key input parameters for soil hydrothermal simulation programs such as HYDRUS.The effects of differentλschemes and SWRC schemes on HYDRUS simulated soil hydrothermal results are unclear,and theλschemes,of which there are only two in HYDRUS,have not received sufficient attention.Therefore,this study proposes to select 24λschemes（including 2HYDRUS built-in schemes）used in 39 mainstream surface,hydrological,and soil-vegetation-atmosphere transport models to codify them into the HYDRUS-1D freeze-thaw model,and then combine them with 4 single-pore SWRC schemes,and then use 5 months of measured data from two 100 cm profiles,A and C,at a test site in the United States to simulate dynamic changes of soil temperature and water content to evaluate the effects of differentλschemes and SWRC schemes on the simulation of soil hydrothermal movement under freeze-thaw conditions.The following main conclusions were obtained:（1）Different SWRC schemes andλschemes have a certain degree of influence on the simulation of soil hydrothermal movement,mainly in terms of soil temperature and water content.In terms of temperature simulations,the analysis in Chapters 3 and 4 shows that eight thermal conductivity schemes（i.e.,DV1963,CS1981,CS1984,CH1987,BB1992,HL1998,DP1998,SA1998 and DI2009）simulated soil temperatures in high agreement with the observed values and significantly better than the two built-in schemes of HYDRUS-1D（CG1985 and CH1987）.Through model effect evaluation and analysis,the simulation accuracy of these eight thermal conductivity schemes is very high,the calculated R is greater than 0.98,the RMSE is kept within 1°C,the SD is also close to the observed value,and the average temperature deviation of the simulation is±1°C.However,some schemes are particularly poorly simulated,and some combinations have a simulation deviation of about6°C,which may be caused by texture and weather-driven data.It was also found that there was no significant stratification in the temperature simulation for each case,indicating that the temperature simulation is less affected by texture when HYDRUS simulates hydrothermal.（2）Soil water content simulation results are less affected by theλscheme and more affected by the SWRC scheme.For Site A,where the soil texture is relatively homogeneous,the simulation results of 24λschemes are basically the same,and the variation of soil liquid water content with time and depth for each scheme is not large,and the simulation accuracy R is in the range of 0.1～0.3;for Site C,where the soil texture is relatively complex,the soil moisture parameters vary greatly at different depths,and the simulated soil liquid water content with time and depth for four different SWRC schemes is relatively large,with R around 0.4.The variation of soil liquid water content with time and depth is relatively large with R around 0.4.Although the overall simulation deviation is large,with the maximum close to 0.11 cm³ cm^-3,HYDRUS still outperforms other models such as CLM,SHAW and Noah.This study deepens our understanding of the hydrothermal movement processes within the soil and the HYDRUS-1D numerical model,and provides a reference for the selection of moisture characteristic curves and thermal conductivity schemes in the future modeling of soil hydrothermal processes.