Simulation of evaporation, coupled liquid water, water vapor and heat transport through the soil medium

• Development of SWCT model for simulation of soil water flux (liquid water, water vapor and heat).
• Improved estimation of actual evaporation from soil surface using the Penman-Monteith method.
• Sensitivity analysis to identify variables requiring accurate measurement.

In this research, the model of SWCT (soil water content and temperature) was developed for simulation of fluxes of soil water (liquid water and water vapor) and heat. The governing equations for water and heat fluxes were solved using numerical finite difference method. Soil water content and temperature were numerically solved using a fully explicit method. The measurement devices of temperature sensors and TDR probes (for measuring water content) were installed at soil depths of 0.05, 0.35, and 0.5 m, at the selected site. The novelty of this study was the simulation of actual evaporation from the soil surface using the Penman–Monteith's equation, in which the value of saturated air vapor pressure was substituted by actual soil vapor pressure at the surface layer. This novelty allowed estimation of the actual evaporation rate based on soil water content and temperature and meteorological data. Simulated water contents followed reasonably well the measured values at three soil depths during the simulation period. Soil water content and temperature were numerically simulated using the SWCT model with coupling liquid water, water vapor, and heat transport and provided reasonably good results with the values of RMSE less than 0.017 cm3 cm−3 and 2.2 °C, respectively.The sensitivity analysis of the simulated soil temperatures were accomplished by changing radiation, input data of air temperatures and initial soil temperatures by ±20% and ±10% of measured values and the maximum values of RMSE changed by 65.7%, 140.3% and 190.7%, respectively. The results showed that air temperature, solar radiation, and initial soil temperature must be measured or predicted with high accuracy in order to produce low error values.