Numerical modeling of soil temperature variation under an extreme desert climate

•Soil temperature variation was numerically estimated from the energy balance at the surface.•A maximum value of 0.61 °C was obtained for the mean absolute error at 1 m depth.•Soil surface water content greatly influences the model's accuracy regardless of depth. The soil temperature profile is useful for different engineering, architecture, and agriculture areas. In recent years, the growth of direct uses of geothermal energy has increased the interest in soil temperature prediction models. This research paper numerically models the soil temperature variation under extreme desert climate conditions and compares it with experimental data measured on-site to evaluate the reliability of numerical models based on the one-dimensional transient state heat conduction equation as a tool for the evaluation of low enthalpy geothermal resource using meteorological data and soil thermal properties. From the study results, it can be concluded that the numerical models present temperature values close to those measured experimentally. The model that presents the more significant mismatches concerning the experimental data is the 1 m model, presenting the following statistical values; R2 of 0.97, mean absolute error of 0.61 °C, RMSE of 0.76 °C, NRMSE of 3.01%, and a MBIAS of 0.23. Although the numerical models present values are close to those measured, some coefficients, such as the soil surface water content greatly influence the model's accuracy regardless of depth.