Validation of the temperature of sandy soil in the block of the active layer of the land of the INM RAS-MSU model according to field observations

The soil temperature is one of the most important characteristics of climatic, soil-forming, permafrost and engineering-geological conditions. It determines the functioning of terrestrial ecosystems, allowing us to determine the sensitivity of landscapes to anthropogenic impact, changes in the natural environment and climate fluctuations. The study of real natural processes, phenomena and objects using the methods of mathematical modeling is topical. Interest is dictated by the growing attention of the scientific community to the problem of global warming. The reproduction of the temperature regime in the soil is an important problem for further research. Particularly, studying the thermodynamic processes in soils and the assessment of their reproduction in models are important objectives. This work is aimed at comparing the results of modeling the temperature regime of soils with data from field observations. The block of models of the active layer of land is used in the INM RAS-MSU climate model. Computational experiments require information on atmospheric forcing (ERA-5 reanalysis), as well as soil and hydrophysical characteristics, which, to fully support the model, were taken from the Global Soil Dataset for Earth System Models). The temperature data of sandy soil obtained with the help of atmospheric-soil measuring complexes [5] installed by the employees of the V. B. Sochava Institute of Geography of the Siberian Branch of the Russian Academy of Sciences in different landscapes of the Tunkinskaya depression (Republic of Buryatia) were used as reference information. The measurements are carried out from October 13, 2011 to the present with an hourly interval. Checking the adequacy of the constructed model to the initial data, the following statistical characteristics were analyzed: the average error, the average error modulus, the root mean square error, the correlation coefficient, and the Kling-Gupta complex efficiency. The first result of calculations of sandy soil temperature simulations recorded a large difference from field measurements. The error could be associated with an overestimated amount of incoming precipitation in winter. The developers corrected the work of the model, rewrote the code. The results of the second experiment (after correcting the model) on average for the year showed a decrease in the average calculation error (from 2 to 1.4), the error modulus (from 10.1 to 4.2), the root mean square error (from 11.1 to 4, 9) and an increase in the correlation coefficient (from 0.3 to 0.7). The KGE coefficient on average for the year changed from -1.4 to -0.2. The most significant improvement in the simulation result is in the winter season at all depths