Bulletin Samara State Agricultural AcademyBulletin Samara State Agricultural Academy1997-32252782-4225Samara State Agrarian University56836110.55170/19973225_2023_8_3_28Research ArticleForecasting of soil-hydrological parameters with the help of artificial intelligenceChelovechkovaAnna V.<p>Candidate of Biological Sciences</p>chelovechkova_2011@mail.ruhttps://orcid.org/0000-0003-0236-9402KomissarovaIrina V.<p>Candidate of Biological Sciences</p>ir.komissarova@mail.ruhttps://orcid.org/0000-0002-9106-5509MiroshnichenkoNatalya V.<p>Candidate of Agricultural Sciences</p>natalya.mir79@mail.ruhttps://orcid.org/0000-0002-7064-6929Kurgan State UniversityKurgan State Agricultural Academy named after T. S. Maltsev – branch of the Kurgan State University220820238328361408202314082023Copyright © 2023, Chelovechkova A.V., Komissarova I.V., Miroshnichenko N.V.2023<p>The purpose of the study is the development of a mathematical model of moisture transfer in soils. The studies were carried out on leached, low-humus, medium-thick, light loamy chernozem. To build the model, the main physical properties of the soil (granulometric composition, density, porosity) were taken. At the initial stage of the work, the construction of the basic hydrophysical characteristic was carried out by laboratory and calculation methods. Since the cost of obtaining soil-hydrophysical information, especially taking into account spatiotemporal variability, is usually high, the urgent task is to simplify and reduce the cost of obtaining soil-hydrophysical information. Therefore, in order to reduce the time spent, a hardware-software technique for constructing the studied graphs was developed. Soil-hydrological constants were determined from the obtained graphs. The values of the main physical properties of the soil and the obtained soil-hydrological constants made it possible to select data for working on a predictive model using artificial intelligence. As a result of forecasting using the TensorFlow framework with single-layer linearization, predictive values were obtained for the maximum soil hygroscopicity W<sub>mg</sub>, the maximum molecular water capacity W<sub>mw</sub>, the lowest water capacity W<sub>nv</sub>, for the value of the yield strength W<sub>pt</sub> and porosity . Based on the results obtained, it can be noted that for the parameter of maximum molecular water capacity, the model predicts a fairly well predictable parameter. The loss function used makes it possible to see that the obtained values of the W<sub>mg </sub>parameter are close to the test data. Most of the values center around the value zero, getting bigger or smaller with about equal probability. In the case of the W<sub>mv</sub> parameter, the loss function takes on a constant value starting from epoch 75. Observation of minor outliers allows us to speak about the reliability of the predictive model of the parameter. In the case of porosity , a large spread of predicted values should be noted. This may be due to the fact that the total porosity was determined without taking into account the change in the profile of active pores occupied by capillary water and aeration pores.</p>modelinggranulometric compositionbasic hydrophysical characteristicsmoisture transferdensityporosityмоделированиегранулометрический составосновная гидрофизическая характеристикавлагопереносплотностьпористость[Kabakov, Z. K. & Kabakov, P. Z. (2005). Application of mathematical modeling in the development of new technologies and in education. Uspekhi sovremennogo yestestvoznaniya (The successes of modern natural science), 6, 85–86 (in Russ).][Mikayilov, F. D. (2014). Modeling of some soil processes. Vestnik Altayskogo gosudarstvennogo agrarnogo universiteta (Bulletin of the Altai State Agrarian University), 7(117), 59–64 (in Russ).][Savin, I. Yu., Zhogolev, A. V. & Prudnikova, E. Yu. (2019). Modern trends and problems of soil cartography. Pochvovedenie (Edaphology), 5, 517–528. doi: 10.1134/S0032180X19050101 (in Russ).][Tiwari, S., Saha, S. & Kumar, S. (2015). Prediction modeling and mapping of soil carbon content using artificial neural network, hyperspectral satellite data and field spectroscopy. Advances in Remote Sensing, 4, 63–72. doi: 10.4236/ars.2015.41006.][Rybalsky, N. N. & Ivanov, A. V. (2020). Digital soil model. Moscow: Publishing Solutions (in Russ).][Mikayilov, F. D. (2014). Direct and inverse problems of the salt transfer model under conditions of stationary water-salt regime of soils. Permskii agrarnii vestnik (Perm Agrarian Journal), 3(7), 52–59 (in Russ).][Golovanov, A. I. & Sotneva, N. I. (2009). Mathematical simulation of water and salt transfer in geosystems of solonetzic soils in the northern caspian region. Eurasian Soil Science, 42, 3, 251–266. doi: 10.1134/S1064229309030028.][Egorov, V. P. & Krivonos, L. A. (1995). Soils of the Kurgan region. Kurgan: Zauralie Publishing House (in Russ).][Chelovechkova, A. V. (2014).Construction of curves of water-holding capacity of leached chernozems of the Trans-Urals by instrumental and calculation methods. Agrarnyi vestnik Urala (Agrarian Bulletin of the Urals), 3 (121), 13–18 (in Russ).][Chelovechkova, A. V. & Komissarova, I. V. (2013).The main hydrophysical characteristics of the leached chernozems of the Trans-Urals. Development of scientific, creative and innovative activities of youth '13: collection of articles based on the materials of the V All-Russian scientific and practical correspondence conference of young scientists. (pp. 22–24). Kurgan (in Russ).][Chelovechkova, A. V. (2022). The use of hydrophysical properties to characterize the soils of the Kurgan region. Extended abstract of candidate’s thesis. Kurgan (in Russ).][Chelovechkova, A. V., Komissarova, I. V. & Eremin, D. I. (2018). Forecasting Water Saturation of Fill Grounds in Urban Infrastructure Conditions by Mathematical Modeling Based on the Main Hydrophysical Characteristic. Journal of Environmental Management and Tourism, 9, 3(27), 485–490. doi: 10.14505/jemt.v9.3(27).08.][Chelovechkova, A. V., Komissarova, I. V. & Eremin, D. I. (2018). Using Basic Hydrophysical Characteristics of Soils in Calculating Capacity of Water-Retaining Fertile Layer in Recultivation of Dumps of Mining and oil Industry. Earth and Environmental Science. LOP Conference Series. Saint-Petersburg Mining University, 194, 9, 092004. doi: 10.1088/1755-1315/194/9/092004.]