The paper presents a comparison of modern approaches to the interpolation of the spatial distribution of the chemical elements in the upper soil layer by the example of heavy metals chromium (Cr) and copper (Cu). Spots with an abnormally high Cr content were found on the examined area. Copper, on the contrary, was distributed evenly. The study is based on the data from soil screening in Tarko-Sale, Russia. For the prediction were selected models based on artificial neural networks (multilayer perceptron (MLP)), random forest (RF) algorithms, and the hybrid method in which MLP is used as a classifier (tree) (RMLPF). Models have been implemented in MATLAB. Approaches involving artificial neural networks (MLP and RMLPF) turned out to be more precise for abnormally distributed Cr. Models based on the RF algorithm are more precise for uniformly distributed Cu. In general, the proposed RMLPF model is showed the best results.
Keywords: artificial neural networks, random forest, random perceptron forest, interpolation, heavy metals, soil.

Buevich Alexander Gennadyevich, engeneer Institute of Industrial Ecology UB RAS. E-mail: bagalex3@gmail.com.

Moskaleva Anastasia Sergeevna, student The Ural State University, Institute of Physics and Technology, Department of Technical Physics. Ekaterinburg, ul. Mira, d. 19. E-mail: amoskalyova03@gmail.com.

Kosachenko Alexandra Ilyinichna, student The Ural State University, Institute of Physics and Technology, Department of Technical Physics. Ekaterinburg, ul. Mira, d. 19. E-mail: alleshch7@gmail.com.

Shichkin Andrey Vasilevich, engeneer Institute of Industrial Ecology UB RAS. Ekaterinburg, ul. S. Kovalevskoj, d. 20. E-mail: and@ecko.uran.ru.

Sergeev Alexander Petrovish, Ph.D. Head of the laboratory of physics and ecology Institute of Industrial Ecology UB RAS. Ekaterinburg, ul. S. Kovalevskoj, d. 20. E-mail: alexanderpsergeev@gmail.com.

1. Analysis of the statistical dependencies of the distribution of pollutants in the surface soil layer of urbanized territories using mathematical models (LUR method ) / A.G. Buevich, A.M. Safina, A.P. Sergeev, A.N. Varaksin, A.N. Medvedev // Geoecology. 2015. No. 3. P. 268-279.
2. Demyanov V.V., Savelyev E.A. Geostatistics: Theory and Practice / Ed. R.V. Harutyunyan ; Institute for Problems of Safe Development of Atomic Energy RAS. M. : Science, 2010. 327 p.
3. Nikitin A.A., Cheremisina E.N., Malinina S.S. Neural network modeling of the depth of the contact surface over a complex of geophysical fields // Geoinformatics. 2018. No. 1. P. 41-42.
4. Prediction of air pollution peaks generated by urban transport networks / M. Bell, A.S. Bergantino, M. Catalano, F. Galatioto // Working papers. SIET, 2015.
5. Kottur S.V., Mantha S.S. An Integrated Model using Artificial Neural Network (ANN) and Kriging for Forecasting Air Pollutants using Meteorological Data // International Journal of Advanced Research in Computer and Communication Engineering. 2015. V. 4, Issue 1. P. 146-152.
6. De Souza A., Aristones F., Goncalves F.V. Modeling of Surface and weather effects ozone concentration using neural networks in West Center of Brazil // Journal of Climatology & Weather Forecasting. 2015. V. 3, Issue 1 (123).
7. Shepherd A.J. Second-Order Methods for Neural Networks: Fast and Reliable Training Methods for Multi-Layer Perceptrons. London : Springer-Verlag, 1997. 145 p.
8. Rosenblatt F. The perceptron: A probabilistic model for information storage and organization in the brain // Psychological Review. 1958. V. 65 (6). P. 386-408.
9. Extensive evaluation of neural network models for the prediction of NO2 and PM10 concentrations, compared with deterministic modelling system and measurements in central Helsinki / J. Kukkonen, L. Partanen, A. Karppinen, J. Ruuskanen, H. Junninen, M. Kolehmainen, H. Niska, S. Dorling, T. Chatterton, R. Foxall, G. Cawley // Atmospheric Environment. 2003. V. 37 (32). P. 4539-4550.
10. Progress in developing an ANN model for air pollution index forecast / D. Jiang, Y. Zhang, X. Hu, Y. Zeng, J. Tan, D. Shao // Atmospheric Environment. 2004. V. 38, Issue 40. P. 7055-7064.
11. Artificial neural networks forecasting of PM2.5 pollution using air mass trajectory based geographic model and wavelet transformation / X. Feng, Q. Li, Y. Zhu, J. Hou, L. Jin, J. Wang // Atmospheric Environment. 2015. V. 107. P. 118-128.
12. Estimation of heavy metal sorption in German soils using artificial neural networks / J. Anagu, J. Ingwersen, J. Utermann, T. Streck // Geoderma. 2009. V. 152. P. 104-112.
13. Study on Spatial Distribution of Soil Heavy Metals in Huizhou City Based on BP–ANN Modeling and GIS / Y. Li, C. Li, J.-J. Tao, L.-D. Wang // Procedia Environmental Sciences. 2011. V. 10. P. 1953-1960.
14. Hilko O.S., Kundas S.P., Gishkeluk I.A. Radionuclides migration modelling using artificial neural networks and parallel computing // European water. 2012. V. 39. P. 3-13.
15. Falamaki A. Artificial neural network application for predicting soil distribution coefficient of nickel // Journal of Environmental Radioactivity. 2013. V. 115. P. 6-12.
16. High Variation Topsoil Pollution Forecasting in the Russian Subarctic: Using Artificial Neural Networks Combined with Residual Kriging / D.A. Tarasov, A.G. Buevich, A.P. Sergeev, A.V. Shichkin. // Applied Geochemistry. 2017. V. 88, Part B. P. 188-197.
17. Breiman L. Random forests // Mach. Learn. 2001. V. 45. P. 5-32.
18. Friedman H., Meulman J.J. Multiple additive regression trees with applica-tion in epidemiology // Stat. Med. 2003. V. 22. P. 1365-1381.
19. Lawrence R.L., Wood S.D., Sheley R.L. Mapping invasive plants using hyper-spectral imagery and Breiman Cutler classifications (Random Forest) // Remote Sens. Environ. 2006. V. 100. P. 356-362.
20. Downscaling MODIS-derived maps using GIS and boosted regression trees: the case of frost occurrence over the arid Andean highlands of Bolivia / R. Pouteau, S. Rambal, J.-P. Ratte, F. Gogé, R. Joffre, T. Winkel // Remote Sens. Environ. 2011. V. 115. P. 117-129.
21. Wetland vegetation distribution modelling for the identification of constraining environmental variables / J. Peters, N.E.C. Verhoest, R. Samson, P. Boeckx, B. De Baets // Landsc. Ecol. 2008. V. 23, Issue 9. P. 1049-1065.
22. Froeschke J.T., Froeschke B.F. Spatiotemporal predictive model based on environmental factors for juvenile spotted seatrout in Texas estuaries using boosted regression trees // Fish. Res. 2011. V. 111. P. 131-138.
23. Carslaw D.C., Taylor P.J. Analysis of air pollution data at a mixed source location using boosted regression trees // Atmospheric Environment. 2009. V. 43. P. 3563-3570.
24. Soil organic carbon concentrations and stocks on Barro Colorado Island-digital soil mapping using Random Forests analysis / R. Grimm, T. Behrens, M. Märker, H. Elsenbeer // Geoderma. 2008. V. 146. P. 102-113.
25. Spatial distribution of soil organic carbon stocks in France / M.P. Martin, M. Wattenbach, P. Smith, J. Meersmans, C. Jolivet, L. Boulonne, D. Arrouays // Biogeosciences. 2011. P. 1053-1065.
26. Spatial assessment of soil organic carbon density through random forests based imputation / K. Sreenivas, G. Sujatha, K. Sudhir, D.V. Kiran, M.A. Fyzee, T. Ravisankar, V.K. Dadhwal // J. Indian Soc. Remote Sen. 2014. V. 42. P. 577-587.
27. Estimation of total organic carbon storage and its driving factors in soils of Bavaria (southeast Germany) / M. Wiesmeier, F. Barthold, P. Spörlein, U. Geuß, E. Hangen, A. Reischl, B. Schilling, G. Angst, M. von Lützow, I. Kögel-Knabner // Geoderma Reg. 2014. V. 1. P. 67-78.
28. A comparative assessment of support vector regression, artificial neural networks, and random forests for predictingand mapping soil organic carbon stocks across an Afromontane landscape / K. Were, D.T. Bui, O.B. Dick, B.R. Singh // Ecological Indicators. 2015. V. 52. P. 394-403.
29. Digital mapping of soil organic matter for rubber plantation at regional scale: An application of random forest plus residuals kriging approach / Peng-Tao Guo, Mao-Fen Li, Wei Luo, Qun-Feng Tang, Zhi-Wei Liu, Zhao-Mu Lin // Geoderma. 2015. 237-238. P. 49-59.
30. Comparison of boosted regression tree and random forest models formapping topsoil organic carbon concentration in an alpine ecosystem / Ren-Min Yang, Gan-Lin Zhang, Feng Liu, Yuan-Yuan Lu, Fan Yang, Fei Yang, Min Yang, Yu-Guo Zhao, De-Cheng Li // Ecological Indicators. 2016. V. 60. P. 870-878.