V.S. Chernyshenko1, C.D. Belov2, L.B. Kulygin3, M.A. Chetyrkina4

1–4 Financial University under the Government of the Russian Federation (Moscow, Russia)
1 vschernyshenko@fa.ru, 2 constantinbelov5903@gmail.com, 3 Leonid2004.ru@mail.ru, 4 m.chetyrkina24@gmail.com

The escalating anthropogenic emissions of greenhouse gases, predominantly CO₂ and methane, are compromising ecological stability, driving climate change, increasing the frequency of extreme weather events, and accelerating biodiversity loss. Conventional carbon sequestration assessment methods face limitations when applied to Russia's boreal landscapes due to inadequate representativeness in training datasets.

This study aims to develop an integrated multilevel model for quantifying CO₂ absorption by forest ecosystems through the synergistic integration of satellite remote sensing data, meteorological observations, and soil parameters using machine learning techniques.

The developed model exhibits superior predictive accuracy for key carbon sequestration parameters. Implementation of ensemble algorithms and neural networks yielded highly precise estimations (RMSE 0.25–0.35 t/ha; R² 0.82–0.88). These performance metrics confirm the model's robustness and applicability across diverse forest biomes spanning extensive regions of the Russian Fe­deration.

The findings demonstrate the operational viability of artificial intelligence for large-scale environmental monitoring and carbon management, offering substantial potential for integration into national frameworks for ecological conservation planning.

Chernyshenko V.S., Belov C.D., Kulygin L.B., Chetyrkina M.A. A machine learning-based model for assessing carbon sequestration in russian forests using integrated satellite, meteorological, and soil data. Neurocomputers. 2025. V. 27. № 6. P. 7−16. DOI: 10.18127/j19997493-202506-01 (in Russian).

1. Intergovernmental Panel on Climate Change. (2022). Sixth Assessment Report. URL: https://www.ipcc.ch
2. International Energy Agency. (2023). Global Energy and CO₂ Status Report. URL: https://www.iea.org/
3. British Petroleum. (2022). BP Statistical Review of World Energy 2022. URL: https://www.bp.com/
4. European Commission, Eurostat. (2022). Greenhouse gas emission statistics. URL: https://ec.europa.eu/eurostat
5. United Nations. (2015). Paris Agreement. URL: https://unfccc.int/process-and-meetings/the-paris-agreement
6. Federal State Statistics Service. (2023). Statistical Compendium «Environmental Protection». URL: https://rosstat.gov.ru/
7. Ministry of Natural Resources and Ecology of Russia. (2022). State of the Environment and Environmental Protection Reports of the Russian Federation. URL: https://www.mnr.gov.ru/
8. State Duma of the Russian Federation. (2021). Federal Law No. 296-FZ. URL: http://publication.pravo.gov.ru/
9. World Meteorological Organization. (2022). WMO Statement on the State of the Global Climate.  URL: https://public.wmo.int/
10. Food and Agriculture Organization of the United Nations. (2021). State of the World’s Forests. URL: https://www.fao.org/publications
11. Matiza C., Mutanga O., Peerbhay K., Odindi J., & Lottering R. (2023), A systematic review of remote sensing and machine learning approaches for accurate carbon storage estimation in natural forests / Southern Forests: a Journal of Forest Science, 85(3-4)
12. Autumn Nguyen, Sulagna Saha, Machine Learning and Multi-source Remote Sensing in Forest Carbon Stock Estimation: A Review. URL: https://arxiv.org/abs/2411.17624
13. Advancing forest carbon stocks' mapping using a hierarchical approach. (2024). Scientific Reports. URL: https://www.nature.com/ articles/s41598-024-71133-8
14. Reiersen G., Dao D., Lütjens B., Klemmer K., Zhu X., & Zhang C. (2021). Tackling the Overestimation of Forest Carbon with Deep Learning and Aerial Imagery. URL: https://arxiv.org/abs/2107.11320
15. Remote Sensing, MDPI, 16[18], 3394. (2023). Comprehensive Multi-Temporal Satellite Data for Boreal Forest Carbon Sequestration Analysis. URL: https://www.mdpi.com/2072-4292/16/18/3394