Estimates of Transboundary Transfer and Balance of Atmospheric Carbon Dioxide Fluxes in Sverdlovsk Region Using the Machine Learning Model

Abstract

In connection with the global warming problem caused by the increase in the concentration of greenhouse gases in the atmosphere, estimating the potential of different ecosystems for the sequestration of atmospheric carbon dioxide on both regional and global scales is a topical problem. In this work, the balance of natural carbon dioxide fluxes throughout Sverdlovsk region is considered. An integral estimate of net CO₂ absorbed from the atmosphere by regional ecosystems over 2020–2022 is obtained for the first time using an original NorthFlux machine learning model, where spectra data from the MODIS satellite sensor, meteorological data of retrospective climate analysis, and satellite data on the classification of the underlying surface vegetation are used as input. The data on the amount of anthropogenic CO₂ emissions is taken from the inventory of greenhouse gas emissions in Sverdlovsk region. The transboundary transfer of carbon dioxide is estimated using a balance equation for CO₂ fluxes in the atmospheric column and data on the average annual rate of increase in the CO₂ concentration in the regional atmosphere obtained from ground-based Bruker IFS 125M high-resolution measurements at the Kourovka Astronomical Observatory in 2012–2024. It has been found that the sequestration of atmospheric CO₂ by ecosystems of Sverdlovsk region ranges from 10.9 to 15.2% and the transboundary transfer of CO₂ through the region boundaries to neighboring regions ranges from 72.5 to 76.7% of the annual industrial CO₂ emissions in the region. The NorthFlux machine learning model can be useful for estimating the sequestering potential of ecosystems in other regions of the planet.