Drivers of atmospheric CO2 concentration in southeast Brazil: Insights from land use change, vegetation, and climate factors

The atmospheric carbon dioxide (CO2) increase is one of the main causes of climate change and has been a global concern. Due to the world's necessity of understanding CO2 sources and sinks, satellites such as the Orbiting Carbon Observatory 2 (OCO-2) have been launched to monitor greenhouse gases (GHG). This study aimed to characterize the temporal variability and the main factors controlling (i.e., vegetation, climate, and land use changes) the column-averaged atmospheric CO2 (Xco2) in Southeast Brazil (925 thousand km2) between 2015 and 2021. Xco2 was detrended by subtracting the world's trend to study regional features (Xco2R). The vegetation component was sampled through the Solar Induced Chlorophyll Fluorescence (SIF) and Gross Primary Production (GPP). As regards climate parameters, we used Precipitation (Prec) and Global Irradiation (Qg), while the Land Use Change (LUC) was characterized by using the land use classification maps available to the region. Xco2R follows a seasonal pattern of maximum averages during autumn-winter and minimum averages spring-summer, with the highest value observed in October 2021 (393.4 ± 1.26 ppm) and the lowest in April 2015 (389.37 ± 1.7 ppm). For the other variables, this seasonality is inverted which means those have negative linear correlations with Xco2R ranging from −0.35 to −0.6. We observed an Xco2R temporal increase trend since 2017 and this tendency is higher over forest areas which could be related to deforestation observed in the same period. In addition, the emission data from the national GHG inventory estimated an increase of CO2 emissions due to land use changes in the same period and was able to explain around 50 % of the annual Xco2R increase in Southeast Brazil. Thus, vegetation has an essential role in the Xco2R absorption, and especially the land use change and deforestation were the main CO2 sources to the atmosphere on a regional scale.