Implications of CO emissions on the main land and forest uses in the Brazilian Amazon.

The emission of soil carbon dioxide (CO) in agricultural areas is a process that results from the interaction of several factors such as climate, soil, and land management practices. Agricultural practices directly affect the carbon dynamics between the soil and atmosphere. Herein, we evaluated the temporal variability (2020/2021 crop season) of soil CO emissions and its relationship with related variables, such as the CO flux model, enhanced vegetation index (EVI), gross primary productivity (GPP), and leaf area index (LAI) from orbital data and soil temperature, soil moisture, and soil CO emissions from in situ collections from native forests, productive pastures, degraded pastures, and areas of high-yield potential soybean and low-yield potential soybean production. A significant influence (p < 0.01) was observed for all variables and between the different land uses and occupation types. September and October had lower emissions of soil CO and low means of soil moisture and soil temperature, and no differences were observed among the treatments. On the other hand, there was a significant effect of the CO flux model in productive pastures, high-yield potential soybean areas, and low-yield potential soybean areas. The months with the highest CO flux values in the model, regardless of land use and land cover, were October and November, which is the beginning of the rainy season. There were positive correlations between soil CO emissions and GPP (0.208), LAI (0.354), EVI (0.363), and soil moisture (0.280) and negative correlations between soil CO emissions and soil temperature (-0.240) and CO flux model (-0.314) values. Land use and land cover showed negative correlations with these variables, except for the CO flux model variable. Soil CO emission values were lower for high-yield potential soybean areas (averages from 0.834 to 6.835 μmol m s) and low-yield potential soybean areas (from 0.943 to 5.686 μmol m s) and higher for native forests (from 2.279 to 8.131 μmol m s), whereas the opposite was true for the CO flux model.