Key Drivers of Soil CO ₂ , CH ₄

ABSTRACT

Carbon dioxide (CO ₂ ), methane (CH ₄ ), and nitrous oxide (N ₂ O) are the most significant greenhouse gases (GHGs), contributing substantially to climate change. Afforestation has been promoted as a nature‐based solution to enhance carbon sequestration, but the impacts of afforestation on atmospheric CH ₄ and N ₂ O are often ignored. In this research, we examined variations in soil CO ₂ , CH ₄ , and N ₂ O fluxes and applied machine learning models to explore the main drivers of those fluxes in the Conservation Reserve Program with bottomland broadleaf and upland conifer tree planting in the southeastern U.S. Bottomland broadleaf stands generally showed higher CO ₂ and CH ₄ emissions compared to conifers, particularly during the growing season. On average, CO ₂ fluxes were 11.5% higher, and CH ₄ fluxes were 70.8% higher in broadleaf stands relative to conifer stands. Random Forest and XGBoost machine learning models revealed that soil temperature, air temperature, air humidity, and water‐filled pore space (WFPS) were the dominant predictors for CO ₂ and CH ₄ in both forest types. However, for N ₂ O, most predictors, such as C:N ratio, total nitrogen, pH, and WFPS, showed weak or inconsistent effects, especially in conifer systems. Our results suggest that the climate mitigation benefits of afforestation programs depend not only on tree carbon sequestration but also on soil GHG responses. These findings underscore the importance of considering both tree biomass and soil GHG fluxes when targeting lands for afforestation to mitigate climate change. Incorporating site‐specific factors such as soil moisture regime, temperature sensitivity of soil GHG fluxes, and forest types (broadleaf and conifer) of selection into planning could help maximize net carbon benefits and avoid unintended trade‐offs.