Assessment of Modeling Tree Stem Methane Fluxes Using Gridded Environmental Data Sets

Abstract

Tree stems are increasingly recognized as important components of methane (CH ₄ ) production and consumption, yet their contribution to the terrestrial CH ₄ budget remains highly uncertain due to sparse observations and challenges in upscaling in situ measurements. Here, we model tree stem CH ₄ (TSCH ₄ ) fluxes across wetland and upland ecosystems using multi‐source gridded environmental data. We compiled 13,547 TSCH ₄ flux observations from 90 published studies, supplemented by in situ measurements from two sites in 2025. Feature importance analysis identified the key structural and environmental drivers of TSCH ₄ fluxes in wetland and upland ecosystems. Random forest models were developed using five predictors selected through feature selection and evaluated with a combined 10‐fold cross‐validation for upland and wetland TSCH ₄ fluxes separately. Results suggest that tree structural traits (e.g., sampling height, diameter at breast height) and subsurface soil variables (e.g., deep soil temperature and moisture (100–289 cm), and belowground biomass) are key predictors of TSCH ₄ fluxes. The models show reasonable predictive skill, with a mean coefficient of determination ( R ² ) of 0.65 and a root mean square error (RMSE) of 3.26 nmol CH ₄ m ⁻²  s ⁻¹ for wetlands, and a mean R ² of 0.64 and an RMSE of 0.70 nmol CH ₄ m ⁻²  s ⁻¹ for uplands. The models capture the seasonal variability of TSCH ₄ fluxes in both wetland and upland ecosystems, with higher and later peaks in wetlands. The models demonstrate the potential for upscaling TSCH ₄ by linking in situ measurements with gridded environmental predictors, improving our understanding of the role of tree stems in the terrestrial CH ₄ budget.