We used a machine learning approach to upscale CH₄ fluxes over time on non-waterlogged soil in a topographically complex mountain forest. Predicted CH₄ fluxes varied significantly across topographic positions, with greater uptake on ridges and slopes than on the plain and foot slopes. Recent past precipitations significantly influenced seasonal CH₄ uptake. Our findings highlight the role of topography and the potential of remote sensing and machine learning to map CH₄ fluxes.