Data-driven modelling of methane fluxes across a mire complex based on replicated eddy covariance measurements and spatially-resolved driver information

• Data-driven models reduce mire complex methane flux uncertainty by up to 50 %. • Simple biomet model approached the performance of spatially-explicit footprint approach. • NDVI emerged as most important predictor of temporal variations in methane fluxes. • Accounting for peat C:N ratios improved mire complex methane flux estimations. Northern mires are significant natural sources of atmospheric methane (CH 4 ), yet estimating CH 4 emissions remains challenging due to their complex spatio-temporal dynamics. While eddy covariance (EC) measurements provide valuable insights into ecosystem-scale CH 4 fluxes (FCH 4 ) over mire areas typically < 0.05 km², the predictability of FCH 4 at the mesoscale (∼0.5 – 20 km²) of a mire complex based on single-site EC measurements has not been explored. In this study, we utilized a network of four EC towers and developed a machine learning approach that integrates these EC data with comprehensive spatial information on drivers to predict FCH 4 across a boreal mire complex in Northern Sweden. For this purpose, environmental driver variables were mapped and area-weighted within dynamic EC flux footprints and related to FCH 4 in a spatially-explicit random forest model (‘footprint-based model’). For comparison, we also considered a standard random forest model used for gapfilling of FCH 4 data that is based on environmental measurements from fixed sensor locations (‘biomet model’). For both models, variable importance analysis revealed NDVI as the strongest predictor of temporal FCH 4 patterns, followed by air pressure, soil temperature and water table. Adjusting for site-specific carbon-to-nitrogen (C:N) ratios substantially improved model performance. Both models significantly improved estimates of the mire complex average FCH 4 compared to simple extrapolation of single-site measurements, reducing the uncertainty from ∼22 % in 2022 and 32 % in 2023 to <10 % and <25 % for the footprint-based model, and to <11 % and <30 % for the biomet model, respectively. Overall, our findings suggest that the comprehensive spatially-resolved driver information resulted in only marginally improved model performance at our study site. In comparison, the biomet model offers practical advantages through simpler implementation and wider applicability. However, we encourage testing the footprint-based model approach at other more heterogenous sites where it might become superior due to its ability to account for complex site conditions.