Effects of spatial heterogeneity within the eddy covariance (EC) footprint on up-scaled methane fluxes across multiple wetland sites

Methane (CH₄) is a strong greenhouse gas that is produced in anoxic soil conditions. Wetlands are the largest natural source of CH₄ globally because their anoxic soils provide suitable habitats for CH₄-producing Archaea. Both global and regional wetland CH₄ budgets remain unconstrained due to large uncertainties in wetland extent and high spatio-temporal variability in CH₄ dynamics that are in part driven by wetland spatial heterogeneity. High wetland spatial heterogeneity often results from the variability in microtopography, soil hydrological and chemical properties, and vegetation and microbial composition. Intertwined together, the different abiotic and biotic variables further contribute to the ratio between CH₄ production, consumption and transport processes in wetland soil, resulting in either net CH₄ emission or uptake. However, the contribution of different abiotic and biotic factors to CH₄ flux variability in wetlands remains unclear, increasing uncertainties in up-scaling CH₄ emissions from plot to ecosystem and regional scales. Therefore, including well-defined spatial heterogeneity into wetland CH₄ bottom-up estimates can help improve the regional and global CH₄ budget calculations.

This study investigates the effect of spatial heterogeneity on observed CH₄ emissions in ten different wetland sites with varying climatic conditions. Our approach will include up-scaling chamber measurements from different land cover classes to the level of the eddy covariance (EC) footprint. The compiled chamber datasets include both manual (n=5) and automatic (n=5) measurements that have been combined with EC measurements (FLUXNET-CH4 database) based on matching timestamps. First, the chamber observations will be compared to the corresponding EC measurements without accounting for spatial heterogeneity. Then, various remotely sensed environmental variables, such as leaf area index (LAI) and topographic wetness index (TWI), in high spatial resolution will be used to create land cover classes and combined with a modeled footprint to include spatial heterogeneity in the up-scaling of point-level measurements in all sites. 

Preliminary results suggest that the chamber and EC observations differ significantly in magnitude between seasons and sites. In general, chamber observations had a larger range than EC, which we expected, given that chambers capture finer spatial heterogeneity than EC. However, no consistent trends emerged in the difference in magnitude between chamber and EC. We expect that the inclusion of spatial heterogeneity into the footprint model will decrease the differences  between up-scaled chamber and EC observations for all sites. Notably, we expect that the inclusion of proxies for soil moisture, plant functional type (PFT) and aerenchyma  will improve footprint-level comparisons to chamber-level data. We will present updated comparisons of EC and chamber data with and without inclusion of spatial heterogeneity. Altogether, this study will establish a workflow for combining wetland CH₄ data from different measurement types (EC and chamber) and will allow global syntheses to use more of the available data to constrain CH₄ budgets.