From sink to source: high inter-annual variability in the carbon budget of a Southern African wetland

We report on three years of continuous monitoring of carbon dioxide (CO(2)) and methane (CH(4)) emissions in two contrasting wetland areas of the Okavango Delta, Botswana: a perennial swamp and a seasonal floodplain. The hydrographic zones of the Okavango Delta possess distinct attributes (e.g. vegetation zonation, hydrology) which dictate their respective greenhouse gas (GHG) temporal emission patterns and magnitude. The perennial swamp was a net source of carbon (expressed in CO(2)-eq units), while the seasonal swamp was a sink in 2018. Despite differences in vegetation types and lifecycles, the net CO(2) uptake was comparable at the two sites studied in 2018/2020 (−894.2 ± 127.4 g m(−2) yr(−1) at the perennial swamp, average of the 2018 and 2020 budgets, and −1024.5 ± 134.7 g m(−2) yr(−1) at the seasonal floodplain). The annual budgets of CH(4) were however a factor of three larger at the permanent swamp in 2018 compared to the seasonal floodplain. Both ecosystems were sensitive to drought, which switched these sinks of atmospheric CO(2) into sources in 2019. This phenomenon was particularly strong at the seasonal floodplain (net annual loss of CO(2) of 1572.4 ± 158.1 g m(−2)), due to a sharp decrease in gross primary productivity. Similarly, drought caused CH(4) emissions at the seasonal floodplain to decrease by a factor of 4 in 2019 compared to the previous year, but emissions from the perennial swamp were unaffected. Our study demonstrates that complex and divergent processes can coexist within the same landscape, and that meteorological anomalies can significantly perturb the balance of the individual terms of the GHG budget. Seasonal floodplains are particularly sensitive to drought, which exacerbate carbon losses to the atmosphere, and it is crucial to improve our understanding of the role played by such wetlands in order to better forecast how their emissions might evolve in a changing climate. Studying such hydro-ecosystems, particularly in the data-poor tropics, and how natural stressors such as drought affect them, can also inform on the potential impacts of man-made perturbations (e.g. construction of hydro-electric dams) and how these might be mitigated. Given the contrasting effects of drought on the CO(2) and CH(4) flux terms, it is crucial to evaluate an ecosystem's complete carbon budget instead of treating these GHGs in isolation. This article is part of a discussion meeting issue ‘Rising methane: is warming feeding warming? (part 2)’.