Hot spots for carbon emissions from Mediterranean fluvial networks during summer drought

Abstract

During summer drought, Mediterranean fluvial networks are transformed into highly heterogeneous landscapes characterized by different environments (i.e., running and impounded waters, isolated river pools and dry beds). This hydrological setting defines novel biogeochemically active areas that could potentially increase the rates of carbon emissions from the fluvial network to the atmosphere. Using chamber methods, we aimed to identify hot spots for carbon dioxide (CO2) and methane (CH4) emissions from two typical Mediterranean fluvial networks during summer drought. The CO2 efflux from dry beds (mean ± SE = 209 ± 10 mmolCO2m-2 d-1) was comparable to that from running waters (120 ± 33 mmol m-2 d-1) and significantly higher than from impounded waters (36.6 ± 8.5 mmol m-2 d-1) and isolated pools (17.2 ± 0.9 mmol m-2 d-1). In contrast, the CH4 efflux did not significantly differ among environments, although the CH4 efflux was notable in some impounded waters (13.9 ± 10.1 mmol CH4 m-2 d-1) and almost negligible in the remaining environments (mean\0.3 mmol m-2 d-1). Diffusion was the only mechanism driving CO2 efflux in all environments and was most likely responsible for CH4 efflux in running waters, isolated pools and dry beds. In contrast, the CH4 efflux in impounded waters was primarily ebullition-based. Using a simple heuristic approach to simulate potential changes in carbon emissions from Mediterranean fluvial networks under future hydrological scenarios, we show that an extreme drying out (i.e., a four-fold increase of the surface area of dry beds) would double the CO2 efflux from the fluvial network. Correspondingly, an extreme transformation of running waters into impounded waters (i.e., a twofold increase of the surface area of impounded waters) would triple the CH4 efflux. Thus, carbon emissions from dry beds and impounded waters should be explicitly considered in carbon assessments of fluvial networks, particularly under predicted global change scenarios, which are expected to increase the spatial and temporal extent of these environments