B13D-0226:
In-lake carbon dioxide concentration patterns in four distinct phases in relation to ice cover dynamics

Monday, 15 December 2014
Blaize A Denfeld, Marcus Wallin, Erik Sahlee, Sebastian Sobek, Jovana Kokic, Hannah Chmiel and Gesa A. Weyhenmeyer, Uppsala University, Uppsala, Sweden
Abstract:
Global carbon dioxide (CO2) emission estimates from inland waters include emissions at ice melt that are based on simple assumptions rather than evidence. To account for CO2 accumulation below ice and potential emissions into the atmosphere at ice melt we combined continuous CO2 concentrations with spatial CO2 sampling in an ice-covered small boreal lake. From early ice cover to ice melt, our continuous surface water CO2 concentration measurements at 2 m depth showed a temporal development in four distinct phases: In early winter, CO2 accumulated continuously below ice, most likely due to biological in-lake and catchment inputs. Thereafter, in late winter, CO2 concentrations remained rather constant below ice, as catchment inputs were minimized and vertical mixing of hypolimnetic water was cut off. As ice melt began, surface water CO2 concentrations were rapidly changing, showing two distinct peaks, the first one reflecting horizontal mixing of CO2 from surface and catchment waters, the second one reflecting deep water mixing. We detected that 83% of the CO2 accumulated in the water during ice cover left the lake at ice melt which corresponded to one third of the total CO2 storage. Our results imply that CO2 emissions at ice melt must be accurately integrated into annual CO2 emission estimates from inland waters. If up-scaling approaches assume that CO2 accumulates linearly under ice and at ice melt all CO2 accumulated during ice cover period leaves the lake again, present estimates may overestimate CO2 emissions from small ice covered lakes. Likewise, neglecting CO2 spring outbursts will result in an underestimation of CO2 emissions from small ice covered lakes.