Southern Ocean carbon trends: Sensitivity to methods

Abstract

The Southern Ocean is highly under-sampled with respect to variables needed to assess trends in air-sea carbon dioxide fluxes. Multiple investigators have made use of sparse measurements of surface ocean partial pressure of CO2 (pCO2), but there remains a lack of consensus as to recent trends in the carbon sink here. We analyze impacts of specific methodological choices on pCO2 trends, with a focus on regional binning, time series length, and spatial biases. We find the most robust approach to include binning based on geographic areas, the removal of spatial gradients in climatological pCO2, the use of annual-average data, and the consideration of trends across multiple combinations of start and end years. Applying this approach, we find that Southern Ocean carbon uptake slowed from about 1990 to 2006 and subsequently strengthened from 2007 to 2010. More data are needed to reduce uncertainties and increase our ability to diagnose change in the carbon sink. Key Points Methodological choices have substantial impact on ocean pCO2 trend estimatesRegional binning, data selection, and spatial bias correction are criticalSouthern Ocean carbon uptake slowed until 2006 but has strengthened since