The surface ocean CO2 mapping inter comparison project in its second phase

Though the ocean is known to be a large sink for anthropogenic carbon dioxide (CO₂), recent estimates of the ocean carbon uptake published within the Global Carbon Budget highlight a problematic development. Since about the year 2000, estimates based on global hindcast model simulations and upscaled pCO₂measurements from the surface ocean ship network (SOCAT) appear to diverge, leaving an unexplained gap between the two estimates equaling, in magnitude ~0.6-0.7 Pg C yr-1, or roughly the emissions of the EU27. Within the upscaled measurement estimate, the role of data sparsity and uncertainty have not yet been sufficiently explored which may help explain some of the differences between the two ocean sink estimates. Early work points toward significant biases in the air-sea CO₂ uptake, caused by the sparsely observed Southern Hemisphere. Therefore, the Surface Ocean CO₂ Mapping intercomparison project (SOCOM) has launched its second phase to better evaluate biases (the accuracy) and uncertainties (the precision) of these upscaling techniques. Here we will present the first results from 4 dedicated experiments that were designed to identify biases, linked to data sparsity, changing observing networks and data, and uncertainties linked to mapping methods, input data and the parameterization of gas transfer of CO₂ across the air-sea interface. These experiments include an extensive geospatial uncertainty analysis assessing precision, and three subsampling experiments based on 1. the GCB hindcast simulations (representing true climate variability), 2. on large ensemble simulations (representing multiple climate states) and 3. on idealized carbon uptake without climate variations. These experiments will shed light on the role of data sparsity and network design on the accuracy and precision of the resulting air-sea CO₂ flux and will thus pave the way for implementing new observing strategies to support the value chain and better inform policy decisions.