Search for: All records

Abstract Few observational platforms are able to sustain direct measurements of all the key variables needed in the bulk calculation of air‐sea carbon dioxide (CO₂) exchange, a capability newly established for some Uncrewed Surface Vehicles (USVs). Western boundary currents are particularly challenging observational regions due to strong variability and dangerous sea states but are also known hot spots for CO₂uptake, making air‐sea exchange quantification in this region both difficult and important. Here, we present new observations collected by Saildrone USVs in the Gulf Stream during the winters of 2019 and 2022. We compared Saildrone data across co‐located vehicles and against the Pioneer Array moorings to validate the data quality. We explored how CO₂flux estimates differ when all variables needed to calculate fluxes from the bulk formulas are simultaneously measured on the same platform, relative to the situation where in situ observations must be combined with publicly‐available data products. We systematically replaced variables in the bulk formula with those often used for local and regional flux estimates. The analysis revealed that when using the ERA‐5 reanalysis wind speed in place of in situ observations, the ocean uptake of CO₂is underestimated by 8%; this underestimate grows to 9% if the NOAA Marine Boundary Layer atmospheric CO₂product and ERA‐5 significant wave height are also used in place of in situ observations. Overall our findings point to the importance of collecting contemporaneous observations of wind speed and oceanpCO₂to reduce biases in estimates of regional CO₂flux, especially during high wind events. 

Abstract A simplifying assumption in many studies of ocean carbon uptake is that the atmosphere is well‐mixed, such that zonal variations in its carbon dioxide (CO₂) content can be neglected in the calculation of air‐sea fluxes. Here, we examine this assumption at various scales to quantify the errors it introduces. For global annual averages, we find that positive and negative errors effectively cancel, so the use of atmospheric zonal‐average CO₂introduces reassuringly small errors in fluxes. However, for millions of square kilometers of the North Pacific and Atlantic that are downwind of the highly industrialized northern hemisphere continents, these biases average to over 6% of the annual ocean uptake and can cause errors of up to 30% on a given day. This work highlights the need to use a high quality, spatially‐resolved atmospheric CO₂product for process studies and for accurate long‐term average maps of ocean carbon uptake.