American Meteorological Society, Journal of Atmospheric and Oceanic Technology, 4(35), p. 911-926, 2018
DOI: 10.1175/jtech-d-17-0086.1
Full text: Download
AbstractIn this study a set of 157 collocated XBT (DB/T7 type) and CTD stations distributed across three different regions of the Southern Ocean is explored using the manufacturer’s fall-rate equation (FRE), which is a classic correction method, and new correction methods to investigate how the regional environment characteristics may impact a probe’s descent and the corresponding depth estimates. Regional coefficients were estimated for all three basins and for the Southern Ocean as a whole. The manufacturer’s FRE proved to perform better in high latitudes than in the rest of the World Ocean, overestimating the true depth by only 2%. The overall depth bias was positive, further supporting the hypothesis of a regional dependence of the XBT fall rate on water temperature, which leads to a general overestimation of ocean heat content in the upper layer (~4.79 × 109 J or ~10%). The pure thermal bias was found to be mostly negative, which is likely to be related to temperature errors. However, the Southern Ocean region is notoriously undersampled when compared to the rest of the World Ocean as well, as it is associated with strong spatial and temporal variability, thus raising the overall uncertainty on that estimate. Moreover, although the manufacturer’s FRE has a satisfying performance in the Southern Ocean, the current community’s recommended correction method still leads to improved temperature values in those waters. Finally, more studies are needed in order to fully understand the XBT regional bias and its implications for climate studies in the region.