Revisiting the data record of High Resolution Sea Surface Temperature drifting buoys: possible implications for climate monitoring

<p>Between 2012 and 2015, the EIG EUMETNET Operational Service for surface marine observations (E-SURFMAR) and European partners funded and deployed drifting buoys equipped to measure High Resolution Sea Surface Temperature. The HRSST referred to three major improvements as compared to standard practices at the time. First, the positioning accuracy was increased, thanks to GPS instead of Argos. Second, the temperature was transmitted and reported at higher resolution, 0.01 K instead of 0.1 K. Third, each buoy used an individually calibrated temperature probe, instead of one picked from a batch calibration. Note the first two improvements (but not the third) are now standard on most drifting buoys. In addition, some of these buoys were also fitted with two digital probes, one in the buoy hull around 17 cm depth, and another located around 45 cm depth (as part of a conductivity and temperature sensor). This configuration, with two accurate sensors, offered near-optimal horizontal and temporal collocation. The data records from the two sensors, located at different depths, were revisited with the help of reanalysis products as a proxy for the sea-state. The differences between the two sensors can be reduced from 0.21 K RMS to 0.06 K RMS after retaining only the situations that likely correspond to well-mixed SST. Accounting for individual buoy biases, the differences are further reduced to 0.02 K RMS. This would tend to suggest a strong relationship of the relevance of SST measurement with depth and sea-state, and the importance of these factors when using in situ SST data for climate monitoring. Another result concerns potentially the temporal stability of buoy SST data. The HRSST buoys were first calibrated individually in certified laboratories. Three buoys were recovered and the temperature sensors were brought back to a certified laboratory in 2016. Comparing the initial laboratory results with those of 2016, a temporal drift of about 0.01K/year is found for each of the three buoys. These results from so few buoys cannot be claimed to be significant. However, if confirmed, these results would have applications on the use of in situ SST buoy data as a reference, without any temporal drift correction, for the calibration and validation of satellite data and to monitor the climate.