A comparison of remote-sensing SST and in situ seawater temperature in near-shore habitats in the western Mediterra nean Sea

14 páginas, 5 tablas, 6 figuras.

Remote sensing of sea surface temperature (SST) is widely used in climate science because it provides a quasi-synoptic coverage of the ocean. However, the use of these data for near-shore habitats is hindered by the proximity of the coast, therefore further investigation is needed. We compared remote-sensing SST from the MODIS sensor (aboard the Aqua satellite) to near-shore seawater temperature (ST) recorded in situ with data loggers at 5 locations in the western Mediterranean Sea. In situ ST data were collected at 5 m depth over a ~6 yr period and at depths below 5 m at 3 of the locations. We evaluated the suitability of MODIS to represent the temperature at shallow subtidal depths relative to different modes of variability. MODIS reproduced seasonal variability with high correlations (r > 0.98) and biases (0.59 ± 0.03°C) only slightly higher than the accuracy of the loggers (0.50°C). MODIS also captured interannual variability with no systematic biases. When evaluated for intra-seasonal temperature variability, MODIS showed limited biases (up to 0.79°C) with a tendency to overestimate the variability (between 4 and 64%) in both cold and warm seasons. Finally, MODIS over-/underestimated only the most extreme unseasonably cold/warm events (by 1.51 and −0.79°C, respectively). The observed limited differences between the 2 methods can be explained by the particular hydro dynamics of the area and by methodological constraints. Overall, MODIS SST data proved to be a reliable proxy for near-shore ST in the western Mediterranean Sea, and are thus considered suitable for studies requiring temperature reconstruction in shallow near-shore environments.

The study was funded in part by the Spanish Government through CSI-Coral Project (CGL2013- 43106-R). This is a contribution from the Marine Biogeochemistry and Global Change research group funded by the Catalan Government (2014SGR1029).

Peer reviewed