Submesoscale variations in sea surface temperature and salinity - a comparison between satellite data and in situ measurements from a mobile platform

Abstract: Sea surface temperature and salinity are important variables for understanding many ocean and climate processes. They can be measured both by in situ sensors and satellite remote sensing, each method with its own advantages and disadvantages. Satellite data provides unmatched spatial and temporal coverage on a global scale, but is limited by its spatial resolution which cannot always resolve variation in the submesoscale range of single digit kilometers or less. For this study, a small, free-standing sensor package was constructed. It was attached to a kayak that was paddled in two areas off the Swedish west coast that would otherwise be hard to access with other platforms. The routes included both transects of satellite sea surface temperature pixels, and alongshore routes to reach very shallow areas (<0.1m depth). The collected in situ data showed a minimum 0.23°C along transect temperature range in the sea surface temperature pixels. The pixel closest to shore showed a greater intra-pixel variability with an average of a 3.27°C along transect temperature range, and variances in the in situ data that were on the order of a hundred times greater than in the pixels further from shore. The alongshore routes paddled in very shallow areas showed temperature variations on similar intra-pixel spatial scales. No full transects of a satellite sea surface salinity pixel was performed, but the one partial transect and alongshore routes paddled showed submesoscale variability in sea surface salinity, as well as increased variability with freshwater influx. The sensor package developed shows that this type of in situ observations can be performed with free standing sensor packages, made with easily accessible parts, and confirms the result of other studies showing variability in sea surface temperature and salinity on scales smaller than what can be captured by satellite data.