Sea Surface Temperature-Salinity Covariability and Its Scale-Dependent characteristics

Sea surface temperature and salinity (SST/SSS) are critical to understanding the ocean's role in the climate system. Because seawater density is controlled by temperature and salinity, characterizing the relationship between these two variables is particularly important for oceanography, climate science and other related fields. By analyzing observational data, we provide the first global picture of covariability between SST and SSS and associated physical processes at interannual timescales. A notable feature of such SST-SSS covariability is its distinct spatial-scale dependence; at small spatial scales (<1,000 km), their covariability is attributed to fluctuations in ocean currents and the associated heat and salt transport, whereas large-scale covariability result from combinations of changes in local atmospheric conditions and large-scale ocean circulations. It is found that the current generation of climate models underestimate such distinct SST-SSS covariations, suggesting that physical processes involving in variations of these variables are not faithfully represented as observed. Plain Language Summary Variations in sea surface temperature and salinity (SST and SSS) reflect changes in the atmosphere and ocean system, and play an important role in the climate system. As the density of seawater, which controls the ocean circulation and distributions of various biogeochemical tracers, is determined by both temperature and salinity, it is important to understand the linkage between the two variables. To address this issue, we analyzed observational data and present the global picture of covariability between SST and SSS at interannual timescales. We found that characteristics of covariability between the two variables at spatial scale smaller than 1,000 km are different from those at larger spatial scales, due to differences in the dominant mechanisms driving variations in these variables at each spatial scale. At small spatial scales, covariations between SST and SSS mostly arise from changes in ocean circulation, while both atmosphere and ocean are important for covariations at large spatial scales. Analysis of numerical climate models that emulate the atmosphere-ocean system reveals that such distinct covariations between SST and SSS are also found in these models, but their covariant intensity vary among models, implying that some parts of observed density variability are still difficult to represent in these models.