Using remote sensing and in situ datasets to distinguish coastal from ocean sea surface temperature forcing over the last 18 years.

Arctic river discharge into surrounding seas has consequential effects on the local-to-synoptic sea surface temperature, including the ocean's temperature memory for seasonal to interannual timescales (Li et al., 2022). Greater than 1/10 of global land surface drains into the Arctic and, as temperatures continue to warm, the volume of river discharge into the Arctic has increased in the winter by 60% and in the spring by 33% during the last 30 years (Feng et al., 2021; Shiklomanov et al., 2021). Characterizing the impact of river flux into the Arctic Ocean will be increasingly important for understanding its impact on ocean circulation and coastal fisheries and other stakeholders. This study uses 18 years (2003-2018) of sea surface temperature (SST) in the upper Bering Sea to characterize the effects of Yukon River discharge (YRD) on the surface temperature at the Yukon River Mouth (165.9° W, 62.4° N), the Lower Bering Sea (55.2° N, 172.4°W), and the Bering Strait (64.2° N, 169.2° W) over the area from 165° W to 180° W and 52° N to 70° N. This study compares the Remote Sensing System's Group for High Resolution Sea Surface Temperature 0.09 degree level 4 (microwave and infrared) dataset with the Arctic Great Rivers Observatory's Yukon River volumetric daily point river discharge dataset over the same time period.

A monthly mean of each dataset was taken and the seasonal cycle removed to reveal the anomaly of the resulting datasets. We examine SST first with respect to successive processing of: monthly mean, monthly mean with seasonal cycle removed, and the principal component of the first empirical orthogonal function (EOF1) of the processed data. The principal component 1 (PC1) of the monthly mean data (including the seasonal cycle) represents greater than 99% of the total variance, whereas the monthly mean of the data with seasonal cycle removed (anomaly) represents nearly 59% of the variance in the area studied. Comparison of the correlations of the anomaly and daily data with YRD suggests that the remaining SST variance is represented by the YRD in the Bering Sea. Future aspects of this study will make use of a numerical model provided by the Estimating the Circulation and Climate of the Ocean (ECCO) project to establish causality between SST variability and Yukon River discharge.