Satellite observed salinity distributions at high latitudes in the Northern Hemisphere: A comparison of four products

Abstract

Global surface ocean salinity measurements have been available since the launch of SMOS in 2009 and coverage was further enhanced with the launch of Aquarius in 2011. In the polar regions where spatial and temporal changes in sea surface salinity (SSS) are deemed important, the data have not been as robustly validated because of the paucity of in situ measurements. This study presents a comparison of four SSS products in the ice‐free Arctic region, three using Aquarius data and one using SMOS data. The accuracy of each product is assessed through comparative analysis with ship and other in situ measurements. Results indicate RMS errors ranging between 0.33 and 0.89 psu. Overall, the four products show generally good consistency in spatial distribution with the Atlantic side being more saline than the Pacific side. A good agreement between the ship and satellite measurements was also observed in the low salinity regions in the Arctic Ocean, where SSS in situ measurements are usually sparse, at the end of summer melt seasons. Some discrepancies including biases of about 1 psu between the products in spatial and temporal distribution are observed. These are due in part to differences in retrieval techniques, geophysical filtering, and sea ice and land masks. The monthly SSS retrievals in the Arctic from 2011 to 2015 showed variations (within ∼1 psu) consistent with effects of sea ice seasonal cycles. This study indicates that spaceborne observations capture the seasonality and interannual variability of SSS in the Arctic with reasonably good accuracy.

Sea Surface Salinity Distribution in the Southern Ocean as Observed From Space

“…The northward transport of freshwater associated with the increasing extent of the sea ice cover has been observed to cause significant changes in salinity distribution in the Southern Ocean (Haumann et al, 2016). In this study, we investigate large-scale variations of SSS in the Southern Ocean using a technique similar to that used in a recent study of SSS in the Northern Hemisphere (Garcia-Eidell et al, 2017). As in the Northern Hemisphere study, we use data from the European Space Agency's Soil Moisture and Ocean Salinity (SMOS) as produced by the Barcelona Expert Center (BEC; Gabarro et al, 2016) and three products from Aquarius/Satelite de Aplicaciones Cientificas (SAC-D), which is a joint venture of the National Aeronautics and Space Administration's (NASA) and Argentina's Comision Nacional de Actividades Espaciales (CONAE).…”

“…Retrieving SSS in the polar regions is especially challenging because of significantly lower sensitivity of changes in brightness temperature to changes in SSS at relatively low sea surface temperature (SST; Lagerloef et al, 2015; Figure 1 of Garcia-Eidell et al, 2017). Validation studies in the Southern Ocean were made to complement the validation of the standard SSS products that have been performed primarily in warm waters (Banks et al, 2012;Ebuchi & Abe, 2014;Tang et al, 2015) and in the Arctic region (Garcia-Eidell et al, 2017). In particular, comparative studies of satellite SSS were performed with available quality-controlled Polarstern thermosalinograph (TSG) measurements and aggregated in situ measurements referred to as Coriolis Ocean database for ReAnalysis (CORA) data (Cabanes et al, 2013;Szekely et al, 2015).…”

“…Large-scale biases in the polar region have indeed been observed (Melnichenko et al, 2014), but some studies indicate that the biases were in part due to land and sea ice contaminations (Brucker et al, 2014;Kohler et al, 2014). For completeness, the technique and quality enhancements used for deriving SSS data as discussed in Garcia-Eidell et al (2017) are summarized. As in Garcia-Eidell et al (2017) and Dinnat and Brucker (2017), a sea ice mask based on ice concentration data derived from Aquarius and other passive microwave sensors is used, and the importance of using an accurate sea ice mask and a concurrent scatterometer is described.…”

“…For completeness, the technique and quality enhancements used for deriving SSS data as discussed in Garcia-Eidell et al (2017) are summarized. As in Garcia-Eidell et al (2017) and Dinnat and Brucker (2017), a sea ice mask based on ice concentration data derived from Aquarius and other passive microwave sensors is used, and the importance of using an accurate sea ice mask and a concurrent scatterometer is described. Interannual and seasonal variations of SSS were also analyzed using the four products.…”

Assessing Coastal SMAP Surface Salinity Accuracy and Its Application to Monitoring Gulf of Maine Circulation Dynamics

“…Thus, we expect that the monthly SSS anomalies used for dynamical analyses in this paper have sufficient accuracy and applicability to extend beyond the tropics to mid-latitude oceans and nearer to the coast (Boutin et al [13]; Lee [14]). This expectation is aided by the relatively large,~1 psu, spatial and temporal salinity changes that are often observed across the NW Atlantic shelf, which are gradients that are known to be detectable by remote sensing salinity (e.g., Kubryakov et al [15]; Garcia-Eidell et al [16]). …”