Estimation of Surface Freshwater Fluxes in the Arctic Ocean Using Satellite-Derived Salinity
- 185 Downloads
The temporal and spatial distribution of sea surface salinity (SSS) is not well known in the Arctic region, and only recent satellite-derived measurements and models have allowed for potential enhancement, though their accuracy remains uncertain. We use NASA’s Soil Moisture Active Passive (SMAP) and the European Space Agency’s Soil Moisture Ocean Salinity (SMOS) to investigate the variability of SSS in the Arctic from 2015 to 2017, as well as to calculate surface advective freshwater fluxes. These data sets are compared with Argo and European Center for Medium-Range Weather Forecasts Ocean Reanalysis version 4 (ORAS) to assess the ability of satellites in detecting freshwater fluxes. Salinity and surface freshwater fluxes are estimated for the Bering Strait and Barents Sea Opening (BSO). In this study, we have compared the Jet Propulsion Laboratory’s (JPL) SMAP salinity to the remote sensing system’s (RSS) SMAP salinity product, as well as the The Centre Aval de Traitement des donnéees SMOS salinity product. There is disagreement between the reanalysis product and satellites on the mean and variability of surface freshwater fluxes in the BSO; however, the meridional fluxes of the satellites and reanalysis product were significantly correlated within the Bering Strait. This shows the capability of using satellites to measure surface freshwater fluxes in this pathway. However, the discrepancies between satellite-derived SSS and fluxes in other regions of the Arctic Ocean emphasizes the need to increase in situ monitoring to help validate satellites in the higher latitudes.
KeywordsArctic Ocean Salinity Freshwater fluxes
The Arctic Ocean plays a very important role in the Earth’s freshwater cycle, essential to both the oceanic and atmospheric components in the Northern Hemisphere. The Arctic Ocean acts as a gateway connecting the North Pacific waters to the North Atlantic and is a key region in the deep-water formation of the meridional overturning circulation. High salinity water from the Atlantic cools as it enters the Arctic, causing overturning circulation. Some studies suggest an addition of freshwater into the Arctic could have an effect on this overturning (e.g., [16, 28]). Monitoring freshwater fluxes in the Arctic is therefore crucial to see the effects of freshwater input on circulation and climate. However, Arctic Ocean processes are difficult to study due to perennial ice and the remoteness of the region. Freshwater fluxes and content can be affected by circulation (e.g., ) and salinity, which are in turn impacted by precipitation, evaporation, rivers, and ice . Changes in circulation can impact the strength of the freshwater fluxes as well as where the freshwater is stored or released. The seasonal release of brine and freshwater from ice melting and formation drives changes in near-ice salinity .
Comparisons between satellite-derived SSS, reanalysis products, and in-situ observations will help assess the capability of using satellites to study surface freshwater processes in the Arctic region. SSS in the polar regions is difficult to monitor from satellites due to the cold-water reducing the sensitivity of L-band instruments to SSS [18, 19], high winds and strong waves affecting surface roughness and the accuracy of SSS [9, 10], and the lack of in-situ measurements for validation [14, 35, 36]. Nevertheless, NASA’s Jet Propulsion Laboratory’s Soil Moisture Active Passive (JPL SMAP) salinity data version 4 utilizes a new correction strategy in the Arctic region to account for the large uncertainties and errors near land and sea ice. It was found to have a root mean square difference less than ~ 1 psu and a correlation coefficient of ~ 0.82 psu with in-situ data north of 50° N . The remote sensing system’s SMAP (RSS SMAP) salinity data version 3 product also provides SSS measurements in the Arctic Ocean; however, more data is removed using their algorithm due to ice bias and land contamination, as shown by the extent of the product into the Arctic and along coasts (Fig. 3). This removes bias associated with the SSS measurements near ice or land; however, important measurements along the sea ice edge are removed. The LOCEAN CATDS SMOS version 3 product also improved high latitudinal biases in SSS as well as land-sea biases close to coasts. With these uncertainties in mind, we examine the Arctic region using a variety of remote sensing datasets, JPL SMAP, RSS SMAP, and SMOS salinity, as well as surface salinity from the ECMWF’s Ocean Reanalysis version 4 (ORAS).
2 Materials and Methods
2.1 Observational Data
SMAP was launched on January 31, 2015, with data availability beginning in April 2015. JPL SMAP version 4 level 3 monthly product (DOI: https://doi.org/10.5067/SMP40-3TMCS) [13, 17] is used to analyze SSS and derive surface freshwater fluxes within the Arctic. The SMAP product has a 60 km feature resolution and provides SSS at 0.25° × 0.25° gridded spatial resolution. We also use RSS SMAP version 3 level 3 monthly product (DOI: https://doi.org/10.5067/SMP3A-3SPCS) [23, 24, 29]. The feature resolution of this product is 70 km and it is provided in 0.25° × 0.25° gridded spatial resolution. We also use version 3 of debiased SMOS level 3 18-day product generated by LOCEAN CATDS [7, 8]. This product is available from January 2010 through December 2017.
In order to calculate the surface freshwater fluxes, meridional and zonal geostrophic currents from Copernicus Marine Environment Monitoring Service (CMEMS) sea surface height (SSH) level 4 products are used for the components of velocity. The product identifier for the SSH from Copernicus is SEALEVEL_GLO_PHY_L4_REP_OBSERVATIONS_008_047 (reprocessed product). These near real-time and reprocessed products allow for quality control checks and cross-calibration processes to remove residual orbit error. This dataset runs from January 1993 to present in 0.25° × 0.25° resolution.
We also use the Global Ocean Argo gridded data set (version 2018), generated by China Argo Real-time Data Center using the Barnes objective analysis , in order to validate the satellite-derived SSS and reanalysis SSS. It covers the global ocean from 79.5° S to 79.5° N from 2004 to 2017 in 1° × 1° spatial resolution. It has 58 vertical levels; however, we will be using the first layer at 0 dbar, or sea surface. When comparing the satellite products to Argo, we regridded the satellites to 1° × 1° spatial resolution.
2.2 Reanalysis Product
To compare to satellite-derived freshwater fluxes and Argo salinity, we use ECMWF’s ORAS (version 4) [4, 25]. The product is in 1° × 1° horizontal resolution from 1959 to present and uses the Nucleus for European Modeling of the Ocean v3.0 ocean model with direct surface forcing from ERA40 and ERA-Interim as well as observational data from Olv2 SST, EN3 in-situ, and AVISO SLA.
2.3 Surface Advective Freshwater Fluxes
U and V are the zonal and meridional geostrophic current components respectively (m s−1), Sfw is the standardized (unitless) freshwater anomaly value from Eq. (1), and londist and latdist are the horizontal expanse of the grid cell (m). This is a similar process to [26, 27]; however, we are only calculating surface fluxes. The relationships between the fluxes are compared through Pearson’s correlation coefficient. The correlations are tested with an alpha of 0.05, resulting in values of 95% confidence level.
2.4 Arctic Pathways
Bering Strait: 60° N–68° N and 160° W–170° E
BSO: 71° N–77° N and 17° E–29° E
3 Results and Discussion
3.1 Comparison of SSS Products
The greater similarities between Argo and ORAS may be due in part to sampling depth. The reanalysis product and Argo provide data at approximately 5 m, whereas satellite-derived salinity is measured in the top millimeters of the water column. The reanalysis product is also likely more similar to Argo because it incorporates in-situ observations in the creation of their datasets.
3.2 Salinity Variability
The variations between the SMAP products are due to the screening for sea ice and land contamination. However, the differences between the SMAP and SMOS products also come from differences in the satellite’s measurements. Sea ice can have a large impact on satellite sensors. More areas were removed from the RSS SMAP product due to sea ice contamination compared to the JPL SMAP. While this removes measurements close to the ice edge, which is an area of importance for freshwater flux measurements, it allows for less bias or error. This is why the Bering Strait and BSO were focused on in this study. Compared with other gateways in the Arctic, like the Davis Strait or Fram Strait, the BSO and Bering Strait have less error due to lower sea ice concentrations, higher water temperatures, and a greater distance from the land.
3.3 Surface Advective Freshwater Fluxes
Zonal and meridional horizontal advective freshwater flux estimates (m2 s−1) from April 2015–December 2017 in the Bering Strait and BSO for JPL SMAP, RSS SMAP, SMOS, and ORAS
Bering (m2 s−1)
BSO (m2 s−1)
These results indicate that these satellites may be capable of measuring the variability of surface freshwater fluxes in the Bering Strait. The BSO, due to its higher latitude and proximity to many land masses, does not show the same capability of satellites in measuring surface freshwater fluxes. The Bering Strait has a much larger amplitude than BSO (Figs. 8 and 9) which is able to be captured by the satellites and ORAS. This large variability is due to changes in local winds (e.g., ) and substantial interannual salinity variability .
Variability in the surface freshwater fluxes are driven by multiple different factors, including SSS and geostrophic currents. These surface advective freshwater fluxes are estimated as the freshwater anomaly compared to SSS of 34.8 psu. Larger freshwater fluxes indicate SSS values lower than this reference. The long-term trends of surface freshwater fluxes are not calculated as we were calculating fluxes during the SMAP time period and investigating the ability of satellites in measuring surface freshwater fluxes. However, a longer SSS trend will prove the utmost importance in monitoring the Arctic climate. The freshwater fluxes from all these datasets are calculated with geostrophic currents from altimetry based on SSH. While this is not an overall measure of the currents, there is a lack of full coverage currents in the Arctic region. This is a very difficult region to get accurate, year-round measurements due to the sea ice and harsh winter conditions. Error associated with geostrophic currents may also affect the freshwater flux calculations.
Satellite-derived SSS can capture overall patterns, like a higher SSS in the Atlantic, a low SSS in the Pacific, and a lower SSS in the Arctic. The satellites and ORAS are similar to Argo SSS in the subarctic seas. However, these satellites have more difficulty accounting for surface freshwater fluxes in smaller regions close to land and sea, like the Arctic gateways. The satellites can detect fluxes in similar magnitude and variability to ORAS in the Bering Strait; however, they have more difficulty in the BSO.
There is large variability, both from seasonal variability and product bias, in the SSS and resulting horizontal advective freshwater fluxes in the Arctic. The lack of consistency between products in these regions emphasizes the need to improve satellite-derived measurements and uncertainties in the Arctic. This can be done by increasing in-situ observations on consistent, broad scales in order to validate the satellite data. As sea ice declines, impacting freshwater distribution in the Arctic, it is important to continue and improve in-situ and satellite monitoring of this area because we do not yet fully know the broader impacts from climate change. While this study focuses on the surface region, it is also important to study potential changes with depth, which is why the accuracy of models and reanalysis products is crucial as well.
This work provides a way to study horizontal advective freshwater fluxes using satellite-derived SSS in the Arctic region. As far as we are aware, there has not been work on surface advective freshwater fluxes derived from remote sensing SSS in the Arctic region due to previous data gaps and high latitude satellite uncertainties. The JPL SMAP product provides SSS measurements closer to the coasts and sea ice than previous sensors, which allows for better coverage of processes, but possesses large uncertainties in these regions. RSS SMAP and SMOS therefore have less uncertainty in the high latitudes due to sea ice contamination but miss critical processes along the sea ice in the Arctic. Satellite-derived salinity may also prove to better account for surface freshwater than ocean or ocean-atmosphere coupled models and can be used to improve future model simulations by assimilating these satellite-derived salinity products.
JPL SMAP data is produced by Jet Propulsion Laboratory, and version 4.0 is obtained from NASA JPL PO.DAAC Drive (https://podaac-tools.jpl.nasa.gov/drive/files/allData/smap/retired/L3/JPL/V4/monthly) (doi: https://doi.org/10.5067/SMP40-3TMCS). RSS SMAP data is produced by Remote Sensing Systems and obtained from NASA JPL PO.DAAC Drive (https://podaac-tools.jpl.nasa.gov/drive/files/SalinityDensity/smap/L3/RSS/V3/monthly/SCI/70KM) (DOI: https://doi.org/10.5067/SMP3A-3SPCS). The L3_DEBIAS_LOCEAN_v3 Sea Surface Salinity maps have been produced by LOCEAN/IPSL (UMR CNRS/UPMC/IRD/MNHN) laboratory and ACRI-st company that participate to the Ocean Salinity Expertise Center (CECOS) of Centre Aval de Traitement des Donnees SMOS (CATDS). This product is distributed by the Ocean Salinity Expertise Center (CECOS) of the CNES-IFREMER Centre Aval de Traitement des Donnees SMOS (CATDS), at IFREMER, Plouzane (France) (http://www.catds.fr/Products/Available-products-from-CEC-OS/CEC-Locean-L3-Debiased-v3). AVISO sea surface height is obtained from Copernicus (http://marine.copernicus.eu/services-portfolio/access-to-products/?option=com_csw&view=details&product_id=SEALEVEL_GLO_PHY_L4_REP_OBSERVATIONS_008_047). ORAS4 is produced by ECMWF (ftp://ftp-icdc.cen.uni-hamburg.de/EASYInit/ORA-S4/). We are thankful for the helpful comments of the editor and two anonymous reviewers, which improved the quality of this paper.
- 2.Aagaard K, Weingartner TJ, Danielson SL, Woodgate RA, Johnson GC, Whitledge TE (2006) Some controls on flow and salinity in Bering Strait. J Geophys Res Lett 33(19). https://doi.org/10.1029/2006GL026612
- 7.Boutin J, Jean-Luc V, and Dmitry K (2018a) SMOS SSS L3 maps generated by CATDS CEC LOCEAN. Debias V3.0. SEANOE. https://doi.org/10.17882/52804#57467
- 11.Coachman LK, Aagaard K (1981) Re-evaluation of water transports in the vicinity of Bering Strait. In: Hood DW, Calder JA (eds) The eastern Bering Sea shelf: oceanography and resources, vol. 1. National Oceanic and Atmospheric Administration, Washington., pp 95–110Google Scholar
- 17.JPL Climate Oceans and Solid Earth group (2018) JPL SMAP Level 3 CAP sea surface salinity standard mapped image monthly V4.0 Validated Dataset. Ver. 4.0. PO.DAAC, CA, USA. Dataset accessed [2018-08-20] at https://doi.org/10.5067/SMP40-3TMCS
- 21.Lu S, Li H, Xu J, Liu Z, Wu X, Sun C, and Cao M (2018) Manual of Global Ocean Argo gridded data set (BOA_Argo) (Version 2018), 13 pp.Google Scholar
- 23.Meissner T, and Wentz FJ (2016) Remote sensing systems SMAP Ocean surface salinities, version 2.0 validated release. Remote sensing systems, Santa RosaGoogle Scholar
- 24.Meissner T, and Wentz FJ (2018) Remote sensing Sytems SMAP Ocean surface salinities, version 3.0 validated release. Remote sensing systems, Santa RosaGoogle Scholar
- 25.Mogensen K, Balmaseda MA, Weaver A (2012) The NEMOVAR Ocean data assimilation system as implemented in the ECMWF Ocean analysis for System4. ECMWF, 668, https://www.ecmwf.int/node/11174
- 27.Nyadjro ES, Subrahmanyam B, Shriver JF (2011) Seasonal variability of salt transport during the Indian Ocean monsoons. J. Geophys. Res.: Oceans 116(C8). https://doi.org/10.1029/2011JC006993
- 29.Remote Sensing Systems (RSS). 2018. RSS SMAP Level 3 sea surface salinity standard mapped image 8-day running mean V3.0 70km validated dataset. Ver. 3.0. PO.DAAC, CA, USA. Dataset accessed [2018-11-16] at https://doi.org/10.5067/SMP3A-3SPCS
- 33.Schmitt RW, Bogden PS, Dorman CE (1989) Evaporation minus precipitation and density fluxes for the North Atlantic. J Phys Oceanogr 19:1208–1221. https://doi.org/10.1175/1520-0485(1989)019<1208:EMPADF>2.0.CO;2 CrossRefGoogle Scholar
- 34.Serreze MC, Barrett AP, Slater AG, Woodgate RA, Aagaard K, Lammers RB, Steele M et al (2006) The large-scale freshwater cycle of the Arctic. J. Geophys. Res.: Oceans 111(C11) https://doi.org/10.1029/2005JC003424
- 38.Widell K, Fer I, Haugan PM (2006) Salt release from warming sea ice. Geophys Research Lett 33(12). https://doi.org/10.1029/2006GL026262
- 39.Woodgate RA, Aagaard K (2005) Revising the Bering Strait freshwater flux into the Arctic Ocean. Geophys Res Lett 32(2). https://doi.org/10.1029/2004GL021747