Drivers of trends in the Bering Strait throughflow 2023

Please contact Cecilia Peralta-Ferriz (ferriz@uw.edu) for use of any of this material

Abstract
The Pacific oceanic input to the Arctic via the Bering Strait (important for western Arctic ice retreat, water properties, and nutrient supply) has been increasing for three decades. Using satellite Ocean Bottom Pressure (OBP) and Dynamic Ocean Topography (DOT) data, we show that long-term trends in mooring data for a well-sampled sub-period (2003-2014) relate to summer OBP and DOT drop in the Arctic's East Siberian Sea (ESS), in turn caused by stronger westward ESS winds, and increased fall westward winds in the Bering Sea. OBP/DOT differences imply strong (0.17psu/yr) ESS salinization, likely caused by hitherto unappreciated increased Pacific inflow to that region. We find ESS OBP trends are (erroneously) reversed in older data versions, and estimate that ESS salinization may significantly mediate Bering Strait flow increase. These facts may explain why models assimilating older OBP data, or with erroneous Bering Strait salinities, fail to simulate observed Bering Strait flow increase.

Plain Language Abstract
Direct year-round, in-water measurements show the Pacific oceanic flow to the Arctic (which is only via the narrow Bering Strait) has been increasing for three decades. This flow is important to the region as it triggers seasonal ice retreat in the Pacific side of the Arctic, the "western Arctic", and strongly influences the temperature, salinity, and nutrient content of western Arctic waters. Using satellite datasets that measure the height of the sea surface, and how it changes, we find we can explain the 2003-2014 observed flow increase by increased westward winds in the Arctic in summer and in the Bering Sea in fall. These satellite datasets also imply parts of the coastal western Arctic have become much saltier over this period, likely due to previously unsuspected increased Pacific input to the region. This increased salinity acts to slow the increasing flow to the Arctic. We find older versions of some of these satellite datasets are inconsistent with the observed flow increase. Most model simulations of the Arctic do not currently capture the observed flow increase. We suggest this may be due to fitting these models to older (erroneous) satellite datasets, or having poor representation of the salinity of Pacific waters.

Figures
For details, see paper

Figure 1

Figure 1. Trends in key Bering Strait parameters. (a) For the extent of the mooring time-series (1990-2021), monthly (gray) and annual mean (black dots) Bering Strait northward velocity (VVEL), split into pressure-head (PH term) and local wind forced (W term) terms, showing significant trends for the entire time-series and our focus period (2003-2014) (ns = not significant). (b) Seasonal trends (with uncertainties, and marked by large dot if fit significant) for these same parameters. (c) Map with the sub-regions of interest. (d-j) Maps of 2003-2014 year-round trends in Ocean Bottom Pressure (OBP, GRACE JPL Mascon Release 6 version 2); Dynamic Ocean Topography (DOT); the average of these (AVG, see section 3); NCEP Sea Level Pressure (SLP), eastward (Uwind) and westward (Vwind) surface winds; and SSM/I sea-ice concentration. White areas in (c-f and j) indicate no significant trend above the 90% confidence level. Significant trends (above 90% confidence level) in (g-i) are highlighted in thick color lines. Black (j) or gray stippled areas around the North Pole and south of 60oN in (e) and (f) indicate locations with no available DOT data. IBCAO bathymetry isobaths (Jakobsson et al., 2008) are shown every 500m from 500m to 4500m (gray lines), with land shaded in gray.

Figure 2

Figure 2. Seasonal trends in key Bering Strait parameters (2003-2014). Row (a-f) variables as per Figure 1 (d-j), column indicating season (i.e., column 1 = JFM (January, February, March); column 2 = AMJ (April, May, June); column 3 = JAS (July, August, September); and column 4 = OND (October, November, December). Trends are plotted only when significant above the 90% confidence level

Figure 3

Figure 3. Drivers of Bering Strait and East Siberian Sea (ESS) change. (a) Relationship between summer (July, August, September) AVG and PH term, showing significant (above 95% confidence level) coefficients of i) correlation and ii) regression (i.e., B of 'PH term = A + BxAVG'), and iii) response (i.e., regression coefficient multiplied by trend). Significant trends in AVG (from Figure 2c, column 3) shown also as dots on middle panel. (b) and (c) as per (a) but for surface ocean stress and PH term for (b) summer and (c) fall (October, November, December). Dots in middle panels show significant trends in ocean surface stress (not shown in Figure 2). (d) 2003-2014 annual mean trend in water column average salinity as derived from difference of OBP and DOT. (e) 2003-2014 averaged surface geostrophic velocity inferred from DOT; and (f) 2003-2014 annual mean trend in surface geostrophic velocity derived from DOT.

Copyright: Polar Science Center, University of Washington, 2023

We gratefully acknowledge financial support for this work from the National Science Foundation (NSF) and NASA.

Back to Bering Strait Homepage
Back to High Latitude Dynamics Homepage

Arctic and sub-Arctic mechanisms explaining observed increasing northward flow through the Bering Strait and why models may be getting it wrong