1. Introduction
Surface air temperatures in the Arctic have been increasing 4 times faster than the global average since 2000 (Chylek et al. 2022). This is due to Arctic amplification, the dynamics of which are controlled by the poleward transport of heat by the atmosphere and the ocean (Previdi et al. 2021), and strong local feedbacks which enhance the warming signal, including the ice–albedo feedback, the lapse-rate feedback, and Planck feedback (Stuecker et al. 2018). In regional analyses of the Bering Sea, no significant trend in sea ice extent (SIE) was observed between 1979 and 2017 (e.g., Peng and Meier 2018; Bliss et al. 2019). Since then, however, strong negative anomalies have emerged; four consecutive winters with below-average ice extent occurred between 2014 and 2017, followed in 2018 and 2019 by record-low SIE in January–March when SIE was 53% and 40% below the 1981–2010 average, respectively (Ballinger and Overland 2022). Anomalous periods of southerly winds in these winters brought warm and humid air and inhibited the more typical (Sullivan et al. 2014) southward advection of ice (Stabeno and Bell 2019). Using the Community Earth System Model Large Ensemble, Thoman et al. (2020) investigated the decadal probability of Bering Sea SIE in winter exceeding the 2018 record low and found that the probability reaches 50% by the decade 2041–50. The duration of the sea ice season in the Bering Sea north of 60°N is also projected to decrease at a rate between −0.5 and −1 day yr−1 between 2015 and 2044 (Wang et al. 2018).
Sea ice advance and retreat in the Bering Sea are strongly influenced by wind forcing (Overland and Pease 1982), and the structure of surface winds is determined by the position and strength of the Aleutian low (Overland and Hiester 1980). The winter mean low pressure extends over the Aleutian Islands, and its central longitude may vary from 160°W to 160°E (Overland et al. 1999). When the Aleutian low is shifted westward, the cyclonic wind pattern creates southerly winds over the Bering Sea. In this regime, storm tracks that typically lead from west to east can acquire a northward component that allows them to enter the Bering Sea (Anderson and Gyakum 1989). Between 2017 and 2021, the average position of the Aleutian low was anomalously far west, leading to southerly winds and low SIE in the Bering Sea (Ballinger and Overland 2022). On average, the southerly winds that inhibit ice advance are projected to increase by the end of the century (Hermann et al. 2019), although changes in the positioning of storm tracks under climate change remain under active research (Shaw et al. 2016).
Considering the coastal communities of the Bering Sea and the economic importance of the region, the changes underway are of societal importance. Recent changes in SIE have had a negative impact on the Bering Sea ecosystem and resulted in the cancellation of the snow crab fishing season in 2022 (Alaska Department of Fish and Game 2022a) and the red king crab season in both 2021 and 2022 (Alaska Department of Fish and Game 2022b). Without the dampening effect of the sea ice, communities in the region are experiencing winter storm–surge events (Walsh 2018), more hazardous wave conditions (Rolph et al. 2018), coastal erosion, flooding, and destruction of crucial infrastructure (Slats et al. 2019), although to date no long-term trend in storminess has been found (Thoman and Walsh 2019).
No recent studies have looked specifically at decadal-scale trends in surface wind and waves in the Bering Sea, leaving a gap in knowledge, which is especially concerning given the recent observed changes in ice extent and its impacts. Moreover, because it lies just outside the Arctic Ocean, it often receives limited attention and neither global ocean studies nor regional Arctic studies describe surface ocean conditions in the Bering Sea in much detail. The region was included in a recent global study of long-term change between 1985 and 2018 (Young and Ribal 2019), but no trends were indicated for the Bering Sea.
This study aims to address the lack of attention on decadal-scale trends and variability in significant wave height (SWH) and surface winds in the Bering Sea, in the context of recent sea ice loss. We do so by updating the time series of winter SIE in the Bering Sea and reporting on multimission satellite altimeter observations of SWH. We compare SWH with wind speeds from reanalysis data. Section 2 will introduce the study region and period and describe the data and methods used. In section 3, we establish the seasonal cycles of winds, waves, and sea ice in the Bering Sea. We then discuss changes in the distributions of the wind speed and SWH over the last 21 winters and look at the differences in SWH between the eastern Bering Sea shelf, where the presence of sea ice affects surface conditions, and the Aleutian Basin, where the ocean surface is ice free year-round. In a case study, we examine surface conditions during extratropical cyclone Merbok, which occurred in September 2022. We use this study to illustrate the impact that severe storms can have in the Bering Sea in the absence of sea ice. Finally, we assess the observational capabilities required to continue monitoring the largest storms in the region using satellite altimetry techniques.
2. Study region, data, and methods
a. Study region and period
The Bering Sea is a semienclosed basin, 2 million km2 in area, bounded to the south by the Aleutian Islands and the North Pacific Ocean and to the north by the Bering Strait, which connects it to the Arctic Ocean (Fig. 1). It can be divided into a continental shelf shallower than 200 m in the northeast and the Aleutian Basin deeper than 3000 m in the southwest. The shelf can be further divided into the coastal, middle, and outer shelf (Fig. 1a), roughly following the bathymetric contours of 50, 100, and 200 m, respectively (Hood and Calder 1981). In this study, we use the Bering Sea domain as defined by the International Hydrographic Organization (IHO) (IHO and Sieger 2012) shown in Fig. 1 (purple outline). Here, we use “shelf region” to refer to the eastern Bering Sea continental shelf at depths shallower than 200 m, 53.5°N and 175°E (Fig. 1a).
Bering Sea study region. (a) Ocean bathymetry [General Bathymetric Chart of the Oceans (GEBCO) Compilation Group 2021, https://doi.org/10.5285/c6612cbe-50b3-0cff-e053-6c86abc09f8f]. Highlighted are the study region domain (purple line), the 50-, 100-, and 200-m (black) isobaths and 3000-m (yellow) isobaths, and the location of NOAA buoys 46035 and 46073. (b) Over 6.69 million satellite altimeter-derived SWH 1-Hz observations collected in winter between October 2002 and April 2023, ordered from lowest to highest SWH. The largest SWHs are shown on top so as to highlight the geospatial distribution of very high and phenomenal wave height observations > 9 m during the 21-yr period.
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
In this study, we consider two periods: an extended, 44-yr period, spanning October 1979–April 2023, during which satellite-derived ice extents and wind speeds from reanalysis data are available, and a shorter, 21-yr period, spanning October 2002–April 2023, during which observations of SWH from satellite altimetry are available (Fig. 1b). In the following, we define the winter season as running from October to April and refer to the winter season by the year which includes the month of January, i.e., the winter spanning October 1979–April 1980 will be referred to as winter 1980.
b. Data
1) SWH
Altimeters on Earth-orbiting satellites have been used for ocean remote sensing since the late 1970s, but began measuring ocean height at a level accurate enough to determine basin-scale ocean circulation in 1992 with the launch of TOPEX/Poseidon (Stammer and Wunsch 1994). Radar altimetry determines the height of a target on Earth’s surface by measuring the satellite-to-surface round-trip travel time of a transmitted radar pulse (Chelton et al. 2001). Combined with knowledge of the instrument position relative to a geodetic reference frame, the geographical position of the radar footprint on Earth’s surface can be determined. Most oceanographic altimeters operate at the Ku band since this radar frequency penetrates clouds and most rain, providing data in nearly all weather conditions (Fu et al. 1994). Sea level is sampled at points along the satellite’s ground track, and depending on the pulse repetition frequency of the instrument, approximately 1000 individual waveforms are returned to the receiver per second and averaged to produce a 1-Hz measurement (Chelton et al. 2001). SWH, defined as the peak-to-trough height of the largest third of surface ocean waves, is approximately 4 times the standard deviation of the sea surface elevation (Sverdrup and Munk 1946). These waves roughen the sea surface and alter the travel time of transmitted radar pulse such that the leading edge of the received radar waveform indicates radar energy intercepted at the wave crests and energy returned from the wave troughs is received slightly later. The time delay in the returned radar energy can thus be related to wave height (Chelton et al. 2001).
To study satellite altimeter observations of the Bering Sea, we use the level 2 along-track SWH data from the multimission Radar Altimeter Database System (RADS) (Scharroo et al. 2013). The number of satellite altimeters observing the surface ocean determines how well SWH conditions are sampled in space and time. The study period commences in 2003 since this is the first winter when observations from at least three satellites are present, following the launch of Envisat in March 2002 (Table 1). Prior to this, with only two satellites sampling the region, sparse observations may not record all occurrences of large waves making it more challenging to determine ocean surface conditions. Later, in section 3g, we investigate the ideal number of altimeters required to adequately sample storm conditions in the Bering Sea. During the study period, data from the 10 altimeters included in the RADS database (Table 1) yielded a total of 6 690 105 1-Hz SWH observations in the Bering Sea, as shown in Fig. 1b. There are >160 000 1-Hz observations available every winter, providing ∼1600 observations to robustly estimate the 1% largest SWHs, a metric we use in our analysis to investigate extremes in SWH.
Satellite radar altimeters used in this study, the temporal availability of data, and their latitudinal extent and orbit repeat period.
We follow the World Meteorological Organization Sea State Code 3700 (WMO 2019) to define surface ocean conditions in the Bering Sea. Figure 1b shows that altimeter-derived SWHs are available across the full study region, including the shelf seas and coastal regions, and that waves between 9 and 14 m tall, defined as “very high seas” (WMO 2019), are observed over the entirety of the Aleutian Basin and in some cases over the outer and middle shelf. “Phenomenal seas” (SWH ≥ 14 m) have been observed in the satellite radar altimeter record on eight occasions and occurred along the western and southern edges of the Bering Sea (Fig. 1b, dark blue dots).
There are two active, open-ocean buoys from the National Oceanic and Atmospheric Administration (NOAA) National Data Buoy Center (NDBC) (NOAA/NDBC 1971) in the Bering Sea (Fig. 1, diamonds) that also provide SWH measurements for the region. These data are available at 10-min intervals dating back to 1985 (station 46035) and 2005 (station 46073), respectively, and are used to assess the range of SWHs in the Bering Sea and to evaluate the quality of the altimeter measurements. Following the approach used in previous studies (e.g., Sepulveda et al. 2015; Queffeulou 2004; Ribal and Young 2019), comparisons are performed by taking the average of all altimetry observations within a radius of 50 km and ±30 min of the buoy observation. Comparisons between the satellite-derived SWH and the buoy measurements (Fig. 2) show that at the site of buoy 46035, collocated SWHs range from 0.4 to 10.9 m, with an average of 3.3 m, while at the site of buoy 46073, collocated SWHs range from 0.5 to 12 m, with an average of 3.1 m, indicating sea states spanning “smooth” to “very high” (WMO 2019). At the two sites, the correlation coefficients between the spatiotemporally collocated datasets are 0.99 and 0.98, respectively, and the root-mean-square errors (RMSEs) are 0.27 and 0.28 m, respectively, about a ∼0 mean bias (Fig. 2). This RMSE is comparable to that reported by Sepulveda et al. (2015) for Jason-2 (0.24 m), where a large number of collocated altimeter and buoy data were analyzed. The strong agreement between the two independent SWH datasets suggests that we can use the altimeter-derived SWH data in our analysis with a high degree of confidence.
Altimeter measurements of SWH in winter between 2003 and 2022 within a 50-km radius and ±30-min window of in situ observations at (a) NDBC buoy 46035 and (b) NDBC buoy 46073. The black line indicates the linear regression, and the gray line is the 1:1 slope. Buoy location is shown in Fig. 1.
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
2) SIE and SIC
We use sea ice concentration (SIC) data from the NOAA/U.S. National Snow and Ice Data Center (NSIDC) Climate Data Record of Passive Microwave Sea Ice Concentration 4 (Meier et al. 2021) to assess sea ice conditions in the Bering Sea. SIC is defined as the fraction of ocean area covered by sea ice in 25 km × 25 km grid cells (Parkinson 2014). We use daily and monthly mean SIC data for the winters 1980–2023. From SIC, we compute SIE as the sum of all grid cells with SIC ≥ 15% following (Parkinson and Cavalieri 2008).
3) Wind speed
We use data from the fifth major global reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ECMWF) (ERA5) (Hersbach et al. 2020) for the winters 1980–2023. Following previous studies, we analyze data at a 0.25° horizontal resolution and a 6-hourly temporal resolution (e.g., Laurila et al. 2021; Aue et al. 2022; Akperov et al. 2020). The atmospheric variables used in this study are the two horizontal components of the 10-m wind speed and the mean sea level pressure (SLP). In the following, we will use “surface wind” to refer to the 10-m wind. For area-average statistics, such as the seasonal and annual distributions, the wind speed data, which are on a latitude–longitude grid, are weighted by the cosine of latitude to account for the change in area with latitude.
4) Cyclone tracking
Extratropical cyclones are low pressure systems poleward of 30° latitude which are a major source of weather extremes (Messmer and Simmonds 2021). Their development depends on the horizontal gradient of temperature as originally described by Bjerknes and Solberg (1922) and revised by Shapiro and Keyser (1990). To investigate the storm conditions that produce the largest wave heights in the Bering Sea, we track extratropical cyclones in the region using the Center for Earth Observation Science (CEOS)/NSIDC Extratropical Cyclone Tracking (CNECT) algorithm, version 12.4 (Crawford et al. 2021). The algorithm tracks closed contours in mean SLP fields from ERA5 to track the surface low associated with the cyclones. It outputs several metrics over the lifespan of the cyclone, such as minimum pressure, radius, depth, and deepening rate. The storms of interest are defined in this study as those with a pressure minimum of ≤980 hPa and ≥21 m s−1 surface wind speeds.
3. Results
a. Bering sea ice extent
At sea ice maximum, which typically occurs in March, seasonal sea ice extends across most of the shelf region, whereas there is no significant occurrence of sea ice in the deeper Aleutian Basin (Fig. 3). As a consequence, air–sea interactions in winter over the shelf are often affected by the presence of sea ice, whereas that is not the case in the Aleutian Basin. The interaction between waves and sea ice consists mainly of wave attenuation by the ice (Horvat et al. 2020) and, conversely, ice breakup by wave action (Thomson 2022).
Decadal variability of Bering Sea winter SIE and SIC. (a) Time series of winter SIE between 1980 and 2023 with the 44-yr mean (thin solid line), 5-yr rolling average (thick gray line), and linear trend (dashed line). (b) Bering Sea winter mean SIC 1980–2023. Decadal SIC anomalies (%) relative to the 1980–2023 mean for the periods (c) 1984–93, (d) 1994–2003, (e) 2004–13, and (f) 2014–23.
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
Figure 3a shows the time series of winter SIE in the Bering Sea between 1980 and 2023. The average winter SIE over the 44-yr period (Fig. 3a, thin solid line) is 528 × 103 km2. The data indicate a small linear decline (Fig. 3a, dashed line) in sea ice extent of −3 × 103 km2 yr−1 for the period 1980–2023, which is statistically significant to the 95% confidence level. The pentad, 5-yr rolling average (Fig. 3a, thick gray line), shows that SIE in the Bering Sea has subdecadal variability. It also reveals that SIE was relatively stable about the long-term average over the first three decades, followed by a recent decline, primarily during the last decade. Record minimum SIE occurred in the winters of 2018 (214 × 103 km2) and 2019 (306 × 103 km2). Winter SIE in 2020 and 2021 remained well below the long-term average, and although SIE in winter 2022 was slightly above average, winter 2023 saw anomalously low SIE ∼77 × 103 km2 below average. Indeed, nine of the last ten winters have seen below-average SIE in the Bering Sea (Fig. 3a).
b. Seasonality
Next, we establish the general characteristics of the Bering Sea in winter and investigate seasonal variability in ocean surface conditions (Fig. 4) using satellite-derived SWH and SIE, and ERA5 wind speed. Annual mean wind speed is 8.0 m s−1, but there is a clear seasonal cycle, with windy winters and relatively calm summer months (Fig. 4a). Between October and April, monthly mean wind speed is at or above the annual mean, and from November to February, it is notably constant at ∼10 m s−1. The monthly 95th and 99th percentiles of wind speed follow very similar seasonal patterns (Fig. 4a). Gale force winds (≥17 m s−1, force 8 on the Beaufort wind scale) occur, and the 99th percentile of wind speed is ≥19 m s−1 during the months of November–February.
Monthly variations in wind speed, SWH, and SIE in the Bering Sea. (a) Annual climatology of ERA5 wind speed averaged monthly for the period October 1979–April 2023 showing the monthly mean (dark blue), 95th (yellow) percentile, and 99th (red) percentile. (b) Annual climatology of altimeter-derived SWH averaged monthly between May 2002 and April 2023 showing the monthly mean, 95th percentile, and 99th percentile of wave height. (c) Average monthly SIE between October 1979 and April 2023 (solid black line with diamonds) given as percentage extent of the Bering Sea region (%, as defined in Fig. 1). Shading indicates the envelope of monthly SIE between the 10th and 90th percentiles. Daily SIE shown for three record-breaking years: 2012 (yellow), the record winter maximum ice extent; 2018 (purple) and 2019 (blue), record low winter ice extents.
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
The seasonal cycle in wave height matches that of the wind speed, with larger waves in the winter than in summer (Fig. 4b). Annual mean SWH is 2.3 m, with a modal height of 1.3 m, indicating that the SWH distribution is skewed to the right with a tail of high extremes. Between October and April, the mean SWH is 3.0 m, twice as high as during the summer months of May–September. During winter, at least 1% of all SWH observations are ≥6.0 m, corresponding to a “high” sea state (WMO 2019), and this period is associated with gale force winds. The largest waves occur in the winter months of December–February, of which December has the largest SWHs across all metrics (Fig. 4b), in agreement with the long-term wind speed climatology (Fig. 4a). Since some altimeter-derived SWH data are assimilated in ERA5 (Hersbach et al. 2020), consistency between the datasets may be expected. If we consider, however, SWH observations from the Sentinel-3A/B and Sentinel-6A missions alone for the period 2016–23, data that were not assimilated in ERA5, we can replicate (not shown) the seasonal SWH cycle in Fig. 4b. We can therefore conclude that December and February are the stormiest months in the Bering Sea.
Sea ice freeze-up typically occurs in November, and at its peak, sea ice can extend across ∼40% of the Bering Sea (Fig. 4c), covering the majority of the surface ocean in the shelf region. This can occur any time between January and April (Stabeno and Bell 2019). Temporal variability in SIE can be large though, and studies show that ice can be advected for ∼1000 km in just 30 days (Sullivan et al. 2014). Synoptic events and weather systems can lead to large swings in ice advection and downward longwave radiation, by bringing either cold, dry air from over the continents or sea ice, or relatively warm, humid southerly air from the Pacific. Such variability is illustrated in the daily SIE for the winters of 2012, 2018, and 2019 (Fig. 4c). Northerly winds between September 2011 and May 2012 advected ice southward (Babb et al. 2013), leading to a record maximum SIE in the Bering Sea in March 2012 (Fig. 3a), when sea ice extended across ∼60% of the region. Conversely, repeated periods of southerly winds in November, February, and March in the winters of 2018 and 2019 (Stabeno and Bell 2019) led to record-low extents (Fig. 3a) when ice only extended across 20% and 30% of the Bering Sea in March 2018 and January 2019, respectively (Fig. 4c).
In summary, there is a strong distinction between summer and winter conditions at the sea surface. Taken together, our decadal-scale results show that the winter period can be defined as spanning October–April. In October and November, however, much of the Bering Sea is ice free, while at least 1% of all SWH observations correspond to high sea states (WMO 2019) and large waves can freely propagate onto most parts of the shelf. In December–February, when the waves reach their maximum height, sea ice extends across much of the northern and coastal shelf areas, dampening wave amplitudes and protecting coastal regions from the high sea states occurring in the Aleutian Basin. On average, most of the shelf is covered by sea ice for the remainder of the winter storm season until the end of April.
c. Annual distributions
To assess interannual variability and changes in ocean surface conditions in winter, we examine the annual distributions of wind speed and SWH between 2003 and 2023 (Fig. 5). The wind speed (Fig. 5a), which can be described by a Weibull distribution (Hennessey 1977; Monahan 2006; Justus et al. 1978), is not very skewed, with a mean of 9.2 m s−1 (±0.23 m s−1) and mode of 9.0 m s−1. The 99th percentile of wind speeds over the 21-yr period is 19 m s−1, roughly twice the mean. Over time, gale force wind speeds (≥17 m s−1) have become more common (Fig. 5a), as indicated by the fatter tails of the distributions for the more recent years (yellow) in comparison with the earlier years (darker colors).
Annual distributions in (a) surface wind speed in winter, area weighted, and (b) SWH, for a 21-yr period spanning winter 2002/03–2022/23. Mode (light blue dashed line), mean (dark blue dashed line), and 99th percentile (red dashed line) of all area-weighted data over 21-yr period are shown. The 21-yr winter average distribution (black dashed line) is also shown for both wind speed and SWH.
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
The SWH distributions are more skewed than the wind speed distributions (Fig. 5b). They are approximately lognormal, with a mode of 2.1 m, a mean of 3.0 m, and a 99th percentile of 7.8 m. As is the case with wind speed, there is no clear indication that modal values have changed during the 21-yr period; for example, modal SWH was 2.0 m in both 2003 and 2022. SWH extremes on the other hand have increased in recent years, indicated by changes in the tail of the distribution, especially between ∼4 and 9 m, which has become notably more elongated in recent years (Fig. 5b).
d. Decadal-scale trends
Next, we look at the time series of wind speed and SWH in the Bering Sea and we consider variability and trends in both the winter (October–April) means (Fig. 6a) and the 99th percentile of their distributions (Fig. 6b).
Trends in Bering Sea wind speed and SWH in winter (annual October–April average). (a) Mean surface wind speed (blue) from 1980 to 2023 and mean SWH (red) from 2003 to 2023. (b) The 99th percentile of surface wind speed from 1980 to 2023 (blue) and 99th percentile of SWH (red) from 2003 to 2023. The thick lines indicate the 5-yr rolling average (pentad) in wind speed (light blue) and SWH (light red). The solid lines indicate linear trends in wind speed for the period 1980–2023 (gray) and SWH for the period 2003–23 (red).
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
We observe an increase in both the mean surface wind speed and the mean SWH with time, as well as high interannual variability (Fig. 6a). Fitting a pentad to both time series shows that the largest increases have occurred in the last decade. The linear trend in mean wind speed is 0.09 m s−1 decade−1 for the period 1980–2023, and for mean SWH, it is 0.13 m decade−1 for the period 2003–23, both statistically significant to 95%. Winter mean wind speeds larger than 9.5 m s−1 have occurred five times in the 44-yr record, and four have occurred in the last 6 years (2018, 2019, 2021, and 2022). Similarly, the three winters with the highest mean SWHs during the 21-yr study period have occurred in 2018, 2021, and 2022. Interannual variability in mean wind speed and mean SWH correlates very strongly during the last 21 winters; they have a Pearson correlation coefficient of 0.90. This suggests a strong coupling between SWH as measured by radar altimeters and the surface wind forcing in ERA5.
To further quantify changes in the largest SWHs, we examine the variability and trends in the 99th percentile of the surface wind speed and SWH (Fig. 6b). The pentad emphasizes the low-frequency signal in the time series and shows that both variables have greatly increased during the study period (Fig. 6b), with a pattern quite similar to that observed for mean wind speed and SWH (Fig. 6a). Wind speed extremes in the Bering Sea increased by a linear rate of 0.19 m s−1 decade−1 during the period 1980–2023, while SWH extremes increased at a rate of 0.54 m decade−1 during the period 2003–23, with both trends statistically significant to 99%.
Interannual variability in the 99th percentile of the SWHs correlates very well with the variability in the 99th percentile of the surface wind speed from ERA5 over the last two decades (Fig. 6b), with a Pearson correlation coefficient of 0.93. The 99th percentile of wintertime wind speed exceeded 19.3 m s−1 eight times during the 44-yr record, including all of the last six winters (2018–23). Similarly, the last seven winters (2017–23) have had a higher 99th percentile of SWH than previous years in the 21-yr period. Notably, in recent years, there has also been an improvement in satellite altimetry coverage of the region following the launch of Jason-3 and Sentinel-3a in early 2016, with at least five active altimeters observing sea state in the region at all times (Table 1). As we will discuss in section 3g, this has improved our capability to capture extreme SWH events. The fact that the variability in the 99th percentile of SWH is strongly correlated with the variability in wind speeds, and the fact that the top 1% of the SWH data consist of ∼1600 1-Hz measurements per year, lends confidence that the altimeter-derived results are a reflection of changes in sea state rather than an effect of the increased satellite sampling in the region.
e. Contrasts between the Aleutian Basin and eastern Bering Sea shelf region
Since the Bering Sea is a large region, we consider the regional differences within it by discerning two subdomains with and without a seasonal sea ice cover: the shallow eastern Bering Sea shelf and the deep Aleutian Basin (Figs. 1 and 7). Figure 7a shows the winter mean SIE (2003–23) for consistency with the time period during which altimeter-derived SWH observations are available. During this period, the pentad analysis reveals rapid ice loss in the last decade (Fig. 7a, thick gray line). The linear trend in SIE decline for 2003–23 was 7.7 × 103 km2 yr−1, which coincides with the increases we found in surface wind speed and wave height extremes (as shown in Fig. 6b), although we note that the 2003–23 linear SIE trend is not a statistically significant trend. Sea ice in the Bering Sea is not geographically uniform and only occurs in the shelf region (Fig. 7b) where it affects air–sea coupling and dampens wave propagation toward the coastal zone. Our results show that the duration of the sea ice season for the period 2003–23 has decreased at a rate of between 1 and 5 days yr−1, with the biggest changes occurring in the middle shelf region near Saint Matthew Island (Fig. 7b). Here, SWH extremes are lower, but the linear increase in SWH remains positive at 0.39 m decade−1, which is statistically significant to 99% (Fig. 7c, blue line). In contrast, SWH extremes in the Aleutian Basin where sea ice is absent are higher: The average 99th percentile of SWH in the Aleutian Basin is 8.1, 1.3 m larger than that over the shelf, at 6.8 m (Fig. 7c). The linear trend in 99th percentile of SWH is more pronounced in the Aleutian Basin at 0.55 m decade−1 (Fig. 7c, red line), consistent with where the trend in surface wind speed is strongest (Fig. 7d). While Fig. 6 revealed strong correlations between the trends and variability in SWH and wind speed, Fig. 7d shows that the recent increases in wind speed are not geographically uniform across the region and peak at 1.0 m s−1 decade−1 over the central Bering Sea.
Winter sea surface trends over the Aleutian Basin and eastern Bering Sea shelf for the 21-yr period 2003–23. (a) Bering Sea winter mean SIE, with the 1980–2023 mean (solid line), 5-yr rolling average (thick gray line), and 2003–23 linear trend (dashed line). (b) Trend in the duration of the Bering Sea ice season for the period 2003–23, defined as the number of days annually when SIC is greater than 15%. (c) The 99th percentile of SWH trends in the Aleutian Basin (red) and over the shelf (blue). (d) Trend in 99th percentile of ERA5 surface wind speed for the period 2003–23, with the 200-m bathymetry contour delineating the shelf region (black).
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
Although variability in the 99th percentile of SWH can be well explained by the surface wind variations (Fig. 6b), the interannual variability of the SWH extremes in the shelf region (Fig. 7c, blue line) is modulated by the winter mean sea ice (Fig. 7a) and is negatively correlated with a Pearson correlation coefficient of −0.57. The record-low SIE in winter 2018 (Fig. 7a) occurred at the same time as the record high (7.6 m) in the 99th percentile of SWH over the shelf (Fig. 7c). There are other years, notably 2012, where the 99th percentile of SWH in the shelf region is high, regardless of a large mean SIE in the same winter. This situation can arise because the sea ice season only partially overlaps with the storm season, which means that the highest SWHs can occur before the onset of freeze-up. For example, in winter 2012, the largest SWHs were observed in November (2011), during a period when SIE was still very low (<10%; Fig. 4c).
f. September 2022 extratropical cyclone Merbok
Another strong indicator of recent changes in Bering Sea storm conditions came in September 2022. Tropical cyclone Merbok reached peak intensity in the North Pacific on 14 September before transitioning into an extratropical cyclone as it moved north and east into the Bering Sea, where it manifested as the strongest storm in 70 years (Eisner et al. 2023). The storm exhibited hurricane-like characteristics, with a pressure minimum of 937 hPa, 45 m s−1 wind gusts, and record-breaking SWHs (Eisner et al. 2023). This resulted in widespread damage of coastal infrastructure due to erosion and flooding (Schwoerer et al. 2023).
Here, we examine SWH conditions during extratropical cyclone Merbok as a representative case study to explore the impacts of storms that generate extreme wave heights when the dampening effects of sea ice are not present (since the storm occurred before the onset of freeze-up). Observations from the satellite radar altimeter suite captured sea state conditions on 16–17 September 2022, during the passage of Merbok through the region (Fig. 8). Of the 5741 SWH observations obtained, 6.5% corresponded to very high sea states (9–14 m), which extended over much of the central Bering Sea from 53.0° to 61.6°N and from 177.3°E to 166.9°W (Fig. 8). 0.3% of altimeter-derived SWH observations were associated with phenomenal seas (>14 m), and notably, these large waves were located over the shelf break at ∼56°N, 173°W and over the middle shelf at ∼59°N, 170°E (Fig. 8). These sea state conditions were exceptional for the time of year. Prior to this, very high seas (>9 m) were observed in only 0.005% of the September SWH observations between 2003 and 2023, and phenomenal seas had not been observed in September before.
Altimeter-derived SWH observations on 16 and 17 Sep 2022 during the passage of extratropical cyclone Merbok through the Bering Sea, including observations from Sentinel-3A, Sentinel-3B, Sentinel-6A Michael Freilich, SARAL, and Jason-3. Inset: Time series of buoy SWH observations over the same period at the locations indicated by the diamonds on the map.
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
The largest satellite-derived SWH recorded was 15.3 m, observed by Sentinel-6A Michael Freilich at 59.1°N, 170.5°W, over the middle shelf where the ocean floor is <100 m deep, setting a record-high SWH for the month of September, where the previous highest value was 10.9 m during the period 2002–22. NOAA buoy 46 035 recorded even larger SWHs of up to 15.8 m (Fig. 8, inset). The unique nature of these phenomenal sea state conditions is further corroborated by comparing the SWH observations during Merbok (Fig. 8) to those over the preceding 20 winters (Fig. 1b). Phenomenal seas (>14 m) have only been observed eight times in the 21-yr study period (Fig. 1b), and none were observed over the shelf. The largest waves previously observed in the shelf region were all <9 m. Notably, this region, south of Saint Matthew Island, coincides with the location where the largest changes in the length of the sea ice season are occurring, at a rate of −5 day yr−1 (Fig. 7b).
g. Satellite coverage during extratropical cyclones
The storms associated with the largest waves are almost always extratropical cyclones. Of the very high (9–14 m) and phenomenal (≥14 m) seas that were observed by radar altimeters, 88% (355/405 days) occurred on days when a cyclone was tracked in the region with a minimum SLP ≤ 980 hPa and maximum surface winds ≥ 21 m s−1. Cyclone duration is determined by the persistence of these characteristics for a minimum duration of 6 h, which is a function of the temporal resolution of the data. Using these atmospheric characteristics to identify cyclones in the ERA5 reanalysis data, we can investigate how many such cyclones occurred between 2003 and 2023 and compare this to how many were observed in the satellite altimeter data to understand the efficacy of the satellite sampling.
During the 21-yr study period, a total of 600 cyclones were recorded in winter in the ERA5 dataset. 60% of cyclones were between 6- and 24-h long, while the remaining 40% persisted for ≥24 h up to a maximum of 126 h. In Fig. 9, we show the satellite observations of SWH during these cyclones: Each point indicates the time and number of satellite observations per hour during the cyclone. When no satellite observations were obtained during a cyclone, the point is shown in cyan (and recorded as a “cyclone missed”; Fig. 9). In all other cases, the color indicates the maximum SWH observed by any altimeter during the cyclone (Fig. 9).
Number of altimeter-derived SWH observations over the Bering Sea within 3 h of cyclones with a minimum pressure ≤ 980 hPa and surface winds ≥ 21 m s−1. The number of 1-Hz observations is normalized by the duration of the cyclone in hours. The color of the dots indicates the maximum SWH observed during the cyclone passage through the Bering Sea. Extratropical cyclone Merbok, which occurred in September 2002, is indicated. When no observations were made during the passage of a cyclone, dots are represented with a cyan color and tagged as a “missed” cyclone. Below the data, the blue bars show the per mission time span of satellite data available in RADS.
Citation: Journal of Climate 37, 23; 10.1175/JCLI-D-23-0485.1
The number of satellites active at any time and their inclination and repeat period (Table 1) determine the sampling frequency of SWH in the Bering Sea. Between 2002 and 2016, there were three to four altimeters observing the region, but following the launch of the Jason-3 and Sentinel-3a missions in 2016, observations from five to six altimeters were available. In the seven winters since 2017, observations of sea surface height were obtained by the satellite suite during all cyclones. In the 14 prior winters, there were 24 cyclones during which no satellite altimeter observations were obtained. With data from only two or three active altimeters available in RADS during many of those winters, longer time gaps between subsequent overpasses allowed some storms to be missed. Notably, all 24 cyclones were short-lived events of ≤6-h duration.
Between 2003 and 2016, the number of satellite observations during cyclones averaged 42 1-Hz measurements per hour, increasing to an average of 93 1-Hz measurements per hour for cyclones that occurred from 2017 to 2023. The maximum SWH observed during these cyclones has increased from an average of 8.2 m during the first 14 winters to an average of 9.3 m during the last seven winters of the study period, indicated by the increased prevalence of very high and phenomenal seas in Fig. 9. In other words, SWHs associated with very high and phenomenal sea states were observed during 35% of cyclones between 2003 and 2016, while that number rose to 56% between 2017 and 2023.
4. Discussion
Our analysis shows that sea state in the Bering Sea has been exceptionally rough in the seven winters between 2017 and 2023, when the 99th percentile of SWH was higher than in any other year during the study period. At the same time, a decline in SIE has emerged. Our results suggest that SIE conditions in the Bering Sea have subdecadal variability, in agreement with Hendricks et al. (2023), who also found a strong relationship to land surface temperatures in southwest Alaska. We have found that winter SIE averaged for the period 2014–23 was 23% below the long-term mean, and a widespread SIC anomaly of –10% was observed across the entire shelf region during that period. Studies by Cox et al. (2019) and Ballinger and Overland (2022) both show that low SIE (in 2015–18 and 2017–21, respectively) is related to a westward shift of the Aleutian low and strengthening of the polar jet stream aloft over the Bering Sea. Repeating the analysis of SLP conducted by Ballinger and Overland (2022, Fig. 2) and extending it to include two additional winter seasons, we investigate how this relationship has developed (Fig. S1 in the online supplemental material). An eastern position of the Aleutian low in 2012–16 (Fig. S1a) promotes southwestward advection and growth of sea ice, which is indicated by an increased SICanom in the western shelf region (Fig. S1d). The westward shift of the Aleutian low in 2017–21 gave rise to warm southerly winds over the Bering Sea that resulted in extreme negative SICanom (Fig. S1e), corroborating the relationship described by Ballinger and Overland (2022). Meanwhile, the frequency of gale force winds and above increased over the Aleutian Basin (Fig. S1h). Extending the analysis to include the winters of 2022 and 2023 reveals a slight eastward shift of the Aleutian low (Fig. S1c), with reduced negative SICanom over the outer shelf and small positive SICanom over the Alaskan coastal shelf and the northern middle shelf (Fig. S1f). Whether this atmospheric regime persists into the future and what consequences this will have for the location and seasonality of storm tracks (Mioduszewski et al. 2018) that originate in the North Pacific are important questions that warrant further investigation.
North of the Bering Strait, in the Chukchi and Beaufort seas, Francis et al. (2011) observed an increase in SWH by 0.2 m decade−1 between 1993 and 2010 and Thomson et al. (2016) show that interannual variability was strongly related to the amount of open water. In summers between 1996 and 2015, mean SWH in the Chukchi Sea increased at >0.1 m decade−1 and the 99th percentile of SWH increased faster, at rates of between 0.1 and 0.5 m decade−1 (Liu et al. 2016). The linear trends we observe in winter in the Bering Sea of 0.13 m decade−1 in mean SWH and 0.54 m decade−1 in the 99th percentile of SWH between 2003 and 2023 are of similar magnitude to those found in summer in the Chukchi and Beaufort seas.
The passage of extratropical cyclone Merbok through the Bering Sea in September 2022 exemplified the consequences of very high and phenomenal sea states in the Bering Sea, and their impacts on coastal communities, when storms hit during ice-free periods. To monitor these storms, which our results show are becoming more prevalent (Figs. 6, 7, and 9), there are only two NOAA buoys in the Bering Sea and two tide gauges on the 1000-km-long western Alaska coastline between Nome and Hooper Bay (NOAA/NDBC 1971). This small number of in situ observations over such a large area means that although atmospheric conditions can be well predicted, forecasts of sea state and storm surge are inadequate in this region (Schwoerer et al. 2023). To improve forecasts for coastal communities vulnerable to storm surge as well as commercial entities operating in the region, additional observations are needed to constrain models (Roach et al. 2019; Cooper et al. 2022). Our results indicate that the current coverage of at least five satellite radar altimeters is needed to maintain observations of sea state during all intense cyclones. To better understand wave–ice interactions in the marginal ice zone during individual storm events in winter when sea ice has formed, advances in both high-resolution altimetry, such as with ICESat-2 (Horvat et al. 2020; Brouwer et al. 2022), and regional modeling (Roach et al. 2019; Boutin et al. 2020, 2022; Cooper et al. 2022) will also be important.
5. Conclusions
Despite the economic importance of the Bering Sea region and myriad societal impacts of changing climate conditions there, recent regional surface ocean conditions are not well documented. Our analyses of satellite-derived SIE and SWH in combination with reanalysis of wind speeds reveal that the months of December and February are the stormiest times of the year with the largest SWHs and wind speeds, averaging 3.4 m and 10 m s−1, respectively. During winter, at least 1% of all SWH observations indicate high sea states (WMO 2019), but by December sea ice extends across ∼20% of the Bering Sea, dampening wave amplitudes and protecting the coasts, particularly in the northern and coastal shelf areas, while the very high and “phenomenal” sea states occur in the deeper Aleutian Basin. Between 2003 and 2023, gale force wind speeds have become more common as SWHs have increased. Our results also show that the duration of the sea ice season has decreased at a rate of 1–5 days yr−1 between 2003 and 2023, which is quicker than the rate of −1 day yr−1 previously observed (Bliss et al. 2019) and faster than the rate projected for the future (−0.5 to −1.0 day yr−1 between 2015 and 2044; Wang et al. 2018). Pentad analysis of the 44-yr SIE time series in the Bering Sea has revealed a rapid loss of sea ice in the last decade. Lower-than-average SIE has been recorded in 9 of the last 10 winters (2014–23).
Because the sea ice season overlaps with the storm season, low SIE in winter and delayed freeze onset have increased air–sea coupling. Concurrent with this, mean and 99th percentile of wind speed and SWH in the Bering Sea have steadily increased. Both interannual variability and longer-term (i.e., decadal) variations in wind speed and SWH are large and are highly correlated (r = 0.9) suggesting a strong coupling between SWH as measured by radar altimeters and the surface wind forcing in ERA5. SWH extremes in the Bering Sea region increased by a linear rate of 0.54 m decade−1 and at a rate of 0.39 m decade−1 over the continental shelf during the period 2003–23. During the record-low SIE in 2018, the 99th percentile of SWH over the shelf peaked at 7.6 m. At the same time, extreme wind speeds have increased, although these increases are not geographically uniform across the region and peak at 1.0 m s−1 decade−1 over the central Bering Sea. This has led to a linear trend in the 99th percentile of SWH of 0.6 m decade−1 in the deep Aleutian Basin. Our results show that the increase in SWHs observed during the satellite radar altimetry era reflects decadal-scale changes in sea state rather than artifacts due to increased satellite sampling.
In September 2022, extratropical cyclone Merbok caused sea state conditions over the shelf that were unprecedented for both the season and the location in the context of the climatology. Satellite radar altimeter observations revealed record-high SWHs of 15.3 m over the shelf where the ocean depth is <100 m, an event that had only been observed on three other occasions during the 21-yr study period (2002–23) in the deeper Aleutian Basin. Merbok led to widespread coastal inundation, flooding, and damage to property and infrastructure (Schwoerer et al. 2023). This demonstrated the threat that storms which generate phenomenal sea state can pose, not only to the coastal environment of the Bering Sea but also to the economic health of the region when the protective sea ice is absent. Should a shortening of the sea ice season and reduction in SIE, as reported here, continue into the future, we expect an increased risk of damaging waves to impact activities in the Bering Sea shelf region. Societal impacts are further exacerbated since many coastal communities are isolated and without connection to national road networks, necessitating evacuation by sea or air in the event of severe flooding or coastal erosion.
Because of these threats and sparse in situ observations in the Bering Sea, satellite observations play an important role in monitoring variability and trends in sea state. The current suite of altimeters provides sufficient observations to capture the sea state during all impactful cyclones of at least 6 h. We found that 24 short-lived cyclones were missed due to sparse sampling when fewer than five altimeters were operational, whereas all storms were captured with five or six altimeter systems observing the area. So as to better understand linkages between ice loss and extreme wave conditions, it is crucial that the satellite altimeter record be maintained at its current coverage at least, to provide observations that can be used to better predict risks to Alaskan coastal communities and safety at sea.
Acknowledgments.
We thank Editor Dr. Stephen Yeager and three anonymous reviewers for their feedback and comments which helped improve the manuscript significantly. This study was supported by NOAA Grant NA19NES4320002 [Cooperative Institute for Satellite Earth System Studies (CISESS)] at the University of Maryland and the Ocean Surface Topography Science Team. The CEOS/NSIDC Extratropical Cyclone Tracking (CNECT) algorithm was made available by Alex Crawford (https://github.com/alexcrawford0927/cyclonetracking). We thank R. Thoman for his helpful discussion in personal communication. The scientific results and conclusions, as well as any views or opinions expressed herein, are those of the authors and do not necessarily reflect those of NOAA or the Department of Commerce.
Data availability statement.
Satellite radar altimetry SWH data from RADS (Scharroo et al. 2013) version 4.4.0 as described at https://github.com/remkos/rads are available at http://rads.tudelft.nl/rads/rads.shtml. NDBC buoy SWH data (NOAA/National Data Buoy Center 1971) are available at https://www.ndbc.noaa.gov/obs.shtml. Sea ice data from the NSIDC Climate Data Record of Passive Microwave Sea Ice Concentration 4 (Meier et al. 2021) are available at https://doi.org/10.7265/efmz-2t65. Horizontal surface wind magnitudes and SLP data from ERA5 (Hersbach et al. 2020) are available at https://doi.org/10.24381/cds.adbb2d47. The CEOS/NSIDC Extratropical Cyclone Tracking (CNECT) algorithm is available at https://github.com/alexcrawford0927/cyclonetracking.
REFERENCES
Akperov, M., V. A. Semenov, I. I. Mokhov, W. Dorn, and A. Rinke, 2020: Impact of Atlantic water inflow on winter cyclone activity in the Barents Sea: Insights from coupled regional climate model simulations. Environ. Res. Lett., 15, 024009, https://doi.org/10.1088/1748-9326/ab6399.
Alaska Department of Fish and Game, 2022a: 2022/23 Bering Sea snow crab season closed. Accessed 17 April 2023, https://www.adfg.alaska.gov/static/applications/dcfnewsrelease/1441272349.pdf.
Alaska Department of Fish and Game, 2022b: 2022/23 Bristol Bay red king crab season closed. Accessed 2 May 2023, https://www.adfg.alaska.gov/static/applications/dcfnewsrelease/1441166194.pdf.
Anderson, J. R., and J. R. Gyakum, 1989: A diagnostic study of Pacific basin circulation regimes as determined from extratropical cyclone tracks. Mon. Wea. Rev., 117, 2672–2686, https://doi.org/10.1175/1520-0493(1989)117<2672:ADSOPB>2.0.CO;2.
Aue, L., T. Vihma, P. Uotila, and A. Rinke, 2022: New insights into cyclone impacts on sea ice in the Atlantic sector of the Arctic Ocean in winter. Geophys. Res. Lett., 49, e2022GL100051, https://doi.org/10.1029/2022GL100051.
Babb, D. G., R. J. Galley, M. G. Asplin, J. V. Lukovich, and D. G. Barber, 2013: Multiyear sea ice export through the Bering Strait during winter 2011–2012. J. Geophys. Res. Oceans, 118, 5489–5503, https://doi.org/10.1002/jgrc.20383.
Ballinger, T. J., and J. E. Overland, 2022: The Alaskan Arctic regime shift since 2017: A harbinger of years to come? Polar Sci., 32, 100841, https://doi.org/10.1016/j.polar.2022.100841.
Bjerknes, J., and H. Solberg, 1922: Life Cycle of Cyclones and the Polar Front Theory of Atmospheric Circulation. Geophysisks Publikationer, 18 pp.
Bliss, A. C., M. Steele, G. Peng, W. Meier, and S. Dickinson, 2019: Regional variability of Arctic sea ice seasonal change climate indicators from a passive microwave climate data record. Environ. Res. Lett., 14, 045003, https://doi.org/10.1088/1748-9326/aafb84.
Boutin, G., C. Lique, F. Ardhuin, C. Rousset, C. Talandier, M. Accensi, and F. Girard-Ardhuin, 2020: Towards a coupled model to investigate wave–sea ice interactions in the Arctic marginal ice zone. Cryosphere, 14, 709–735, https://doi.org/10.5194/tc-14-709-2020.
Boutin, G., T. Williams, C. Horvat, and L. Brodeau, 2022: Modelling the Arctic wave-affected marginal ice zone: A comparison with ICESat-2 observations. Philos. Trans. Roy. Soc., A380, 20210262, https://doi.org/10.1098/rsta.2021.0262.
Brouwer, J., and Coauthors, 2022: Altimetric observation of wave attenuation through the Antarctic marginal ice zone using ICESat-2. Cryosphere, 16, 2325–2353, https://doi.org/10.5194/tc-16-2325-2022.
Chelton, D. B., J. C. Ries, B. J. Haines, L.-L. Fu, and P. S. Callahan, 2001: Satellite altimetry. Satellite Altimetry and Earth Sciences: A Handbook of Techniques and Applications, L.-L. Fu and A. Cazenave, Eds., International Geophysics Series, Vol. 69, Academic Press, 1–131.
Chylek, P., C. Folland, J. D. Klett, M. Wang, N. Hengartner, G. Lesins, and M. K. Dubey, 2022: Annual mean Arctic amplification 1970–2020: Observed and simulated by CMIP6 climate models. Geophys. Res. Lett., 49, e2022GL099371, https://doi.org/10.1029/2022GL099371.
Cooper, V. T., L. A. Roach, J. Thomson, S. D. Brenner, M. M. Smith, M. H. Meylan, and C. M. Bitz, 2022: Wind waves in sea ice of the western Arctic and a global coupled wave-ice model. Philos. Trans. Roy. Soc., A380, 20210258, https://doi.org/10.1098/rsta.2021.0258.
Cox, C. J., R. S. Stone, D. C. Douglas, D. M. Stanitski, and M. R. Gallagher, 2019: The Aleutian Low-Beaufort Sea Anticyclone: A climate index correlated with the timing of springtime melt in the Pacific Arctic cryosphere. Geophys. Res. Lett., 46, 7464–7473, https://doi.org/10.1029/2019GL083306.
Crawford, A. D., E. A. P. Schreiber, N. Sommer, M. C. Serreze, J. C. Stroeve, and D. G. Barber, 2021: Sensitivity of Northern Hemisphere cyclone detection and tracking results to fine spatial and temporal resolution using ERA5. Mon. Wea. Rev., 149, 2581–2598, https://doi.org/10.1175/MWR-D-20-0417.1.
Eisner, L., E. Lamagie, and E. Siddon, 2023: The Bering Sea: Current status and recent trends. PICES Press, 31, 96–100.
Francis, O. P., G. G. Panteleev, and D. E. Atkinson, 2011: Ocean wave conditions in the Chukchi Sea from satellite and in situ observations. Geophys. Res. Lett., 38, L24610, https://doi.org/10.1029/2011GL049839.
Fu, L.-L., E. J. Christensen, C. A. Yamarone Jr., M. Lefebvre, Y. Ménard, M. Dorrer, and P. Escudier, 1994: TOPEX/POSEIDON mission overview. J. Geophys. Res., 99, 24 369–24 381, https://doi.org/10.1029/94JC01761.
Hennessey, J. P., Jr., 1977: Some aspects of wind power statistics. J. Appl. Meteor., 16, 119–128, https://doi.org/10.1175/1520-0450(1977)016<0119:SAOWPS>2.0.CO;2.
Hermann, A. J., G. A. Gibson, W. Cheng, I. Ortiz, K. Aydin, M. Wang, A. B. Hollowed, and K. K. Holsman, 2019: Projected biophysical conditions of the Bering Sea to 2100 under multiple emission scenarios. ICES J. Mar. Sci., 76, 1280–1304, https://doi.org/10.1093/icesjms/fsz043; Corrigendum, 76, 1937, https://doi.org/10.1093/icesjms/fsz111.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146A, 1999–2049, https://doi.org/10.1002/qj.3803.
Hood, D. W., and J. A. Calder, 1981: The Eastern Bering Sea Shelf: Oceanography and Resources. US Department of Commerce, National Oceanic and Atmospheric Administration, Office of Marine Pollution Assessment, 625 pp.
Horvat, C., E. Blanchard-Wrigglesworth, and A. Petty, 2020: Observing waves in sea ice with ICESat-2. Geophys. Res. Lett., 47, e2020GL087629, https://doi.org/10.1029/2020GL087629.
IHO, and R. Sieger, 2012: Limits of oceans and seas in digitized, machine readable form, accessed 22 January 2022, https://doi.org/10.1594/PANGAEA.777975.
Justus, C. G., W. R. Hargraves, A. Mikhail, and D. Graber, 1978: Methods for estimating wind speed frequency distributions. J. Appl. Meteor., 17, 350–353, https://doi.org/10.1175/1520-0450(1978)017<0350:MFEWSF>2.0.CO;2.
Laurila, T. K., V. A. Sinclair, and H. Gregow, 2021: Climatology, variability, and trends in near-surface wind speeds over the North Atlantic and Europe during 1979–2018 based on ERA5. Int. J. Climatol., 41, 2253–2278, https://doi.org/10.1002/joc.6957.
Liu, Q., A. V. Babanin, S. Zieger, I. R. Young, and C. Guan, 2016: Wind and wave climate in the Arctic Ocean as observed by altimeters. J. Climate, 29, 7957–7975, https://doi.org/10.1175/JCLI-D-16-0219.1.
Meier, W. N., F. Fetterer, A. K. Windnagel, and J. S. Stewart, 2021: NOAA/NSIDC climate data record of passive microwave sea ice concentration, version 4. Accessed 1 July 2023, https://doi.org/10.7265/efmz-2t65.
Messmer, M., and I. Simmonds, 2021: Global analysis of cyclone-induced compound precipitation and wind extreme events. Wea. Climate Extremes, 32, 100324, https://doi.org/10.1016/j.wace.2021.100324.
Mioduszewski, J., S. Vavrus, and M. Wang, 2018: Diminishing Arctic Sea ice promotes stronger surface winds. J. Climate, 31, 8101–8119, https://doi.org/10.1175/JCLI-D-18-0109.1.
Monahan, A. H., 2006: The probability distribution of sea surface wind speeds. Part I: Theory and SeaWinds observations. J. Climate, 19, 497–520, https://doi.org/10.1175/JCLI3640.1.
NOAA/NDBC, 1971: Meteorological and oceanographic data collected from the National Data Buoy Center Coastal-Marine Automated Network (C-MAN) and moored (weather) buoys, https://www.ncei.noaa.gov/archive/accession/NDBC-CMANWx.
Overland, J. E., and T. R. Hiester, 1980: Development of a synoptic climatology for the Northeast Gulf of Alaska. J. Appl. Meteor. Climatol., 19, 1–14, https://doi.org/10.1175/1520-0450(1980)019<0001:DOASCF>2.0.CO;2.
Overland, J. E., and C. H. Pease, 1982: Cyclone climatology of the Bering Sea and its relation to sea ice extent. Mon. Wea. Rev., 110, 5–13, https://doi.org/10.1175/1520-0493(1982)110<0005:CCOTBS>2.0.CO;2.
Overland, J. E., J. M. Adams, and N. A. Bond, 1999: Decadal variability of the Aleutian low and its relation to high-latitude circulation. J. Climate, 12, 1542–1548, https://doi.org/10.1175/1520-0442(1999)012<1542:DVOTAL>2.0.CO;2.
Parkinson, C. L., 2014: Spatially mapped reductions in the length of the Arctic sea ice season. Geophys. Res. Lett., 41, 4316–4322, https://doi.org/10.1002/2014GL060434.
Parkinson, C. L., and D. J. Cavalieri, 2008: Arctic sea ice variability and trends, 1979–2006. J. Geophys. Res., 113, C07003, https://doi.org/10.1029/2007JC004558.
Peng, G., and W. Meier, 2018: Temporal and regional variability of Arctic sea-ice coverage from satellite data. Ann. Glaciol., 59, 191–200, https://doi.org/10.1017/aog.2017.32.
Previdi, M., K. L. Smith, and L. M. Polvani, 2021: Arctic amplification of climate change: A review of underlying mechanisms. Environ. Res. Lett., 16, 093003, https://doi.org/10.1088/1748-9326/ac1c29.
Queffeulou, P., 2004: Long-term validation of wave height measurements from altimeters. Mar. Geod., 27, 495–510, https://doi.org/10.1080/01490410490883478.
Ribal, A., and I. R. Young, 2019: 33 years of globally calibrated wave height and wind speed data based on altimeter observations. Sci. Data, 6, 77, https://doi.org/10.1038/s41597-019-0083-9.
Roach, L. A., C. M. Bitz, C. Horvat, and S. M. Dean, 2019: Advances in modeling interactions between sea ice and ocean surface waves. J. Adv. Model. Earth Syst., 11, 4167–4181, https://doi.org/10.1029/2019MS001836.
Rolph, R. J., A. R. Mahoney, J. Walsh, and P. A. Loring, 2018: Impacts of a lengthening open water season on Alaskan coastal communities: Deriving locally relevant indices from large-scale datasets and community observations. Cryosphere, 12, 1779–1790, https://doi.org/10.5194/tc-12-1779-2018.
Scharroo, R., E. W. Leuliette, J. L. Lillibridge, D. Byrne, M. C. Naeije, and G. T. Mitchum, 2013: RADS: Consistent multi-mission products. Proc. of the Symp. on 20 Years of Progress in Radar Altimetry 2012, Venice, Italy, European Space Agency Special Publication, ESA SP-710, 4 pp., https://ui.adsabs.harvard.edu/abs/2013ESASP.710E..69S.
Schwoerer, T., K. Berry, S. M. Rasmus, J. P. Joe, N. W. Naneng, and T. M. Ravens, 2023: Climate policy must account for community-specific socio-economic, health, and biophysical conditions—Evidence from coastal Alaska. Reg. Environ. Change, 23, 90, https://doi.org/10.1007/s10113-023-02080-9.
Sepulveda, H. H., P. Queffeulou, and F. Ardhuin, 2015: Assessment of SARAL/AltiKa wave height measurements relative to Buoy, Jason-2, and Cryosat-2 data. Mar. Geod., 38, 449–465, https://doi.org/10.1080/01490419.2014.1000470.
Shapiro, M. A., and D. Keyser, 1990: Fronts, jet streams and the tropopause. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. O. Holopainen, Eds., Amer. Meteor. Soc., 167–191.
Shaw, T. A., and Coauthors, 2016: Storm track processes and the opposing influences of climate change. Nat. Geosci., 9, 656–664, https://doi.org/10.1038/ngeo2783.
Slats, R., and Coauthors, 2019: Voices from the front lines of a Changing Bering Sea. Arctic Report Card 2019, NOAA Tech. Rep., 88–94, https://arctic.noaa.gov/wp-content/uploads/2023/04/ArcticReportCard_full_report2019.pdf.
Stabeno, P. J., and S. W. Bell, 2019: Extreme conditions in the Bering Sea (2017–2018): Record-breaking low sea-ice extent. Geophys. Res. Lett., 46, 8952–8959, https://doi.org/10.1029/2019GL083816.
Stammer, D., and C. Wunsch, 1994: Preliminary assessment of the accuracy and precision of TOPEX/POSEIDON altimeter data with respect to the large-scale ocean circulation. J. Geophys. Res., 99, 24 584–24 604, https://doi.org/10.1029/94JC00919.
Stuecker, M. F., and Coauthors, 2018: Polar amplification dominated by local forcing and feedbacks. Nat. Climate Change, 8, 1076–1081, https://doi.org/10.1038/s41558-018-0339-y.
Sullivan, M. E., N. B. Kachel, C. W. Mordy, S. A. Salo, and P. J. Stabeno, 2014: Sea ice and water column structure on the eastern Bering Sea shelf. Deep-Sea Res. II, 109, 39–56, https://doi.org/10.1016/j.dsr2.2014.05.009.
Sverdrup, H. U., and W. Munk, 1946: Empirical and theoretical relations between wind, sea, and swell. Eos, Trans. Amer. Geophys. Union, 27, 823–827, https://doi.org/10.1029/TR027i006p00823.
Thoman, R., and J. E. Walsh, 2019: Alaska’s changing environment: Documenting Alaska’s physical and biological changes through observations. International Arctic Research Center, University of Alaska Fairbanks, 16 pp.
Thoman, R., and Coauthors, 2020: The record low Bering Sea ice extent in 2018: Context, impacts, and an assessment of the role of anthropogenic climate change. Bull. Amer. Meteor. Soc., 101, S53–S58, https://doi.org/10.1175/BAMS-D-19-0175.1.
Thomson, J., 2022: Wave propagation in the marginal ice zone: Connections and feedback mechanisms within the air–ice–ocean system. Philos. Trans. Roy. Soc., A380, 20210251, https://doi.org/10.1098/rsta.2021.0251.
Thomson, J., and Coauthors, 2016: Emerging trends in the sea state of the Beaufort and Chukchi seas. Ocean Model., 105, 1–12, https://doi.org/10.1016/j.ocemod.2016.02.009.
Walsh, J., 2018: Sea ice in the Alaska region: The remarkable winter of 2017–2018. Alaska Climate Dispatch. A State-wide Seasonal Summary and Outlook, 18 pp., https://uaf-accap.org/wp-content/uploads/2019/10/climate_dispatch_May_2018_FINAL.pdf.
Wang, M., Q. Yang, J. E. Overland, and P. Stabeno, 2018: Sea-ice cover timing in the Pacific Arctic: The present and projections to mid-century by selected CMIP5 models. Deep-Sea Res. II, 152, 22–34, https://doi.org/10.1016/j.dsr2.2017.11.017.
WMO, 2019: Manual on Codes - International Codes, Volume I.1, Annex II to the WMO Technical Regulations: Part A-Alphanumeric Codes. WMO Tech. Note 306, 454 pp., https://doi.org/10.25607/OBP-109.
Young, I. R., and A. Ribal, 2019: Multiplatform evaluation of global trends in wind speed and wave height. Science, 364, 548–552, https://doi.org/10.1126/science.aav9527.