Dependence of Beaufort Sea Low Ice Condition in the Summer of 1998 on Ice Export in the Prior Winter

Yongli Zhang School of Marine Science and Technology, Tianjin University, Tianjin, China

Search for other papers by Yongli Zhang in
Current site
Google Scholar
PubMed
Close
,
Hao Wei School of Marine Science and Technology, Tianjin University, Tianjin, China

Search for other papers by Hao Wei in
Current site
Google Scholar
PubMed
Close
,
Youyu Lu Bedford Institute of Oceanography, Fisheries and Oceans Canada, Dartmouth, Nova Scotia, Canada

Search for other papers by Youyu Lu in
Current site
Google Scholar
PubMed
Close
,
Xiaofan Luo School of Marine Science and Technology, Tianjin University, Tianjin, China

Search for other papers by Xiaofan Luo in
Current site
Google Scholar
PubMed
Close
,
Xianmin Hu Bedford Institute of Oceanography, Fisheries and Oceans Canada, Dartmouth, Nova Scotia, Canada

Search for other papers by Xianmin Hu in
Current site
Google Scholar
PubMed
Close
, and
Wei Zhao School of Marine Science and Technology, Tianjin University, Tianjin, China

Search for other papers by Wei Zhao in
Current site
Google Scholar
PubMed
Close
Free access

Abstract

Four events of distinctly low summer ice coverage in the Beaufort Sea, in 1998, 2008, 2012, and 2016, have been identified using satellite-observed concentration between 1979 and 2017. Previous studies have revealed that these four minima were impacted by preconditioning of the ice cover, and specifically the 1998 event was preconditioned toward thinner ice by anomalous southeasterly winds during winter. This study further investigates the 1998 event through analyzing the solution of a coupled ocean and sea ice model. Compared with the mean condition during 1995–2015, the net ice loss in the melt season (May–September) of 1998 was not particularly high. In the preceding fall (October–December 1997), the ice conditions and processes contributing to ice changes were neither significantly different from the mean condition nor unique in the time series during 1995–2015. In the preceding winter (January–April 1998), over the southeastern part of the Beaufort Sea, the ice was 1.5 m thinner than the mean condition on average, and the increase in ice thickness due to freezing was nearly offset by the decrease due to lateral advection, which was the result of high westward ice export and limited southerly import. The dynamic process in preceding winter was also the cause of low ice in summer 2016 according to a recent study. Model analyses suggest that the 2008 event was due to the small regional ice volume at the end of summer 2007 and ice export during the preceding fall, whereas the 2012 event was caused by the excessive summer melting.

Corresponding author: Xiaofan Luo, xiaofan.luo@tju.edu.cn.

Abstract

Four events of distinctly low summer ice coverage in the Beaufort Sea, in 1998, 2008, 2012, and 2016, have been identified using satellite-observed concentration between 1979 and 2017. Previous studies have revealed that these four minima were impacted by preconditioning of the ice cover, and specifically the 1998 event was preconditioned toward thinner ice by anomalous southeasterly winds during winter. This study further investigates the 1998 event through analyzing the solution of a coupled ocean and sea ice model. Compared with the mean condition during 1995–2015, the net ice loss in the melt season (May–September) of 1998 was not particularly high. In the preceding fall (October–December 1997), the ice conditions and processes contributing to ice changes were neither significantly different from the mean condition nor unique in the time series during 1995–2015. In the preceding winter (January–April 1998), over the southeastern part of the Beaufort Sea, the ice was 1.5 m thinner than the mean condition on average, and the increase in ice thickness due to freezing was nearly offset by the decrease due to lateral advection, which was the result of high westward ice export and limited southerly import. The dynamic process in preceding winter was also the cause of low ice in summer 2016 according to a recent study. Model analyses suggest that the 2008 event was due to the small regional ice volume at the end of summer 2007 and ice export during the preceding fall, whereas the 2012 event was caused by the excessive summer melting.

Corresponding author: Xiaofan Luo, xiaofan.luo@tju.edu.cn.

1. Introduction

Since 1979, satellite observations have revealed significant variations of sea ice in the Arctic Ocean including the Beaufort Sea (BS; Rigor and Wallace 2004; Babb et al. 2016; Howell et al. 2016; Babb et al. 2019). Starting from the mid-1990s, the summer (September) ice edge in the Beaufort Sea has retreated northward rapidly (Stroeve et al. 2005; Maslanik et al. 2007; Onarheim et al. 2018), accompanied with a sharp decrease in the coverage of multiyear ice (MYI) (Kwok and Cunningham 2010; Maslanik et al. 2011; Galley et al. 2016). Significant decreases in sea ice concentration have occurred in the western Beaufort Sea (Tivy et al. 2011), and the breakup in both the western and eastern regions shifted to earlier dates in a more synchronous way (Steele et al. 2015).

Overlaid on this general trend of sea ice decrease, interannual variation is also evident. The annual sea ice extent from 1953 to the present in the Beaufort Sea reached an extremely low level in 1998 (Maslanik et al. 1999) and another one in 2008 (Perovich et al. 2011). The Beaufort Sea became ice free for 31 consecutive days in the summer of 2012 (Babb et al. 2016) and only 4 years later, it had a greater area of open water in the summer of 2016 (Babb et al. 2019). Among the identified extremely low ice years, the summer of 1998 is unique in the sense that it followed from previous decades with the presence of heavy ice on average, and is the first year with low ice under the condition of a younger and thinner ice pack. Previous studies have identified 1998 as a clear transition of dominant ice coverage in the Beaufort Sea from MYI to first-year ice (FYI) (e.g., Galley et al. 2008; Hutchings and Rigor 2012; Galley et al. 2016).

The long-term and interannual variations of sea ice coverage in summer are related to the dynamic and thermodynamic processes not only in the present summer but also in the preceding seasons. For a regular seasonal cycle, FYI develops in fall and winter and melts in summer in the southern Beaufort Sea (Galley et al. 2008, 2016), while MYI is advected into the Beaufort Sea and then exported into the Chukchi Sea (CS) by the anticyclonic Beaufort Gyre (e.g., Howell et al. 2016). This sea ice advection is generally strong during October–December, weakened from January to March, and strengthened again from April to June. The seasonal variation of ice advection is related to the variability in size and location of the Beaufort Gyre (Proshutinsky et al. 2002) and also the internal stresses within the ice cover (Steele et al. 2015). These dynamic and thermodynamic processes undergo anomalous variations above the seasonal cycle, which affect the state of the ice cover at the end of winter and thereby condition it for either enhanced or reduced summer melt. For example, a larger loss of MYI during the preceding fall and winter shall result in a smaller regional ice volume of thinner ice type in the early melt season. This will make the ice pack more vulnerable to atmospheric forcing (Petty et al. 2016; Babb et al. 2019) and cause the reduction in the ice–albedo feedback and hence the low ice condition in summer (Perovich et al. 2007).

A number of studies have been devoted to understanding the causes of the extremely low sea ice conditions in the Beaufort Sea during the years identified. The low ice condition in the summer of 1998 was related to the low value of the Arctic Oscillation index (Rigor and Wallace 2004) and the export of MYI by persistent southerly and easterly winds from November 1997 through April 1998 (Maslanik et al. 1999). From February to April in 1998, anomalously thin ice was observed in the southern Beaufort Sea, quite different from the years before and after (Melling et al. 2005). The low ice record of September 2008 was related to the ice and ocean condition in the previous summer. In the summer of 2007, the ice cover in the Beaufort Sea was lower than that in previous summers after 1998, leading to a delayed freeze-up in the following winter of 2008 (Perovich et al. 2008; Steele et al. 2010) and an early open-water onset in the summer of 2008 (Kwok and Cunningham 2010). Based on observations in the Canada Basin (CB), Timmermans (2015) revealed that the low ice cover in summer of 2007 resulted in the anomalously large solar heat input into the ocean and excessive heat storage in the near-surface temperature maximum layer (NSTM) underlying the mixed layer. The release of this stored solar heat caused reduced sea ice thickness before the summer of 2008. The 2012 ice minimum was attributed to the ice pack being more vulnerable to the “Great Arctic Cyclone” in early August 2012 (Simmonds and Rudeva 2012), as a consequence of the preconditioning through decades of overall ice reductions (Parkinson and Comiso 2013), the heat storage in the NSTM (Zhang et al. 2013), and the Mackenzie River discharge (Nghiem et al. 2014). The 2016 ice minimum was related to the export of sea ice in the Beaufort Sea from January to April of 2016 that was far higher than the average level of 38 years during 1979–2016 and was the maximum of the identified extreme events. This is associated with the divergence of the ice cover that created a large area of open water and new ice, causing the ice pack to break up prematurely in summer (Babb et al. 2019).

Hence, regarding the causes of the four extremely low ice events identified, previous studies have revealed some common aspects but also some subtle differences in terms of how the ice cover was conditioned for summer melt based on lingering effects of the previous summer and the combination of dynamic and thermodynamic processes during the preceding fall and winter. Quantitative analyses of these processes require examining the budget of ice volume changes. This is difficult to achieve based on observational data, but it is feasible through analyzing the simulation results of coupled ocean and sea ice models. In this study, we carry out a modeling analysis with an objective to quantify the relative contributions of the thermodynamic and dynamic processes to the formation of the extremely low ice events. The focus is put on the summer 1998 event, but the similarity and difference of this event with the other three events will also be discussed.

Section 2 provides the description of the coupled ocean and sea ice model, model evaluation, and analysis methods. Section 3 describes the observed and modeled sea ice condition in the summer of 1998. Sections 4 and 5 quantify the contributions of the thermodynamic and dynamic processes to ice thickness and volume changes in summer and the preceding fall and winter. In particular, section 4 compares the contributions in 1998 and the mean condition over 1995–2015, and section 5 examines the interannual variations during 1995–2015. A summary of conclusions is provided in section 6.

2. Model and analysis methods

a. Model configuration

The coupled ocean and sea ice model is a configuration created using version 3.6 of the Nucleus for European Modelling of the Ocean (NEMO3.6; Madec 2008), with the sea ice component being version 3 of the Louvain-la-Neuve sea ice Model (LIM3; Vancoppenolle et al. 2009; Rousset et al. 2015). LIM3 supports multiple ice thickness categories that can better resolve the intense growth and melting of the thin ice, and ice thickness redistribution caused by ridging and rafting (Uotila et al. 2017). The default five categories within LIM3 (defined for thickness ranges of 0–0.45, 0.45–1.13, 1.13–2.14, 2.14–3.67, and >3.67 m) are used in the simulation. The model domain covers the Arctic and the North Atlantic (north of 26°N) and North Pacific (north of 45°N) Oceans. The horizontal mesh is extracted from the global tripolar ORCA grid (Madec and Imbard 1996) created by the Drakkar Group (2007) with a nominal horizontal resolution of 1/4° in latitude and longitude. The vertical space is discretized with a maximum of 75 geopotential z levels, with high resolution in the upper ocean (19 levels for the top 50 m).

The model is initialized with temperature and salinity from the October fields of the World Ocean Atlas (WOA, version 2) (Locarnini et al. 2013; Zweng et al. 2013), and sea surface height and sea ice concentration and thickness from October 1993 fields of the Global Ocean Reanalysis and Simulations (GLORYS2v4) produced by the Mercator Ocean, France (Ferry et al. 2012). The inputs to the lateral open boundary conditions include monthly ocean temperature, salinity, and horizontal ocean currents from GLORYS2v4, and tidal elevation and currents of five major constituents (M2, S2, N2, K1, and O1) from the global tidal solution TPXO8 of the Oregon State University (http://volkov.oce.orst.edu/tides). The inputs for surface forcing are obtained from the DRAKKAR Forcing Sets, version 5.2 (DFS5.2; https://www.drakkar-ocean.eu/; Dussin et al. 2016), including the 3-hourly wind at 10-m height, air temperature and specific humidity at 2-m height, and daily mean shortwave and longwave radiation and precipitation. The freshwater inputs from rivers use the monthly climatology of Dai and Trenberth (2002). The prognostic equations for water temperature and salinity are solved without including any restoring terms.

The model simulation starts on 1 October 1993 and ends on 31 December 2015. The simulation duration corresponds to the availability of the GLORYS2v4 data (since 1993) that provides the sea ice initial condition and latera l open boundary condition. Five-day-averaged model fields are saved for analysis.

b. Model evaluation

The model evaluation, presented previously in Luo et al. (2019), demonstrates that the model reproduces the main characteristics of sea ice, hydrography, and circulation in the Arctic Ocean. A follow-on study on the heat budget during the ice melt season by Wang et al. (2019) compared the simulated ice thickness with drill-hole observations during the melt season of 2002–04 in the Chukchi and Beaufort Seas. The following evaluations focus on sea ice in the Beaufort Sea, the area confined by the boundaries at 75°N (north), 150°W (west), 125°W (east), and the coastline (south), denoted by red lines in Fig. 1. For a specific variable, the observed and modeled values are denoted as Xoi and Xmi, where i is the index of a sample from a total number of N samples. Besides correlations, the mean bias (BIAS) and the relative average error (RAE) are according to

BIAS=X¯mX¯o,
RAE=i=1N(XmiXoi)2i=1N(|XmiX¯o|2+|XoiX¯o|2)×100%,

where an overbar denotes the time average over the N samples. Throughout this paper we define September in each year as the end of summer, May–September as the melt season, January–April as the preceding winter, and October–December of the previous year as the preceding fall.

Fig. 1.
Fig. 1.

(left) Bathymetry of the study area. BS is delineated with red lines. (top right) Ice thickness from 5-day-averaged model simulation (red solid circles) and drill-hole observations (mean plus and minus one standard deviation denoted by blue line and vertical bars) at locations shown by red circles with station numbers in white on map in the left panel (Shirasawa et al. 2009). (bottom right) The anomalies of the observed (blue line) and the simulated (red line) ice draft averaged over January–April at a mooring site (70°48.6′N, 133°43.4′W) on the MS (marked by the red star on map in the left panel), based on observations during 1992–2003 (Melling et al. 2005) and model results during 1995–2003, respectively.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

Figure 1 (top right) compares the simulated ice thickness with drill-hole observations in the Pacific sector of the Arctic. The drill-hole data are taken from Shirasawa et al. (2009) based on observations during 2002–04 by the Western Arctic Shelf-Basin Interaction Project, at locations shown on the left panel of Fig. 1. The modeled ice thickness is selected from the 5-day-averaged output at the time and location close to that of observations. The modeled values are close to the observed mean values (within the range of the observed standard deviations), with BIAS = 0.1 m and RAE = 19.9%, and the two have a significant correlation of r = 0.8 (p < 0.001). Figure 1 (bottom right) compares the anomalies of the observed and the simulated ice draft averaged over January–April. The observed ice draft data were taken from Melling et al. (2005) based on ice profiling sonar during 1992–2003 at a mooring site (70°48.6′N, 133°43.4′W) near the shelf break. The modeled ice draft is calculated by multiplying a factor of 0.9 (ratio of ice density vs water density) to the ice thickness from the model grid covering the mooring site, and the anomaly is computed from the model results of 1995–2003. The two time series show similar interannual variations with a significant correlation of r = 0.7 (p < 0.05), and both identify the extremely thin ice condition in 1998. The modeled and observed anomaly values are close, with BIAS = 0.1 m and RAE = 32.7%.

The sea ice concentration data used for evaluation are the Nimbus-7 SMMR and DMSP SSM/I-SSMIS Passive Microwave Data, version 1 (Cavalieri et al. 1996). This dataset is obtained from the U.S. National Snow and Ice Data Center (NSIDC) (https://nsidc.org/data/nsidc-0051). It is derived from the daily brightness temperatures in the polar stereographic projection with a grid size of 25 km × 25 km. The data have been provided with daily intervals since August 1987. From January 1979 to July 1987, the data are available in alternate days, and the values on days without data are filled through linearly interpolating the data of the surrounding 2 days.

The open water is defined as where ice concentration is less than 15%. Figure 2 shows the variations of open-water area in the Beaufort Sea from 1979 to 2017 calculated from the satellite data. From May–September, the regional mean open-water area was typically below 1.7 × 105 km2, yet while open-water area has been increasing over the last decade there are four distinct peaks exceeding 2.7 × 105 km2 during 1998, 2008, 2012 and 2016. During these 4 years, the open-water area averaged over May–September occupied more than half of the total area of the Beaufort Sea. In 1998, the open-water area exceeded 3.0 × 105 km2 from mid-June to mid-October, although the maximum value (4.1 × 105 km2 on 12 September) was less than that in 2008, 2012, and 2016. Figure 2 (dashed blue curve) shows the model-simulated open-water area averaged over May–September in each year during 1995–2015. The model well reproduces the observed interannual variations (solid blue curve) including the three peaks in 1998, 2008, and 2012, and the two time series during 1995–2015 have a correlation of r = 0.87 (p < 0.01) and RAE = 27.9%. During 1995–2015, the mean plus and minus one standard deviation of open-water area averaged over May–September are (0.9 ± 0.6) × 105 km2 according to the model, compared with (1.4 ± 0.7) × 105 km2 based on the satellite data.

Fig. 2.
Fig. 2.

Open-water area of the BS (125°–150°W and south of 75°N, outlined by red lines in Fig. 1). The satellite observational data cover 1979–2017, with color shading showing the daily variations and the solid blue curve representing the averaged values during the melt season (May–September) of each year. The dashed blue curve represents the modeled variations averaged over the melt season during 1995–2015. The red vertical boxes highlight the 4 years with the extremely low summer sea ice conditions: 1998, 2008, 2012, and 2016.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

Focusing on the melt season in 1998, Fig. 3 compares the monthly ice concentration based on satellite observations (top row) and model simulation (middle row). The model reasonably reproduces the observed formation and expansion of the open-water area from May to September, and the general pattern of ice retreat from the southeastern to the northwestern parts, accompanied by the decrease in sea ice concentration. For each month, the BIAS between modeled and observed ice concentration is computed using the monthly mean values (i.e., those presented in the top and middle rows of Fig. 3) over the ice covered regions defined according to either the modeled or observed concentration being greater than 15%. From May to September, the BIAS values are 19%, 16%, 23%, 33%, and 28%, respectively. Figure 4 (top and middle rows) presents the observed and modeled ice concentration averaged over May–September. For both the mean condition over 1995–2015 and in 1998, the model underestimates the open-water area and overestimates the ice concentration over the ice covered regions in the melt season. On the other hand, the model well reproduces the observed anomalously low ice concentration in 1998 relative to the 1995–2015 mean, in both the magnitude and spatial distribution.

Fig. 3.
Fig. 3.

Monthly mean sea ice condition from May to September in 1998: (top) satellite-observed ice concentration (Cice), model-simulated (middle) ice concentration, and (bottom) thickness (Hice).

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

Fig. 4.
Fig. 4.

Sea ice condition averaged over May–September: (top) satellite-observed ice concentration (Cice) and model-simulated (middle) ice concentration and (bottom) thickness (Hice) (left) for the mean condition during 1995–2015, (center) in 1998, and (right) for 1998 minus the 1995–2015 mean.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

Finally, the observed and simulated sea ice drift are compared. The observational dataset, version 4 of the Polar Pathfinder Daily 25 km Equal-Area Scalable Earth Grid (EASE-Grid) Sea Ice Motion Vectors (Tschudi et al. 2019), is retrieved from NSIDC (https://nsidc.org/data/NSIDC-0116/versions/4). The daily observational data are interpolated onto the model grid before the comparison. Figure 5 presents the sea ice drift averaged for three seasons: fall (October–December), winter (January–April), and the melt season (May–September), for the 1995–2015 mean condition and in 1998, respectively. Note that the preceding fall of 1998 is October–December 1997. The model well captures the observed ice drift, in terms of directions and magnitudes, their spatial and seasonal variations, and the anomalous condition of 1998 relative to the 1995–2005 mean. Quantitatively, we compare the magnitudes and directions of ice drift averaged over the Beaufort Sea area for each season. For the 1995–2015 mean condition, the observed (modeled) average magnitudes are 5.7 (5.8) cm s−1 for fall, 2.5 (2.6) cm s−1 for winter, and 2.6 (3.0) cm s−1 for the melt season. For 1998, the corresponding values are 6.2 (7.0) cm s−1 for the preceding fall, 5.0 (4.4) cm s−1 for the preceding winter, and 5.1 (4.3) cm s−1 for the melt season. The ice drift direction is 0° if the drift is northward and increases as the turns in a clockwise direction. The BIAS for the ice drift direction in the three seasons are −3.1°, −12.1°, and −7.4° for the 1995–2015 mean condition, and −9.9°, −10.2°, and −13.8° for 1998. This indicates that the modeled drift directs slightly to the right of the observed drift.

Fig. 5.
Fig. 5.

The spatial distribution of sea ice drift averaged for the preceding (left) fall and (center) winter and for (right) the melt season. (top) The mean condition during 1995–2015; (bottom) condition in 1998. Blue and red vectors represent satellite-observed and model sea ice drift, respectively.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

c. Diagnostics of model results

The following diagnostics will be derived from model results. First, the ice thickness changes, denoted as ΔHtotal, are calculated as the difference between the last and the first model ice thickness output (5-day average) for a given period. The thermodynamic contribution to the ice thickness change (ΔHthermal) represents the vertical heat fluxes through the atmosphere–ice and ice–ocean interfaces, provided by a diagnostic variable from model output. The ice thickness change due to dynamic processes (i.e., advection and deformation), denoted as ΔHdynamic, is defined as ΔHtotal minus ΔHthermal. This simple approach to quantify the thermodynamic and dynamic ice thickness change is similar to that used in previous studies (e.g., Lindsay and Zhang 2005; Steele et al. 2015; Hu et al. 2018). By integrating ΔHthermal, ΔHdynamic, and ΔHtotal over the Beaufort Sea for a specific time interval, we obtain the corresponding ice volume changes denoted as ΔVthermal, ΔVdynamic, and ΔVtotal. The lateral ice volume transport (ΔVlateral) is decomposed into three components, that is, the transport across the northern boundary (F75°N), western boundary (F150°W), and eastern boundary (F125°W), respectively.

Based on satellite-derived ice drift during 1997–2012, Kwok (2006) estimated the ice-area fluxes through major gates of the Canadian Arctic Archipelago (CAA). The main CAA gate included in our eastern boundary (125°W) is the entrance to the Amundsen Gulf (Fig. 1). Taking January–April as an example, Kwok (2006) estimated a total ice-area flux of about 40 × 103 km2 over the 4 months. Taking the mean ice thickness as 2 m, the corresponding total ice volume transport is about 80 km3. Our model solution provides an estimate of 27 ± 45 km3 (mean ± standard deviation) of the total ice volume transport across the eastern boundary, during January–April and averaged over 1997–2002. Hence our model underestimates the observation-derived estimate. However, even the observation-derived magnitude across the eastern boundary is negligible compared with that across the northern and western boundaries (section 5).

Next, the ice cover evolution is analyzed in the thickness space (with a bin width of 0.2 m) based on the cell-average ice thickness from model output. The ice cover in each thickness bin is weighted by the total area of the Beaufort Sea (4.85 × 105 km2). This diagnostic variable is similar to the concept of ice thickness distribution (ITD; e.g., Haas et al. 2010), but with the information of the fraction of ice cover included.

3. Sea ice condition during the melt season of 1998

The satellite-observed and model-simulated monthly spatial maps of sea ice concentration from May to September 1998 in the Beaufort Sea are shown in the top and middle rows of Fig. 3. In May, sea ice concentration was low in the southeastern part of the Beaufort Sea [e.g., on the Mackenzie Shelf (MS)] and was high in the rest areas. In June, the open water already formed and occupied the southeastern part, while the sea ice concentration in the northwestern part was still high. In July, the whole area south of 73°N became ice free. In August and September, the ice edge was similar to that in July, though the ice concentration within the ice edge was lower than in July.

Figure 3 (bottom row) presents the modeled monthly ice thickness from May to September. In May, the ice thinner than 2 m occupied about half of the Beaufort Sea in the southeast, while ice thicker than 3 m existed north of 73°N. Thus, a “front” of ice thickness existed, with the MYI in the northern area and the FYI in the southern area. From June to September, this ice thickness front kept its position, while the thin ice at the southern side gradually melted from east to west and the open-water area grew.

Figure 4 presents the observed (top row) and modeled sea ice concentration (middle row) averaged over May–September in 1998 and the mean condition during 1995–2015. According to observations, in 1998 the ice concentration was 15%–30% smaller than the 1995–2015 mean and a larger open-water area existed in the southeastern part. Despite overestimating the observed values, the modeled ice concentration in 1998 was 15%–30% less than the 1995–2015 mean south of 74°N extending from the Beaufort Sea westward to the Chukchi Plateau (CP). Figure 4 (bottom row) shows the modeled ice thickness. There is a good correspondence between the ice thickness and ice concentration in terms of the “anomalous” low in 1998. Over the areas extending from the southern Beaufort Sea to the Chukchi Plateau, the ice thickness in 1998 was 1.0–2.0 m lower than the 1995–2015 mean.

4. Contributions of thermodynamic and dynamic processes to ice thickness change: 1998 versus mean condition over 1995–2015

a. Melt season (May–September)

Figure 6 shows the spatial distributions of ΔHtotal, ΔHthermal, and ΔHdynamic over the ice melt season (May–September). Averaged over 1995–2015, the thermodynamic process (represented by ΔHthermal) reduced the ice thickness by 1.5–2.5 m in most parts of the Beaufort Sea. The dynamic process (represented by ΔHdynamic) contributed to an ice thickness increase of ~1.0 m in the northern part and a decrease of less than 1.0 m in the southern part. In 1998, the thermodynamic process caused a thickness reduction of about 3.0 m in the northern part and less than 1.0 m in the southern part. The dynamic process caused a thickness increase of about 2.5 m along the edge of the Arctic perennial sea ice and about 3.0 m near 73°N, and a decrease of less than 1.5 m in the southern part.

Fig. 6.
Fig. 6.

Model-simulated spatial distributions of (right) the total sea ice thickness change, and the contributions from (left) the thermodynamic and (center) dynamic processes, integrated from May to September (top) for the mean condition during 1995–2015 and (bottom) in 1998.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

This distribution of ΔHdynamic is associated with the ice drift field (Fig. 5, right column), driven by the Beaufort Gyre (Petty et al. 2016). For both the mean condition during 1995–2015 and in 1998, the ice drift displays an anticyclonic circulation pattern, with southward motion in the northern area and westward motion in the southern area. Compared with the mean condition, the summer of 1998 showed a stronger ice drift that favored larger increases of ice thickness near 73°N.

The combination of thermodynamic and dynamic processes caused a net decrease in sea ice thickness in most areas of the Beaufort Sea (Fig. 6, right column). Compared with the mean condition, in the melt season of 1998 more ice was lost in the southern and southwestern regions due mostly to the dynamic process. Particularly in the southwestern region, more ice was exported to the Chukchi Sea. In the northern part, the contributions of thermodynamic and dynamic processes are generally in opposite signs. In 1998, the significant increase due to the dynamic process was mostly offset by the ice loss due to the thermodynamic process. Averaged over the study area, the net decrease in ice thickness was 1.4 m in 1998 compared with 1.5 m during 1995–2015.

Besides the melt season (May–September), Fig. 5 also presents the modeled and observed ice drift averaged over the preceding fall (October–December) and winter (January–April). Hence, the bottom-left panel represents the condition in October–December 1997. The ice drift was strong during October–December, weakened during January–April, and strengthened again during May–September. The comparison between the mean condition of 1995–2015 and in 1998 for the preceding fall and winter will be discussed in the following section.

b. preceding fall and winter (October–April)

Previous studies (e.g., Lindsay and Zhang 2005; Hutchings and Rigor 2012) have pointed out that thinner ice before the melt season can enhance the summer ice loss via the reduced ice–albedo feedback. This was indeed the case for 1998. In May, the ice was anomalously thin, and thin ice covered more than half of the Beaufort Sea (Fig. 3). But what causes the thin ice condition in May? Here we examine the conditions during the preceding fall and winter from October 1997 to April 1998.

Figure 7 presents the ice cover evolution in the ice thickness space from October to the following May. Averaged over 1995–2015, from October to December, the thickness distribution showed a bimodal pattern, with the thin and newly formed ice growing thicker to 0.8–1.6 m and the thicker ice of 2.0–3.0 m gradually increasing its coverage. During January–May the distribution became unimodal of thicker ice types spreading the range of 2.0–4.0 m and the dominant thickness range of 2.4–3.6 m. Hence, the ice cover evolution in thickness space presented different characteristics in the preceding fall (October–December) and winter (January–April). The area-averaged ice thickness (red curve) decreased from 3.2 m at the start of October to the minimum value of 1.6 m in mid-November and gradually increased to 3.0 m at end of May. By comparison, prior to the summer of 1998, the area-averaged ice thickness decreased from 3.2 m at the start of October 1997 to the minimum value of 1.6 m near the end of December, increased to 2.5 m in mid-February and kept nearly constant until the end of April, and decreased to 2.2 m at the end of May. The ice thickness distribution presented the typical bimodal pattern during October–December 1997 and also in January 1998, with both the thin/new ice and thick ice occupying larger areas in the Beaufort Sea than the mean condition during 1995–2015. During January–February 1998, the thinner seasonal ice grew to 1.4–2.0 m thick, and the thicker ice of 2.6–3.0 m had a large coverage. During March–May 1998, the thickness distribution spread over a wide range of 0.2–4.2 m, with the dominant thickness of about 2.0 m. It is worth noting that the ice of about 2 m by April in the Beaufort Sea was seasonal, while the thick ice was either deformed FYI or MYI. This is consistent with the observation-derived ITD by Haas et al. (2010). The larger coverage of FYI in the Beaufort Sea during the preceding winter of 1998 is in agreement with the shift of ice type based on analysis of the ice charts from the Canadian Ice Services (Babb et al. 2019).

Fig. 7.
Fig. 7.

Ice-cover evolution in the thickness space, in terms of the percentage (color shading) of the ice coverage with different thickness (y axis, calculated for bin width of 0.2 m) relative to the total area of the Beaufort Sea from October to the following May (x axis), and the area-averaged ice thickness (red curve): (top) averaged over 1995–2015 and (bottom) from October 1997 to May 1998.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

Figure 8 shows the spatial distributions of sea ice thickness in the preceding fall and winter. Compared with the mean condition during 1995–2015, in the preceding fall of 1998, the ice was about 0.5 m thinner in the thin-ice-covered southern area and about 0.5 m thicker in the northern area. In the preceding winter averaged over 1995–2015, most areas of the Beaufort Sea were covered by ice thicker than 2.5 m, while in 1998 sea ice thinner than 2.0 m ruled a large area in the southeastern part. The anomalously thin ice, occupying a large area of the southeastern part of the Beaufort Sea during the preceding winter, was the precondition that led to the reduced ice cover during the summer of 1998.

Fig. 8.
Fig. 8.

Model-simulated ice thickness averaged over (top) the preceding fall (October–December) and (bottom) preceding winter (January–April): (left) mean condition during 1995–2015, (center) in 1998, and (right) 1998 minus the 1995–2015 mean.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

The thermodynamic and dynamic contributions to ice thickness changes in the preceding fall and winter of 1998 are estimated (Fig. 9). In the preceding fall, the thermodynamic process increased ice thickness by 0.5–2.5 m. For the dynamic process, the distribution of ΔHdynamic corresponds to the advection of ice from the northern part to the southern part, and then westward to the Chukchi Sea, consistent with the ice drift direction (Fig. 5, left column). In the preceding winter, surface cooling represented in the thermodynamic contribution caused a similar amount of ice growth as in the preceding fall, while the dynamic process caused an overall ice loss. Compared to the mean condition during 1995–2015, the preceding winter of 1998 showed stronger westward velocities (Fig. 5, center column), implying more ice exported to the Chukchi Sea. Contributions of the two processes were largely offset, leading to no significant change in ice thickness in the preceding winter of 1998.

Fig. 9.
Fig. 9.

As in Fig. 6, but for (top) the preceding fall and (bottom) preceding winter of 1998.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

5. Interannual variations of thermodynamic and dynamic contributions to ice volume changes during 1995–2015

Figure 10 shows the interannual variations of ΔVthermal, ΔVdynamic, and ΔVtotal for the preceding fall and winter, and the melt season and the regional ice volume at the end of each phase. In all these years, the thermodynamic process contributed to an increase in ice volume in the preceding fall and winter and a decrease of ice volume in the melt season. While ΔVthermal shows interannual variations for all three phases, the thermodynamic process alone does not offer an obvious explanation for the extremely low ice conditions in the summers of 1998, 2008, and 2012. The contributions of the dynamic process show significant interannual variations. The year of 1998 is unique in the sense that the sea ice export occurred both in the preceding fall and winter, and nearly offsets the ice volume increase due to the thermodynamic process. The opposite contributions of the two processes can also be found in the preceding fall for about 1/3 of the study years (e.g., 1995, 1997, 2003, 2004, 2005, 2008, 2010), but in the preceding winter only in 1998. Regarding the regional ice volume at the end of each season, the values at the end of September and December 1997 are close to the mean seasonal values during 1995–2015, but anomalously low values are found at the end of April and September 1998. The low value of regional ice volume at the end of summer (September) 1998 can be attributed to that at the end of the preceding winter (April) because the total ice volume loss in the summer of 1998 was not particularly high. Therefore, the extremely low ice in the summer of 1998 was the result of ice export in the preceding seasons, particularly in the preceding winter.

Fig. 10.
Fig. 10.

Sea ice volume changes in the Beaufort Sea during 1995–2015 in the (top) preceding fall, (middle) preceding winter, and (bottom) melt season. Red solid lines represent regional ice volume at the end of each phase, with dotted lines representing the 1995–2015 mean values. Green, dark blue, and light blue bars denote the total ice volume changes and the contributions due to thermodynamic and dynamic processes, respectively. A positive value means an increase in the ice volume within the area defined in Fig. 1. The gray vertical bars highlight the 3 years with extremely low summer sea ice conditions in the Beaufort Sea: 1998, 2008, and 2012.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

In a recent study, Babb et al. (2019) attributed the low ice event in the summer of 2016 to the precondition due to the dynamic process in the preceding winter, similar to the cause of the low ice event of 1998 discussed above. However, this mechanism of 1998 does not apply to the extremely low ice conditions in the summers of 2008 and 2012. Here the model results (Fig. 10) suggest that the 2008 event can be attributed to the relatively small regional ice volume at the end of summer 2007 and large ice export during the preceding fall. The 2012 event was mainly caused by the ice melt in summer but not the preconditioning because the regional ice volume at the end of preceding winter was nearly the same as the mean value during 1995–2015.

Next, Fig. 11 shows the time series of the lateral ice volume transport ΔVlateral and its three components, F75°N, F150°W, and F125°W, respectively. Overall, the ice volume transport through the eastern boundary was negligible, and the net lateral advection (ΔVlateral) was approximately the sum of the opposite contributions by the import across the northern boundary and export through the western boundary. In the preceding winter of 1998, the uniquely large net ice volume export out of the Beaufort Sea was the result of small ice import through the northern boundary relative to the large export through the western boundary. In the preceding winter of all other years, the import and export can both be large (e.g., in 2013), but they were largely offset thus resulting in smaller values of net lateral transport. In the preceding fall of 1998, the magnitude of export was also larger than that of import, but similar situations also happened in other years, for example, 2003, 2004, 2005, 2008, and 2010. Hence, the most obvious explanation for the extremely low sea ice condition in the summer of 1998 is the fact that the westward export was considerably greater than the southward import in the preceding seasons, particularly in the preceding winter (F150°W = −9.4 × 102 km3 vs F75°N = 2.4 × 102 km3). This can be related to the occurrence of consecutively strong easterly winds during this period, as noted previously by Maslanik et al. (1999).

Fig. 11.
Fig. 11.

Lateral sea ice volume transport into the Beaufort Sea during 1995–2015 in the (top) preceding fall, (middle) preceding winter, and (bottom) melt season. Blue lines represent the net lateral ice transports. Gray, purple, and yellow bars denote transport through sections along 125°W, 75°N, and 150°W, respectively. A positive value means sea ice advected into the Beaufort Sea.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0943.1

6. Conclusions

Time series of the open-water area in the Beaufort Sea during 1979–2017, derived from satellite observations of ice concentration, shows four distinct peaks in summers of 1998, 2008, 2012, and 2016. The observed sea ice variability during 1995–2015 is reasonably reproduced by the simulation results of a medium-resolution coupled ocean and sea ice model. The model results show that in the southern Beaufort Sea, both the ice area and thickness in the summer of 1998 were substantially smaller than the mean condition during 1995–2015.

The model-simulated ice thickness changes are analyzed to quantify the contributions of thermodynamic and dynamic processes. In the melt season (May–September) of 1998, the net ice volume loss was slightly smaller than the mean condition in 1995–2015. The extremely low ice condition in the summer of 1998 was caused by the thin ice condition before the melt season. This points to the importance of preconditioning of the ice cover during the prior seasons. Through analysis of the model results, the abnormally large reduction in ice volume due to the dynamic process, during the preceding winter (January–April), is identified to be the main cause of the extremely low sea ice condition in the summer of 1998. The excessively large net ice export was the consequence of the westward export being relatively larger than the southward import. The ice condition before the preceding winter had limited influence on this event.

The dynamic process in the preceding winter was also the cause of low ice condition in the summer of 2016, according to the recent study by Babb et al. (2019). However, the causes for the 2008 and 2012 events were different. Our model analyses suggest that the 2008 event was due to the small regional ice volume at the end of the summer of 2007 and ice export during the preceding fall while the 2012 event was caused by the excessive melting in summer.

The causes of the extremely low ice events in 1998, 2008, and 2012, derived from budget analyses of model simulation results, are in general agreement with the findings of previous studies based on analyses of observations and model simulations (e.g., Maslanik et al. 1999; Melling et al. 2005; Perovich et al. 2008; Parkinson and Comiso 2013; Zhang et al. 2013). For the 1998 event, our modeling analysis complements and also offers more insight than the study by Maslanik et al. (1999) based on examining the atmospheric forcing data. Through identifying the anomalously persistent southerly and easterly winds from November 1997 through April 1998, Maslanik et al. (1999) predicted the anomalous export of MYI prior to the summer of 1998. Model results suggest that the excessive ice export mainly occurred during January–April 1998 and further ruled out the preconditioning at the end of melt season (September 1997) and preceding fall (December 1997) to be important for the summer 1998 event. On the other hand, in the fall of 1997 near the center of the Beaufort Gyre (north of our study area), the observed sea ice was thinner, and the upper ocean was warmer and fresher than in previous years (McPhee et al. 1998). It remains interesting to analyze if the preceding fall anomaly can affect the ice and ocean condition locally in the preceding winter of 1998, and then subsequently influenced the summer of 1998 ice melting in the southern area of Beaufort Sea. Last, while the present study demonstrates the usefulness of modeling analysis for understanding the forcing mechanisms of Arctic sea ice variations, the models need to be continuously improved with the guidance of observational data for long-term simulations of historical events and prediction of future changes.

Acknowledgments

This study was supported by the National Natural Science Foundation of China (41630969, 41941013, and 41806225) and the National Key Research and Development Program of China (2016YFC1401401). We are grateful to the NEMO development team for providing the state-of-the-art model, and the expert advice from colleagues within the CONCEPTS network of Canada and the Mercator-Ocean International. We thank Dr. Jingen Xiao for useful discussion and Ms. Yali Wang for assistance in model evaluation and analysis. We thank three anonymous reviewers for very constructive comments that guided the revision of the original manuscript.

REFERENCES

  • Babb, D. G., R. J. Galley, D. G. Barber, and S. Rysgaard, 2016: Physical processes contributing to an ice free Beaufort Sea during September 2012. J. Geophys. Res. Oceans, 121, 267283, https://doi.org/10.1002/2015JC010756.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Babb, D. G., J. C. Landy, D. G. Barber, and R. J. Galley, 2019: Winter sea ice export from the Beaufort Sea as a preconditioning mechanism for enhanced summer melt: A case study of 2016. J. Geophys. Res. Oceans, 124, 65756600, https://doi.org/10.1029/2019JC015053.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cavalieri, D. J., C. L. Parkinson, P. Gloersen, and H. Zwally, 1996: Sea ice concentrations from Nimbus-7 SMMR and DMSP SSM/I passive microwave data, version 1. Subset used: Northern Hemisphere daily data (updated yearly), National Snow and Ice Data Center, accessed 31 January 2017, https://doi.org/10.5067/8GQ8LZQVL0VL.

    • Crossref
    • Export Citation
  • Dai, A., and K. E. Trenberth, 2002: Estimates of freshwater discharge from continents: Latitudinal and seasonal variations. J. Hydrometeor., 3, 660687, https://doi.org/10.1175/1525-7541(2002)003<0660:EOFDFC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Drakkar Group, 2007: Eddy-permitting ocean circulation hindcasts of past decades. CLIVAR Exchanges, No. 12, International CLIVAR Project Office, Southampton, United Kingdom, 8–10.

  • Dussin, R., B. Barnier, L. Brodeau, and J. Molines, 2016: The making of the Drakkar Forcing Set DFS5. DRAKKAR/MyOcean Rep., 01-04-16, 34 pp., https://www.drakkar-ocean.eu/publications/reports/report_DFS5v3_April2016.pdf.

  • Ferry, N., and Coauthors, 2012: GLORYS2V1 global ocean reanalysis of the altimetric era (1993–2009) at meso scale. Mercator Quarterly Newsletter, No. 44, Mercator Ocean, Ramonville Saint-Agne, France, 29–39.

  • Galley, R. J., E. Key, D. G. Barber, B. J. Hwang, and J. K. Ehn, 2008: Spatial and temporal variability of sea ice in the southern Beaufort Sea and Amundsen Gulf: 1980-2004. Geophys. Res. Lett., 113, C05S95, https://doi.org/10.1029/2007JC004553.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Galley, R. J., D. G. Babb, M. Ogi, B. G. T. Else, N.-X. Geilfus, O. Crabeck, D. G. Barber, and S. Rysgaard, 2016: Replacement of multiyear sea ice and changes in the open water season duration in the Beaufort Sea since 2004. J. Geophys. Res. Oceans, 121, 18061823, https://doi.org/10.1002/2015JC011583.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Haas, C., S. Hendricks, H. Eicken, and A. Herber, 2010: Synoptic airborne thickness surveys reveal state of Arctic sea ice cover. Geophys. Res. Lett., 37, L09501, https://doi.org/10.1029/2010GL042652.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Howell, S. E. L., M. Brady, C. Derksen, and R. E. J. Kelly, 2016: Recent changes in sea ice area flux through the Beaufort Sea during the summer. J. Geophys. Res. Oceans, 121, 26592672, https://doi.org/10.1002/2015JC011464.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hu, X., J. Sun, T. O. Chan, and P. G. Myers, 2018: Thermodynamic and dynamic ice thickness contributions in the Canadian Arctic Archipelago in NEMO-LIM2 numerical simulations. Cryosphere, 12, 12331247, https://doi.org/10.5194/tc-12-1233-2018.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hutchings, J. K., and I. G. Rigor, 2012: Role of ice dynamics in anomalous ice conditions in the Beaufort Sea during 2006 and 2007. J. Geophys. Res., 117, C00E04, https://doi.org/10.1029/2011JC007182.

    • Search Google Scholar
    • Export Citation
  • Kwok, R., 2006: Exchange of sea ice between the Arctic Ocean and the Canadian Arctic Archipelago. Geophys. Res. Lett., 33, L16501, https://doi.org/10.1029/2006GL027094.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kwok, R., and G. F. Cunningham, 2010: Contribution of melt in the Beaufort Sea to the decline in Arctic multiyear sea ice coverage: 1993-2009. Geophys. Res. Lett., 37, L20501, https://doi.org/10.1029/2010GL044678.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lindsay, R. W., and J. Zhang, 2005: The thinning of Arctic sea ice, 1988–2003: Have we passed a tipping point? J. Climate, 18, 48794894, https://doi.org/10.1175/JCLI3587.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Locarnini, R. A., and Coauthors, 2013: Temperature. Vol. 1, World Ocean Atlas 2013, NOAA Atlas NESDIS 73, 40 pp., http://data.nodc.noaa.gov/woa/WOA13/DOC/woa13_vol1.pdf.

  • Luo, X., and Coauthors, 2019: Evaluation of hindcast simulation with the ocean and sea-ice model covering the Arctic and adjacent oceans (in Chinese with English abstract). Haiyang Xuebao, 41, 112, http://english.hyxb.org.cn/EN/abstract/abstract5578.shtml.

    • Search Google Scholar
    • Export Citation
  • Madec, G., 2008: NEMO ocean engine, version 3.6. Note du Pôle de Modélisation, Institut Pierre-Simon Laplace 27, 386 pp.

  • Madec, G., and M. Imbard, 1996: A global ocean mesh to overcome the North Pole singularity. Climate Dyn., 12, 381388, https://doi.org/10.1007/BF00211684.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maslanik, J. A., M. C. Serreze, and T. Agnew, 1999: On the record reduction in 1998 western Arctic Sea-ice cover. Geophys. Res. Lett., 26, 19051908, https://doi.org/10.1029/1999GL900426.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maslanik, J. A., C. Fowler, J. Stroeve, S. Drobot, J. Zwally, D. Yi, and W. Emery, 2007: A younger, thinner Arctic ice cover: Increased potential for rapid, extensive sea-ice loss. Geophys. Res. Lett., 34, L24501, https://doi.org/10.1029/2007GL032043.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maslanik, J. A., J. Stroeve, C. Fowler, and W. Emery, 2011: Distribution and trends in Arctic sea ice age through spring 2011. Geophys. Res. Lett., 38, L13502, https://doi.org/10.1029/2011GL047735.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McPhee, M. G., T. P. Stanton, J. H. Morison, and D. G. Martinson, 1998: Freshening of the upper ocean in the Arctic: Is perennial sea ice disappearing? Geophys. Res. Lett., 25, 17291732, https://doi.org/10.1029/98GL00933.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Melling, H., D. A. Riedel, and Z. Gedal, 2005: Trends in the draft and extent of seasonal pack ice, Canadian Beaufort. Geophys. Res. Lett., 32, L24501, https://doi.org/10.1029/2005GL024483.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nghiem, S. V., D. K. Hall, I. G. Rigor, P. Li, and G. Neumann, 2014: Effects of Mackenzie River discharge and bathymetry on sea ice in the Beaufort Sea. Geophys. Res. Lett., 41, 873879, https://doi.org/10.1002/2013GL058956.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Onarheim, I. H., T. Eldevik, L. H. Smedsrud, and J. C. Stroeve, 2018: Seasonal and regional manifestation of Arctic sea ice loss. J. Climate, 31, 49174932, https://doi.org/10.1175/JCLI-D-17-0427.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Parkinson, C. L., and J. C. Comiso, 2013: On the 2012 record low Arctic sea ice cover: Combined impact of preconditioning and an August storm. Geophys. Res. Lett., 40, 13561361, https://doi.org/10.1002/grl.50349.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perovich, D. K., B. Light, H. Eicken, K. F. Jones, K. Runciman, and S. V. Nghiem, 2007: Increasing solar heating of the Arctic Ocean and adjacent seas, 1979-2005: Attribution and role in the ice-albedo feedback. Geophys. Res. Lett., 34, L19505, https://doi.org/10.1029/2007GL031480.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perovich, D. K., J. A. Richter-Menge, K. F. Jones, and B. Light, 2008: Sunlight, water, and ice: Extreme Arctic sea ice melt during the summer of 2007. Geophys. Res. Lett., 35, L11501, https://doi.org/10.1029/2008GL034007.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perovich, D. K., and Coauthors, 2011: Arctic sea-ice melt in 2008 and the role of solar heating. Ann. Glaciol., 52, 355359, https://doi.org/10.3189/172756411795931714.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Petty, A. A., J. K. Hutchings, J. A. Richter-Menge, and M. A. Tschudi, 2016: Sea ice circulation around the Beaufort Gyre: The changing role of wind forcing and the sea ice state. J. Geophys. Res. Oceans, 121, 32783296, https://doi.org/10.1002/2015JC010903.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Proshutinsky, A., R. H. Bourke, and F. A. McLaughlin, 2002: The role of the Beaufort Gyre in Arctic climate variability: Seasonal to decadal climate scales. Geophys. Res. Lett., 29, 2100, https://doi.org/10.1029/2002GL015847.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rigor, I. G., and J. M. Wallace, 2004: Variations in the age of Arctic sea-ice and summer sea-ice extent. Geophys. Res. Lett., 31, L09401, https://doi.org/10.1029/2004GL019492.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rousset, C., and Coauthors, 2015: The Louvain-La-Neuve sea ice model LIM3.6: Global and regional capabilities. Geosci. Model Dev., 8, 29913005, https://doi.org/10.5194/gmd-8-2991-2015.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shirasawa, K., H. Eicken, K. Tateyama, T. Takatsuka, and T. Kawamura, 2009: Sea-ice-thickness variability in the Chukchi Sea, spring and summer 2002-2004. Deep-Sea Res. II, 56, 11821200, https://doi.org/10.1016/j.dsr2.2008.10.015.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Simmonds, I., and I. Rudeva, 2012: The great Arctic cyclone of August 2012. Geophys. Res. Lett., 39, L23709, https://doi.org/10.1029/2012GL054259.

  • Steele, M., J. Zhang, and W. Ermold, 2010: Mechanisms of summertime upper Arctic Ocean warming and the effect on sea ice melt. J. Geophys. Res., 115, C11004, https://doi.org/10.1029/2009JC005849.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Steele, M., S. Dickinson, J. Zhang, and R. W. Lindsay, 2015: Seasonal ice loss in the Beaufort Sea: Toward synchrony and prediction. J. Geophys. Res. Oceans, 120, 11181132, https://doi.org/10.1002/2014JC010247.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stroeve, J. C., M. C. Serreze, F. Fetterer, T. Arbetter, W. Meier, J. Maslanik, and K. Knowles, 2005: Tracking the Arctic’s shrinking ice cover: Another extreme September minimum in 2004. Geophys. Res. Lett., 32, L04501, https://doi.org/10.1029/2004GL021810.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Timmermans, M. L., 2015: The impact of stored solar heat on Arctic sea ice growth. Geophys. Res. Lett., 42, 63996406, https://doi.org/10.1002/2015GL064541.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tivy, A., S. E. L. Howell, B. Alt, S. McCourt, R. Chagnon, G. Crocker, T. Carrieres, and J. J. Yackel, 2011: Trends and variability in summer sea ice cover in the Canadian Arctic based on the Canadian Ice Service Digital Archive, 1960-2008 and 1968-2008. J. Geophys. Res., 116, C03007, https://doi.org/10.1029/2009JC005855.

    • Search Google Scholar
    • Export Citation
  • Tschudi, M., W. N. Meier, J. S. Stewart, C. Fowler, and J. Maslanik. 2019: Polar Pathfinder Daily 25 km EASE-Grid Sea Ice Motion Vectors, Version 4. Subset used: Northern Hemisphere daily data (updated yearly), National Snow and Ice Data Center, accessed 16 April 2019, https://doi.org/10.5067/INAWUWO7QH7B.

    • Crossref
    • Export Citation
  • Uotila, P., D. Iovino, M. Vancoppenolle, M. Lensu, and C. Rousset, 2017: Comparing sea ice, hydrography and circulation between NEMO 3.6 LIM3 and LIM2. Geosci. Model Dev., 10, 10091031, https://doi.org/10.5194/gmd-10-1009-2017.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Vancoppenolle, M., T. Fichefet, H. Goosse, S. Bouillon, G. Madec, and M. A. M. Maqueda, 2009: Simulating the mass balance and salinity of Arctic and Antarctic sea ice. 1. Model description and validation. Ocean Modell., 27, 3353, https://doi.org/10.1016/j.ocemod.2008.10.005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, Y., X. Luo, Y. Zhang, Y. Qin, W. Zhao, Z. Zheng, H. Nie, and H. Wei, 2019: Heat budget analysis during the ice-melting season in the Chukchi Sea based on a model simulation (in Chinese with English abstract). Chin. Sci. Bull., 64, 34853497, https://doi.org/10.1360/TB-2019-0060.

    • Search Google Scholar
    • Export Citation
  • Zhang, J., R. Lindsay, A. Schweiger, and M. Steele, 2013: The impact of an intense summer cyclone on 2012 Arctic sea ice retreat. Geophys. Res. Lett., 40, 720726, https://doi.org/10.1002/grl.50190.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zweng, M., and Coauthors, 2013: Salinity. Vol. 2, World Ocean Atlas 2013, NOAA Atlas NESDIS 74, 39 pp., http://data.nodc.noaa.gov/woa/WOA13/DOC/woa13_vol2.pdf.

Save
  • Babb, D. G., R. J. Galley, D. G. Barber, and S. Rysgaard, 2016: Physical processes contributing to an ice free Beaufort Sea during September 2012. J. Geophys. Res. Oceans, 121, 267283, https://doi.org/10.1002/2015JC010756.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Babb, D. G., J. C. Landy, D. G. Barber, and R. J. Galley, 2019: Winter sea ice export from the Beaufort Sea as a preconditioning mechanism for enhanced summer melt: A case study of 2016. J. Geophys. Res. Oceans, 124, 65756600, https://doi.org/10.1029/2019JC015053.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cavalieri, D. J., C. L. Parkinson, P. Gloersen, and H. Zwally, 1996: Sea ice concentrations from Nimbus-7 SMMR and DMSP SSM/I passive microwave data, version 1. Subset used: Northern Hemisphere daily data (updated yearly), National Snow and Ice Data Center, accessed 31 January 2017, https://doi.org/10.5067/8GQ8LZQVL0VL.

    • Crossref
    • Export Citation
  • Dai, A., and K. E. Trenberth, 2002: Estimates of freshwater discharge from continents: Latitudinal and seasonal variations. J. Hydrometeor., 3, 660687, https://doi.org/10.1175/1525-7541(2002)003<0660:EOFDFC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Drakkar Group, 2007: Eddy-permitting ocean circulation hindcasts of past decades. CLIVAR Exchanges, No. 12, International CLIVAR Project Office, Southampton, United Kingdom, 8–10.

  • Dussin, R., B. Barnier, L. Brodeau, and J. Molines, 2016: The making of the Drakkar Forcing Set DFS5. DRAKKAR/MyOcean Rep., 01-04-16, 34 pp., https://www.drakkar-ocean.eu/publications/reports/report_DFS5v3_April2016.pdf.

  • Ferry, N., and Coauthors, 2012: GLORYS2V1 global ocean reanalysis of the altimetric era (1993–2009) at meso scale. Mercator Quarterly Newsletter, No. 44, Mercator Ocean, Ramonville Saint-Agne, France, 29–39.

  • Galley, R. J., E. Key, D. G. Barber, B. J. Hwang, and J. K. Ehn, 2008: Spatial and temporal variability of sea ice in the southern Beaufort Sea and Amundsen Gulf: 1980-2004. Geophys. Res. Lett., 113, C05S95, https://doi.org/10.1029/2007JC004553.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Galley, R. J., D. G. Babb, M. Ogi, B. G. T. Else, N.-X. Geilfus, O. Crabeck, D. G. Barber, and S. Rysgaard, 2016: Replacement of multiyear sea ice and changes in the open water season duration in the Beaufort Sea since 2004. J. Geophys. Res. Oceans, 121, 18061823, https://doi.org/10.1002/2015JC011583.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Haas, C., S. Hendricks, H. Eicken, and A. Herber, 2010: Synoptic airborne thickness surveys reveal state of Arctic sea ice cover. Geophys. Res. Lett., 37, L09501, https://doi.org/10.1029/2010GL042652.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Howell, S. E. L., M. Brady, C. Derksen, and R. E. J. Kelly, 2016: Recent changes in sea ice area flux through the Beaufort Sea during the summer. J. Geophys. Res. Oceans, 121, 26592672, https://doi.org/10.1002/2015JC011464.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hu, X., J. Sun, T. O. Chan, and P. G. Myers, 2018: Thermodynamic and dynamic ice thickness contributions in the Canadian Arctic Archipelago in NEMO-LIM2 numerical simulations. Cryosphere, 12, 12331247, https://doi.org/10.5194/tc-12-1233-2018.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hutchings, J. K., and I. G. Rigor, 2012: Role of ice dynamics in anomalous ice conditions in the Beaufort Sea during 2006 and 2007. J. Geophys. Res., 117, C00E04, https://doi.org/10.1029/2011JC007182.

    • Search Google Scholar
    • Export Citation
  • Kwok, R., 2006: Exchange of sea ice between the Arctic Ocean and the Canadian Arctic Archipelago. Geophys. Res. Lett., 33, L16501, https://doi.org/10.1029/2006GL027094.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kwok, R., and G. F. Cunningham, 2010: Contribution of melt in the Beaufort Sea to the decline in Arctic multiyear sea ice coverage: 1993-2009. Geophys. Res. Lett., 37, L20501, https://doi.org/10.1029/2010GL044678.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lindsay, R. W., and J. Zhang, 2005: The thinning of Arctic sea ice, 1988–2003: Have we passed a tipping point? J. Climate, 18, 48794894, https://doi.org/10.1175/JCLI3587.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Locarnini, R. A., and Coauthors, 2013: Temperature. Vol. 1, World Ocean Atlas 2013, NOAA Atlas NESDIS 73, 40 pp., http://data.nodc.noaa.gov/woa/WOA13/DOC/woa13_vol1.pdf.

  • Luo, X., and Coauthors, 2019: Evaluation of hindcast simulation with the ocean and sea-ice model covering the Arctic and adjacent oceans (in Chinese with English abstract). Haiyang Xuebao, 41, 112, http://english.hyxb.org.cn/EN/abstract/abstract5578.shtml.

    • Search Google Scholar
    • Export Citation
  • Madec, G., 2008: NEMO ocean engine, version 3.6. Note du Pôle de Modélisation, Institut Pierre-Simon Laplace 27, 386 pp.

  • Madec, G., and M. Imbard, 1996: A global ocean mesh to overcome the North Pole singularity. Climate Dyn., 12, 381388, https://doi.org/10.1007/BF00211684.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maslanik, J. A., M. C. Serreze, and T. Agnew, 1999: On the record reduction in 1998 western Arctic Sea-ice cover. Geophys. Res. Lett., 26, 19051908, https://doi.org/10.1029/1999GL900426.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maslanik, J. A., C. Fowler, J. Stroeve, S. Drobot, J. Zwally, D. Yi, and W. Emery, 2007: A younger, thinner Arctic ice cover: Increased potential for rapid, extensive sea-ice loss. Geophys. Res. Lett., 34, L24501, https://doi.org/10.1029/2007GL032043.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maslanik, J. A., J. Stroeve, C. Fowler, and W. Emery, 2011: Distribution and trends in Arctic sea ice age through spring 2011. Geophys. Res. Lett., 38, L13502, https://doi.org/10.1029/2011GL047735.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McPhee, M. G., T. P. Stanton, J. H. Morison, and D. G. Martinson, 1998: Freshening of the upper ocean in the Arctic: Is perennial sea ice disappearing? Geophys. Res. Lett., 25, 17291732, https://doi.org/10.1029/98GL00933.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Melling, H., D. A. Riedel, and Z. Gedal, 2005: Trends in the draft and extent of seasonal pack ice, Canadian Beaufort. Geophys. Res. Lett., 32, L24501, https://doi.org/10.1029/2005GL024483.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nghiem, S. V., D. K. Hall, I. G. Rigor, P. Li, and G. Neumann, 2014: Effects of Mackenzie River discharge and bathymetry on sea ice in the Beaufort Sea. Geophys. Res. Lett., 41, 873879, https://doi.org/10.1002/2013GL058956.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Onarheim, I. H., T. Eldevik, L. H. Smedsrud, and J. C. Stroeve, 2018: Seasonal and regional manifestation of Arctic sea ice loss. J. Climate, 31, 49174932, https://doi.org/10.1175/JCLI-D-17-0427.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Parkinson, C. L., and J. C. Comiso, 2013: On the 2012 record low Arctic sea ice cover: Combined impact of preconditioning and an August storm. Geophys. Res. Lett., 40, 13561361, https://doi.org/10.1002/grl.50349.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perovich, D. K., B. Light, H. Eicken, K. F. Jones, K. Runciman, and S. V. Nghiem, 2007: Increasing solar heating of the Arctic Ocean and adjacent seas, 1979-2005: Attribution and role in the ice-albedo feedback. Geophys. Res. Lett., 34, L19505, https://doi.org/10.1029/2007GL031480.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perovich, D. K., J. A. Richter-Menge, K. F. Jones, and B. Light, 2008: Sunlight, water, and ice: Extreme Arctic sea ice melt during the summer of 2007. Geophys. Res. Lett., 35, L11501, https://doi.org/10.1029/2008GL034007.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perovich, D. K., and Coauthors, 2011: Arctic sea-ice melt in 2008 and the role of solar heating. Ann. Glaciol., 52, 355359, https://doi.org/10.3189/172756411795931714.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Petty, A. A., J. K. Hutchings, J. A. Richter-Menge, and M. A. Tschudi, 2016: Sea ice circulation around the Beaufort Gyre: The changing role of wind forcing and the sea ice state. J. Geophys. Res. Oceans, 121, 32783296, https://doi.org/10.1002/2015JC010903.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Proshutinsky, A., R. H. Bourke, and F. A. McLaughlin, 2002: The role of the Beaufort Gyre in Arctic climate variability: Seasonal to decadal climate scales. Geophys. Res. Lett., 29, 2100, https://doi.org/10.1029/2002GL015847.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rigor, I. G., and J. M. Wallace, 2004: Variations in the age of Arctic sea-ice and summer sea-ice extent. Geophys. Res. Lett., 31, L09401, https://doi.org/10.1029/2004GL019492.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rousset, C., and Coauthors, 2015: The Louvain-La-Neuve sea ice model LIM3.6: Global and regional capabilities. Geosci. Model Dev., 8, 29913005, https://doi.org/10.5194/gmd-8-2991-2015.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shirasawa, K., H. Eicken, K. Tateyama, T. Takatsuka, and T. Kawamura, 2009: Sea-ice-thickness variability in the Chukchi Sea, spring and summer 2002-2004. Deep-Sea Res. II, 56, 11821200, https://doi.org/10.1016/j.dsr2.2008.10.015.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Simmonds, I., and I. Rudeva, 2012: The great Arctic cyclone of August 2012. Geophys. Res. Lett., 39, L23709, https://doi.org/10.1029/2012GL054259.

  • Steele, M., J. Zhang, and W. Ermold, 2010: Mechanisms of summertime upper Arctic Ocean warming and the effect on sea ice melt. J. Geophys. Res., 115, C11004, https://doi.org/10.1029/2009JC005849.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Steele, M., S. Dickinson, J. Zhang, and R. W. Lindsay, 2015: Seasonal ice loss in the Beaufort Sea: Toward synchrony and prediction. J. Geophys. Res. Oceans, 120, 11181132, https://doi.org/10.1002/2014JC010247.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stroeve, J. C., M. C. Serreze, F. Fetterer, T. Arbetter, W. Meier, J. Maslanik, and K. Knowles, 2005: Tracking the Arctic’s shrinking ice cover: Another extreme September minimum in 2004. Geophys. Res. Lett., 32, L04501, https://doi.org/10.1029/2004GL021810.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Timmermans, M. L., 2015: The impact of stored solar heat on Arctic sea ice growth. Geophys. Res. Lett., 42, 63996406, https://doi.org/10.1002/2015GL064541.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tivy, A., S. E. L. Howell, B. Alt, S. McCourt, R. Chagnon, G. Crocker, T. Carrieres, and J. J. Yackel, 2011: Trends and variability in summer sea ice cover in the Canadian Arctic based on the Canadian Ice Service Digital Archive, 1960-2008 and 1968-2008. J. Geophys. Res., 116, C03007, https://doi.org/10.1029/2009JC005855.

    • Search Google Scholar
    • Export Citation
  • Tschudi, M., W. N. Meier, J. S. Stewart, C. Fowler, and J. Maslanik. 2019: Polar Pathfinder Daily 25 km EASE-Grid Sea Ice Motion Vectors, Version 4. Subset used: Northern Hemisphere daily data (updated yearly), National Snow and Ice Data Center, accessed 16 April 2019, https://doi.org/10.5067/INAWUWO7QH7B.

    • Crossref
    • Export Citation
  • Uotila, P., D. Iovino, M. Vancoppenolle, M. Lensu, and C. Rousset, 2017: Comparing sea ice, hydrography and circulation between NEMO 3.6 LIM3 and LIM2. Geosci. Model Dev., 10, 10091031, https://doi.org/10.5194/gmd-10-1009-2017.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Vancoppenolle, M., T. Fichefet, H. Goosse, S. Bouillon, G. Madec, and M. A. M. Maqueda, 2009: Simulating the mass balance and salinity of Arctic and Antarctic sea ice. 1. Model description and validation. Ocean Modell., 27, 3353, https://doi.org/10.1016/j.ocemod.2008.10.005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, Y., X. Luo, Y. Zhang, Y. Qin, W. Zhao, Z. Zheng, H. Nie, and H. Wei, 2019: Heat budget analysis during the ice-melting season in the Chukchi Sea based on a model simulation (in Chinese with English abstract). Chin. Sci. Bull., 64, 34853497, https://doi.org/10.1360/TB-2019-0060.

    • Search Google Scholar
    • Export Citation
  • Zhang, J., R. Lindsay, A. Schweiger, and M. Steele, 2013: The impact of an intense summer cyclone on 2012 Arctic sea ice retreat. Geophys. Res. Lett., 40, 720726, https://doi.org/10.1002/grl.50190.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zweng, M., and Coauthors, 2013: Salinity. Vol. 2, World Ocean Atlas 2013, NOAA Atlas NESDIS 74, 39 pp., http://data.nodc.noaa.gov/woa/WOA13/DOC/woa13_vol2.pdf.

  • Fig. 1.

    (left) Bathymetry of the study area. BS is delineated with red lines. (top right) Ice thickness from 5-day-averaged model simulation (red solid circles) and drill-hole observations (mean plus and minus one standard deviation denoted by blue line and vertical bars) at locations shown by red circles with station numbers in white on map in the left panel (Shirasawa et al. 2009). (bottom right) The anomalies of the observed (blue line) and the simulated (red line) ice draft averaged over January–April at a mooring site (70°48.6′N, 133°43.4′W) on the MS (marked by the red star on map in the left panel), based on observations during 1992–2003 (Melling et al. 2005) and model results during 1995–2003, respectively.

  • Fig. 2.

    Open-water area of the BS (125°–150°W and south of 75°N, outlined by red lines in Fig. 1). The satellite observational data cover 1979–2017, with color shading showing the daily variations and the solid blue curve representing the averaged values during the melt season (May–September) of each year. The dashed blue curve represents the modeled variations averaged over the melt season during 1995–2015. The red vertical boxes highlight the 4 years with the extremely low summer sea ice conditions: 1998, 2008, 2012, and 2016.

  • Fig. 3.

    Monthly mean sea ice condition from May to September in 1998: (top) satellite-observed ice concentration (Cice), model-simulated (middle) ice concentration, and (bottom) thickness (Hice).

  • Fig. 4.

    Sea ice condition averaged over May–September: (top) satellite-observed ice concentration (Cice) and model-simulated (middle) ice concentration and (bottom) thickness (Hice) (left) for the mean condition during 1995–2015, (center) in 1998, and (right) for 1998 minus the 1995–2015 mean.

  • Fig. 5.

    The spatial distribution of sea ice drift averaged for the preceding (left) fall and (center) winter and for (right) the melt season. (top) The mean condition during 1995–2015; (bottom) condition in 1998. Blue and red vectors represent satellite-observed and model sea ice drift, respectively.

  • Fig. 6.

    Model-simulated spatial distributions of (right) the total sea ice thickness change, and the contributions from (left) the thermodynamic and (center) dynamic processes, integrated from May to September (top) for the mean condition during 1995–2015 and (bottom) in 1998.

  • Fig. 7.

    Ice-cover evolution in the thickness space, in terms of the percentage (color shading) of the ice coverage with different thickness (y axis, calculated for bin width of 0.2 m) relative to the total area of the Beaufort Sea from October to the following May (x axis), and the area-averaged ice thickness (red curve): (top) averaged over 1995–2015 and (bottom) from October 1997 to May 1998.

  • Fig. 8.

    Model-simulated ice thickness averaged over (top) the preceding fall (October–December) and (bottom) preceding winter (January–April): (left) mean condition during 1995–2015, (center) in 1998, and (right) 1998 minus the 1995–2015 mean.

  • Fig. 9.

    As in Fig. 6, but for (top) the preceding fall and (bottom) preceding winter of 1998.

  • Fig. 10.

    Sea ice volume changes in the Beaufort Sea during 1995–2015 in the (top) preceding fall, (middle) preceding winter, and (bottom) melt season. Red solid lines represent regional ice volume at the end of each phase, with dotted lines representing the 1995–2015 mean values. Green, dark blue, and light blue bars denote the total ice volume changes and the contributions due to thermodynamic and dynamic processes, respectively. A positive value means an increase in the ice volume within the area defined in Fig. 1. The gray vertical bars highlight the 3 years with extremely low summer sea ice conditions in the Beaufort Sea: 1998, 2008, and 2012.

  • Fig. 11.

    Lateral sea ice volume transport into the Beaufort Sea during 1995–2015 in the (top) preceding fall, (middle) preceding winter, and (bottom) melt season. Blue lines represent the net lateral ice transports. Gray, purple, and yellow bars denote transport through sections along 125°W, 75°N, and 150°W, respectively. A positive value means sea ice advected into the Beaufort Sea.

All Time Past Year Past 30 Days
Abstract Views 98 0 0
Full Text Views 400 107 14
PDF Downloads 479 84 6