Abstract

The authors investigate the interannual variations of freshwater content (FWC) and sea surface height (SSH) in the Beaufort Sea, particularly their increases during 2004–09, using a coupled ice–ocean model (CIOM), adapted for the Arctic Ocean to simulate the interannual variations. The CIOM simulation exhibits a (relative) salinity minimum in the Beaufort Sea and a warm Atlantic water layer in the Arctic Ocean, which is similar to the Polar Hydrographic Climatology (PHC), and captures the observed FWC maximum in the central Beaufort Sea, and the observed variation and rapid decline of total ice concentration, over the last 30 years. The model simulations of SSH and FWC suggest a significant increase in the central Beaufort Sea during 2004–09. The simulated SSH increase is about 8 cm, while the FWC increase is about 2.5 m, with most of these increases occurring in the center of the Beaufort gyre. The authors show that these increases are due to an increased surface wind stress curl during 2004–09, which increased the FWC in the Beaufort Sea by about 0.63 m yr−1 through Ekman pumping. Moreover, the increased surface wind is related to the interannual variation of the Arctic polar vortex at 500 hPa. During 2004–09, the polar vortex had significant weakness, which enhanced the Beaufort Sea high by affecting the frequency of synoptic weather systems in the region. In addition to the impacts of the polar vortex, enhanced melting of sea ice also contributes to the FWC increase by about 0.3 m yr−1 during 2004–09.

1. Introduction

The objective of this paper is to investigate the interannual variations of freshwater content (FWC) and sea surface height (SSH) in the Beaufort Sea, particularly their increases during 2004–09. Changes in the Arctic freshwater balance play an important role in the North Atlantic circulation. A 25% increase in freshwater discharge through Fram Strait maintained for two years can account for the salinity deficit observed in the North Atlantic during the “Great Salinity Anomaly” of the 1970s (Aagaard and Carmack 1989; Dickson et al. 1988; McPhee et al. 2009). Similar studies suggest that salinity anomalies can have profound impacts on the intensity of the Atlantic Ocean meridional overturning circulation (Curry and Mauritzen 2005).

A significant change in the FWC in the central Beaufort Sea has been detected during 2004–09 (Proshutinsky et al. 2009; McPhee et al. 2009). In 2008, the FWC increased by as much as 11 m at some stations, compared to the FWC climatology, and the location of the maximum shifted to the southeast. Measurements in the Canada and Markarov Basins suggest a FWC increase in the surveyed area by about 8500 km3. Accordingly, the steric sea level increased about 75% (McPhee et al. 2009).

Several factors could be responsible for the FWC increase, which include ice melting (Kwok and Cunningham 2010; Yamamoto-Kawai et al. 2009), river runoff (Yamamoto-Kawai et al. 2009), surface moisture flux, and surface wind stress. After 2000, a significant amount of multiyear sea ice (MYI) was lost in the Beaufort Sea. In particular, the net melt area of 490 × 103 km2 between 2005 and 2008 accounts for nearly 32% of the net loss of 1.54 × 106 km2 of Arctic Ocean multiyear sea ice coverage over the same period (Kwok and Cunningham 2010). The increased freshwater from the melted sea ice in the central Beaufort Sea in 2006 and 2007 corresponds to melting an additional 1.3 m yr−1 sea ice in the region (Yamamoto-Kawai et al. 2009). Although the surface freshening near the coast in 2007 was largely due to river runoff from the Mackenzie River, the composite runoff component showed large variability and there is no obvious trend (Yamamoto-Kawai et al. 2009). Moreover, Proshutinsky et al. (2009) speculated that one of the major causes for the variability in the FWC could be related to changes in the atmospheric circulation and Ekman pumping. In the Arctic Ocean, the largest freshwater storage is located in the Beaufort gyre (BG), a dominant anticyclonic circulation in the Beaufort Sea (Aagaard and Carmack 1989). On decadal time scales, the wind-driven circulation alternates between cyclonic and anticyclonic circulation regimes, with each regime persisting for 5–7 yr (Proshutinsky and Johnson 1997). The BG accumulates significant amounts of freshwater during the anticyclonic regime and releases it during the cyclonic regime (Proshutinsky et al. 2002; Häkkinen and Proshutinsky 2004). However, to fully understand the causes for the FWC increase during 2004–09, a model simulation has to be conducted to estimate the freshwater budget and to understand the causes for FWC variations.

To understand interannual variations in FWC and SSH, particularly their sharp increases during 2004–09, we implemented a coupled ice–ocean model (CIOM) in the Arctic Ocean and conducted a 40-yr simulation. Section 2 describes the CIOM and the experimental design. Section 3 validates the model climate with available observations. Section 4 shows the simulated interannual variations of the FWC and SSH during 2004–09. Section 5 presents the conclusions. We show that most of the changes in FWC and SSH are due to surface wind stress curl during 2004–09 and that enhanced ice melting is a second contributor. In addition, the enhanced anticyclonic surface wind during 2004–09 can be related to the interannual variations of the Arctic polar vortex.

2. Model description

In this study, the CIOM consists of two components, the Princeton Ocean Model (POM) (Blumberg and Mellor 1987) and a multicategory ice model (Hibler 1979; Hibler 1980). This model system is based on the coupled ice–ocean models for the Labrador Sea (Yao et al. 2000) and the pan-Arctic Ocean (Wang et al. 2005). A detailed description of the CIOM is given by Yao et al. (2000) and Wang et al. (2005).

a. Ocean model

POM is a three-dimensional, primitive equation model with complete thermohaline dynamics, a sigma (σ) vertical coordinate, and a free surface. A second-order turbulence closure scheme (Mellor and Yamada 1982) is used to represent the mixed layer dynamics. The model domain and bathymetry derived from ETOP2 are shown in Fig. 1. To minimize pressure gradient errors, the bottom topography in the model was smoothed such that the difference in the depths of adjacent grid points divided by their means is less than 0.2 (Mellor et al. 1994). In the experiments described here, 23 vertical sigma levels are used with higher resolution in the upper mixed layer and lower resolution in the deep ocean. In the central Beaufort Sea where the depth is about 3500 m, the vertical resolution decreases from about 6 m for the upper seven layers to as much as 764 m in the deep ocean. The model grids are distributed on a rotated spherical surface with the North Pole at 8°N, 131.5°E and the equator indicated by the thin solid line in Fig. 1. The horizontal resolution is 0.29° × 0.25°, and the time step is 30 min.

Fig. 1.

CIOM bathymetry (m) and model domain: the thin solid line is the equator of the rotated spherical coordinate system, and the box indicated by the thick dashed line is the area where the variables are averaged to show their interannual variations for the Beaufort Sea area. Transect in Fig. 4 is the thick black line.

Fig. 1.

CIOM bathymetry (m) and model domain: the thin solid line is the equator of the rotated spherical coordinate system, and the box indicated by the thick dashed line is the area where the variables are averaged to show their interannual variations for the Beaufort Sea area. Transect in Fig. 4 is the thick black line.

Along the open boundaries, we use radiation boundary conditions for the baroclinic current, and volume transports are specified (Fig. 1). The inflow is prescribed as 0.8 Sv (1 Sv ≡ 106 m3 s−1) through Bering Strait, and the same amount is prescribed as outflow through the Canadian Archipelago. We also prescribe an inflow of 2 Sv Atlantic water into the Arctic Ocean via the Norwegian Sea, and an outflow of 2 Sv along the east coast of Greenland. Along the lateral boundary, the water temperature and salinity are relaxed to the Polar Science Center Hydrographic Climatology (PHC) within a 9-point buffer zone. In this study, the SSH gradient normal to the open boundary is zero.

b. Ice model

The sea ice component of the CIOM was developed at the Bedford Institute of Oceanography (Yao et al. 2000) using a thermodynamic model based on a multicategory ice thickness distribution function (Thorndike et al. 1975; Hibler 1980) and a viscous–plastic sea ice dynamics model (Hibler 1979). The model considers mechanical and thermodynamic redistributions of ice, ice ridging, and the formation of frazil ice. The equations describing these processes and the finite difference implementation are given by Tang et al. (2008). The ice thickness has seven categories (0.4, 0.8, 1.2, 2, 3, 5, and 7 m); heat and salt fluxes at the ice–ocean interface are governed by appropriate boundary processes, as discussed by Mellor and Kantha (1989). An empirical formulation for clear-sky incoming shortwave radiation by Shine and Crane (1984), and the cloud correction of Reed (1977) are used to estimate shortwave radiation. The longwave radiation is given by the Smith and Dobson (1984) formulation. Bulk formulations are used to estimate latent and sensible heat fluxes and the wind stress. Both air–ice and air–water drag coefficients are set to 1.3 × 10−3, while the ice–water drag coefficient is 5.5 × 10−3. Therefore, the composite water surface stress can be represented as

 
formula

where τw and τa are the stresses of ice to water and air to water, and A is the percentage of water surface covered with ice.

The sea ice albedo scheme is based on the formula suggested by Køltzow (2007):

 
formula
 
formula

and

 
formula

where αsea-ice and Ts represent the sea ice albedo and ice surface temperature, respectively. Here, the melt pond fraction is parameterized in terms of the surface temperature (°C). We assume that there is no melt pond if the ice surface temperature is below −2°C, and the melt pond fraction is given by

 
formula

The albedo of the melt pond is given by

 
formula

and the total albedo is given by

 
formula

c. Experiment design

For the experiments described here, boundary conditions for temperature and salinity are obtained from the PHC ocean database version 3.0 (PHC3.0) by Steele et al. (2001). River runoff climatology is prescribed along the Arctic coast (Prange and Lohmann 2004; Prange 2003). Neither restoring nor flux adjustment for sea surface temperature or sea surface salinity is applied. We use National Centers for Environmental Prediction (NCEP) reanalysis data to provide the CIOM with surface air temperature, surface specific humidity, precipitation rate, total cloud cover, sea level pressure (SLP), and 10-m winds.

First, the initial conditions for both ocean and sea ice came from a run that we used to tune the model. Therefore, to investigate the model drift, the average climatology from the NCEP data from 1979 to 2008 is used to drive the CIOM for 20 years, as shown in Fig. 2. These results show that the model estimates for the domain-averaged FWC and total ice volume gradually decrease, as the SST increases. In fact, the CIOM almost reaches its overall equilibrium state after nine years of integration (Fig. 2). Given that the initial conditions are derived from the tuning run, where the CIOM essentially reaches its equilibrium state, it is not surprising to see that the simulation reaches stability relatively quickly.

Fig. 2.

Time series (domain averaged) from the CIOM simulation forced with average NCEP daily climatology fields from 1979 to 2008, showing (top) FWC (m), (middle) SST (°C), and (bottom) total ice volume (km3).

Fig. 2.

Time series (domain averaged) from the CIOM simulation forced with average NCEP daily climatology fields from 1979 to 2008, showing (top) FWC (m), (middle) SST (°C), and (bottom) total ice volume (km3).

To study the interannual variation of FWC and SSH, we integrate the CIOM for 40 years from 1970 to 2009 with the same initial conditions, using daily NCEP reanalysis data as drivers for these years. While the NCEP reanalysis data are available back to 1948, the data from the satellite era beginning in 1979 have relatively higher quality (Serreze and Barrett 2011). The first 9 years of the run are discarded as spinup of the model. Thus, we focus on the CIOM’s performance and outputs for the 31 years from 1979 to 2009 for the remainder of this study.

3. Model climatology

a. Sea ice

In the Arctic, sea ice is a very important component of the climate system. It influences the salinity through brine rejection when ice forms. It causes surface freshening when ice melts. Observations show that the total sea ice area has steadily decreased since the 1990s, reaching a record low in 2007 (Wang et al. 2009). In Figs. 3a and 3b, we compare the model simulations of ice concentration contours and ice extent (defined as the area with at least 15% concentration) for April and September to corresponding estimates from the Hadley Centre data. The data is archived at the Met Office Hadley Centre and can be accessed online (http://hadobs.metoffice.com/hadisst/; Rayner et al. 2003). Overall, the observed ice concentration contours and ice extent in the Arctic in both summer and winter are well captured by the CIOM (Figs. 3a and 3b). Moreover, compared with the time series of Hadley data for the total sea ice extent shown in Fig. 3c, the CIOM has a reasonable ability to simulate the observed negative trend of ice extent as well as its interannual variations. Although the overall correlation between the Hadley data and the CIOM simulation is about 0.9 in Fig. 3c, a caveat is that the CIOM overestimates the ice extent on the Atlantic side (Figs. 3a and 3b). On average, the ice extent simulated by the CIOM is comparable with the Hadley Centre data after 2000 and overestimates the observed ice extent by about 14% in the summers before 1999 (Fig. 3c).

Fig. 3.

Ice concentration fraction as simulated by CIOM (shaded and solid lines) compared to Hadley Centre data (dashed lines) in contour increments of 0.15 in (a: top) April and (b: middle) September. (c: bottom) Total ice extent in September from CIOM simulation (dashed) and Hadley Centre data (solid) (106 km2). The ice edge is defined as the 15% ice concentration contour, and ice extent is the area with at least 15% ice concentration.

Fig. 3.

Ice concentration fraction as simulated by CIOM (shaded and solid lines) compared to Hadley Centre data (dashed lines) in contour increments of 0.15 in (a: top) April and (b: middle) September. (c: bottom) Total ice extent in September from CIOM simulation (dashed) and Hadley Centre data (solid) (106 km2). The ice edge is defined as the 15% ice concentration contour, and ice extent is the area with at least 15% ice concentration.

b. Water temperature and salinity

Simulated temperature and salinity averaged during 1979–2009 are compared with the PHC3.0 climatology in Fig. 4. In winter, the surface wind stress drives the Arctic Ocean anticyclonically, transporting surface freshwater to the Beaufort Sea through Ekman convergence (Proshutinsky and Johnson 1997). The CIOM simulation shows a salinity minimum above 200 m in the Beaufort Sea (Fig. 4a). Although this minimum is similar to the observations seen in Fig. 4c in the Beaufort Sea, it is underestimated. Moreover, both the PHC data and the model simulation show a warm Atlantic water layer (AWL) between 200 and 900 m (Figs. 4b and 4d). Related studies suggest that the AWL is transported into the central Arctic Ocean through Fram Strait and the Barents Sea (Golubeva and Platov 2007), which represents an important heat and salt source for the Arctic Ocean. However, it is clear that there are some differences in the details between the model simulations and the PHC climatology. For example, compared with estimates from the PHC climatology, the CIOM tends to slightly underestimate the maximum temperature between 200 and 900 m (Figs. 4b and 4d), which could be related to weak vertical mixing in the region (Zhang and Steele 2007) and smaller air–ice drag coefficients, as shown in section 4e.

Fig. 4.

Cross section for September as a function of depth (m) along the transect indicated in Fig. 1, showing (a: top left) simulated salinity, (b: top right) simulated temperature, (c: bottom left) PHC salinity, and (d: bottom right) PHC temperature. Units of x axis are × 103 km.

Fig. 4.

Cross section for September as a function of depth (m) along the transect indicated in Fig. 1, showing (a: top left) simulated salinity, (b: top right) simulated temperature, (c: bottom left) PHC salinity, and (d: bottom right) PHC temperature. Units of x axis are × 103 km.

c. Freshwater content

In this study, the FWC is calculated using the formula , if S < 34.8 psu, where S is salinity, z is water depth (m), and L is the uppermost level where S reaches 34.8 psu. Following Proshutinsky et al. (2009), the FWC calculations were carried out for the grid points with total depths greater than 300 m (Fig. 5). In the central Beaufort Sea, both the model simulation and the PHC data show a relative maximum FWC at 76°N, 150°W. However, compared to the PHC data, the CIOM results slightly underestimate the extent of FWC, which has values that are greater than 16 m, and overestimate a maximum near 76°N, 150°W.

Fig. 5.

Comparison of annual FWC (m) (left) simulated by CIOM and (right) estimated from PHC3.0 salinity. The reference salinity used to compute FWC is 34.8 psu, and FWC is estimated for the level with the salinity below 34.8.

Fig. 5.

Comparison of annual FWC (m) (left) simulated by CIOM and (right) estimated from PHC3.0 salinity. The reference salinity used to compute FWC is 34.8 psu, and FWC is estimated for the level with the salinity below 34.8.

4. Interannual variations of FWC and SSH

a. Freshwater content

The simulated FWC distributions for the period from 2003 to 2008 are compared with the observations (Proshutinsky et al. 2009) in Figs. 6a and 6b so as to understand interannual FWC variations in the central Beaufort Sea. Observations suggest that there was a significant increase in FWC in the central Beaufort Sea during 2003–08. The observed maximum FWC increase was about 2 m from 2003 to 2008 and reached a record high in 2008 (Fig. 6b). A similar change in FWC can be seen in the simulated results shown in Fig. 6a. During 2003–08, the CIOM simulated about a 6-m increase of FWC in the central Beaufort Sea. In addition, the CIOM simulates well the main features seen in the spatial and temporal variations in the observations of Fig. 6b. Both observations and model simulations suggest a westward extension of the maximum FWC center during 2003–06 and the southeast extension in 2007 and 2008. However, the CIOM significantly underestimates the extent of the FWC increase. The causes for the underestimation will be discussed in section 4f through sensitivity studies.

Fig. 6a.

Simulated FWC (m) annual average: the box indicated by dashed lines is the area where variables are averaged for the Beaufort Sea area, as in Fig. 1: (top left) 2003, (top right) 2004, (middle left) 2005, (middle right) 2006, (bottom left) 2007, and (bottom right) 2008.

Fig. 6a.

Simulated FWC (m) annual average: the box indicated by dashed lines is the area where variables are averaged for the Beaufort Sea area, as in Fig. 1: (top left) 2003, (top right) 2004, (middle left) 2005, (middle right) 2006, (bottom left) 2007, and (bottom right) 2008.

Fig. 6b.

As in Fig. 6a, but for observations from Proshutinsky et al. (2009).

Fig. 6b.

As in Fig. 6a, but for observations from Proshutinsky et al. (2009).

Figure 7 shows the time series of the simulated FWC averaged over the Beaufort region, shown in the boxed area in Figs. 1 and 6. Although the Beaufort gyre is partially outside of the box, it covers most of the Beaufort gyre, and the FWC change during 2004–09 is not sensitive to its exact position. Overall, FWC has increased significantly since 1997. The rate of increase appears to have accelerated during 2004–09. We estimate that there has been an increase of almost 2.5 m in FWC between 1997 and 2009 after a decrease of about 1 m during the previous decade, which is about 16% of the total FWC in the region. However, the CIOM simulation significantly underestimated the increase in FWC during 2004–09. The underestimation is mainly related to the smaller air–ice drag coefficient and weak vertical mixing; other factors, such as freshwater transport along the boundary and river runoff, were not found to represent dominant factors in our simulations. Therefore, we need to be somewhat careful in estimating changes in simulated FWC values. Proshutinsky et al. (2009) suggested that a major cause for the FWC increase is associated with the changes in Ekman pumping generated by the anticyclonic atmospheric circulation over the Canada Basin. During the last decade, the magnitude of the observed wind stress curl increased significantly, which could contribute to the increase in Ekman pumping and, consequently, in the freshwater accumulation in the central Beaufort Sea (Proshutinsky et al. 2009).

Fig. 7.

(top) Simulated annual FWC (m) and (bottom) SSH (cm) averaged over the box indicated by dashed lines in Fig. 1. The dashed line in the top panel represents observations from Proshutinsky et al. (2009), and the solid line is the CIOM simulation.

Fig. 7.

(top) Simulated annual FWC (m) and (bottom) SSH (cm) averaged over the box indicated by dashed lines in Fig. 1. The dashed line in the top panel represents observations from Proshutinsky et al. (2009), and the solid line is the CIOM simulation.

b. Sea surface height

Sea surface height is the cumulative result of all dynamic and thermodynamic, terrestrial, oceanic, atmospheric, and cryospheric processes and is an important indicator for Arctic climate change (Proshutinsky et al. 2002). Figure 8a shows that the CIOM simulation suggests a SSH maximum in the central Beaufort Sea associated with the anticyclonic Beaufort gyre (Fig. 8b) and a cyclonic circulation located in the North Atlantic region. These two features are consistent with the SSH in the Pacific side being higher than in the Atlantic side (Fig. 8a). The simulated SSH maximum in the Beaufort Sea is also consistent with the results shown in the dynamic surface topography fields (McPhee et al. 2009) and the steric sea level (Steele and Ermold 2007). In the Arctic, SSH is largely determined by salinity (Steele and Ermold 2007), and it is not surprising to see a SSH maximum in the Beaufort Sea (Fig. 8a), given the relatively low salinity in the region.

Fig. 8.

(a; top) Average annual SSH (m) and (b: bottom) average surface current (m s−1) during 1979–2009.

Fig. 8.

(a; top) Average annual SSH (m) and (b: bottom) average surface current (m s−1) during 1979–2009.

Moreover, averaged over the Beaufort Sea, SSH shows a similar interannual variation as FWC (Fig. 7), with a correlation coefficient of 0.97. Although the averaged SSH steadily decreased from 17 to 15 cm during 1979–94, it increased by about 8 cm during the last decade in the simulation and reached a maximum in 2008 (Fig. 6b). In terms of horizontal distributions, most of the SSH increase is located near the center of the Beaufort gyre (Fig. 9). Compared to the average annual SSH between 1979 and 2009 (Fig. 8a), the SSH increased by about 10 cm in the central Beaufort gyre in 2008. Moreover, the data from observations also indicate an increase of about 75% in steric sea level in the Beaufort Sea and a shift of several hundred kilometers to the southeast of the maximum (McPhee et al. 2009).

Fig. 9.

Annual SSH (m) in 2008: the box indicated by dashed lines is the area where variables are averaged for the Beaufort Sea area, as shown in Fig. 1.

Fig. 9.

Annual SSH (m) in 2008: the box indicated by dashed lines is the area where variables are averaged for the Beaufort Sea area, as shown in Fig. 1.

c. Estimates of freshwater sources in the Beaufort Sea

In this section, we discuss the possible freshwater sources in the Beaufort Sea. The source terms for the FWC in the Beaufort Sea include local river runoff (R), precipitation (P), surface evaporation (E), ice formation and melting (M), and horizontal advection associated with Ekman pumping (EP). Therefore, the budget equation can be written as

 
formula

Here EP is estimated as , where A is the horizontal area, the integration over z is from L, the depth at which S = 34.8, to the surface, where S is below 34.8 psu, V is current, and S is water salinity. Local river runoff climatology is prescribed along the Arctic coast and does not contribute to the simulated FWC interannual variations. Compared to M and EP, the contribution of PE is very small (Fig. 10a). Freshwater from ice formation and melting decreased from the maximum 0.4 m yr−1 in the early 1980s to −0.4 m yr−1 in the early 1990s. During 2004–09, ice melting recovered somewhat, contributing about 0.3 m yr−1 to the FWC in the Beaufort Sea (Fig. 10a). The biggest FWC contribution is from horizontal advection associated with Ekman pumping transport (Fig. 10a). On average, EP contributed about 0.63 m yr−1 to the FWC reservoir in the Beaufort Sea for the period 2004–09.

Fig. 10.

(a: top) Anomalies (relative to long-term means) of freshwater flux (m yr−1) into the box, indicated by dashed lines in Fig. 1, due to surface moisture flux (PE) (thick solid line), ice formation and melting (M) (dashed line), and Ekman pumping (EP) (thin solid line); (b: bottom) average wind stress curl [10−5 kg(m2 s−2)−1] over Beaufort Gyre region where positive value represents anticyclonic surface forcing. Anomalies in the top panel are computed relative to the 1979–2009 mean.

Fig. 10.

(a: top) Anomalies (relative to long-term means) of freshwater flux (m yr−1) into the box, indicated by dashed lines in Fig. 1, due to surface moisture flux (PE) (thick solid line), ice formation and melting (M) (dashed line), and Ekman pumping (EP) (thin solid line); (b: bottom) average wind stress curl [10−5 kg(m2 s−2)−1] over Beaufort Gyre region where positive value represents anticyclonic surface forcing. Anomalies in the top panel are computed relative to the 1979–2009 mean.

d. Impacts of surface wind stress

Based on the above estimates, it is clear that Ekman pumping contributes most of the FWC increase during 2004–09, and about one-third of the FWC increase is related to the enhanced ice melting in summer. Proshutinsky et al. (2009) speculated that the interannual variations in wind stress could be one of the main causes for the FWC increase in the 2000s. Through this process, the increased anticyclonic surface forcing in the 2000s enabled the accumulation of freshwater in the Beaufort Sea through Ekman pumping. The surface wind stress curl over the Beaufort gyre region gradually increased during the 1980s. It shows a strong interannual variation in the early 1990s and a peak in 2007, as shown in the time series in Fig. 10b. Meanwhile, Ekman transport shows a significant increase during 2004–09, but it has no obvious change during 1979–2003. The correlation between the wind stress curl (Fig. 10b) and Ekman pumping transport (black line in Fig. 10a) is about 0.6, suggesting significant impacts of surface wind stress on the horizontal advection of the FWC. Therefore, at least in this model, the enhanced cyclonic wind is the main driving force for the interannual variation of FWC during 2004–09. However, in 1980–87 and 1994–97, Ekman transport and ice melting and formation contributed similar changes to the FWC in the Beaufort Sea.

To further understand the impacts of surface wind stress on the interannual variations of FWC, we conducted a sensitivity study for 2007, when the maximum FWC increase occurred (Fig. 7a). In this sensitivity experiment, we replace the wind annual average with its climatology, and we use realistic daily variability. The annual wind in 2007 is more anticyclonic than the annual climatological wind, and the surface wind used in the sensitivity study has a smaller annual average than that used for the control run, described in section 2. Comparisons between this experiment and the control run enable us to understand the impacts of reduced surface wind curl on the FWC, as shown in Fig. 11. In December 2007, the simulation with reduced wind curl has about 1 m less FWC in the central Beaufort Sea than the simulation with the realistic wind. The magnitude is consistent with the estimated Ekman pumping transport (Fig. 10a). In addition, there is no significant difference between the ice melting and formation in the two experiments; essentially all the reduction in FWC by 1 m is due to changes in the Ekman transport.

Fig. 11.

Freshwater content (m) in December 2007: (top) the control run and (bottom) FWC when average annual wind is replaced by climatology, which is more cyclonic than the reanalysis winds used in 2007.

Fig. 11.

Freshwater content (m) in December 2007: (top) the control run and (bottom) FWC when average annual wind is replaced by climatology, which is more cyclonic than the reanalysis winds used in 2007.

e. Impacts of polar vortex at 500 hPa

Here, we will examine the link between the sharp increase of FWC in the Beaufort Sea and the interannual variation of the 500-hPa Arctic polar vortex, which is a prominent atmospheric feature (Fig. 12a). The polar vortex is a persistent large-scale cyclonic circulation with a center near the pole, a trough over the east coast of each continent, and a ridge over western North America. The polar vortex has an impact on surface winds over the western Arctic by affecting the strength of the Beaufort Sea high (BSH). For example, an amplified western North American ridge at 500 hPa is often accompanied by a strong BSH (Serreze and Barrett 2011).

Fig. 12.

Mean 500-hPa geopotential height fields (10 m) in April–September, averaged from (a: top) 1979 to 2009 and (b: bottom) 2004 to 2009.

Fig. 12.

Mean 500-hPa geopotential height fields (10 m) in April–September, averaged from (a: top) 1979 to 2009 and (b: bottom) 2004 to 2009.

Figure 13 shows the second rotated EOF mode of geopotential height at 500 hPa for April–September. The data used to derive the rotated EOF modes are the monthly NCEP–National Center for Atmospheric Research (NCAR) reanalysis data from 1979 to 2009. Comparison between Fig. 12a and 13a suggests that the loading pattern of the second mode mainly represents the variability of the polar vortex. It has a negative center near the North Pole and Greenland, while several weak positive centers can be seen at middle latitudes (Fig. 13a). In Fig. 13b, the time series of the second mode shows a significant weakening of the polar vortex during 2004–09. The green line in Fig. 13b shows the annual change of FWC, which is defined as the FWC difference between the current year and the previous year; it is well correlated with the time series of the second rotated EOF mode after 1997 (Fig. 13b), suggesting that anomalous atmospheric circulation patterns at 500 hPa have impacts on the FWC. In Fig. 13b, the biggest FWC increase occurred in 2008 when the polar vortex reached a record weak point. Moreover, the weakening of the polar vortex can also be directly seen in Fig. 12b.

Fig. 13.

Loading pattern (a: top) of the second rotated EOF mode in geopotential height at 500 hPa during April–September. The time series (b: bottom) is of the second rotated EOF mode (red) and annual change of FWC (m) averaged over the box in Fig. 1 (green).

Fig. 13.

Loading pattern (a: top) of the second rotated EOF mode in geopotential height at 500 hPa during April–September. The time series (b: bottom) is of the second rotated EOF mode (red) and annual change of FWC (m) averaged over the box in Fig. 1 (green).

Consistent with the weakening of the polar vortex, there was a strong BSH during the period 2004–09. Figure 14 shows the SLP anomalies for April–September relative to the mean SLP for the period 1979–2009. Similar to the spatial pattern of the second rotated EOF mode, there are two positive SLP centers, one over Greenland and the other over the Beaufort Sea, suggesting the impacts of an anomalous polar vortex. However, there are no significant SLP anomalies over the Beaufort Sea in winter and autumn during 2004–09 (figures not shown).

Fig. 14.

April–September SLP anomalies (hPa) averaged during 2004–09: anomalies are computed with respect to the mean SLP for the period 1979–2009.

Fig. 14.

April–September SLP anomalies (hPa) averaged during 2004–09: anomalies are computed with respect to the mean SLP for the period 1979–2009.

It is now clear that the interannual variation of the polar vortex is the main cause for the sharp increase of FWC after 2003. A weak polar vortex causes strengthening of the BSH by affecting the frequency of the synoptic weather systems in the Beaufort Sea (Serreze and Barrett 2011). The anticyclonic surface wind associated with a strong BSH further increases the freshwater convergence by resulting in enhanced Ekman transport in the region.

f. Discussion

There are large uncertainties in the air–ice drag coefficient and the sea ice compressive strength parameter (P*) used in model studies. Observed values for the air–ice drag coefficient range from 0.5 × 10−3 to 5 × 10−3, depending on the atmospheric boundary layer stability and surface roughness (Prinsenberg and Peterson 2002). Moreover, the optimal value for P* depends on the choice of air–ice drag coefficient; also, uncertainties in the air–ice drag coefficient are a major source of the uncertainties in P* (Tremblay and Hakakian 2006). For example, a value of 1.5 × 104 N m−2 was used by Kreyscher et al. (1997), whereas Hibler and Walsh (1982) use 2.75 × 104 N m−2 in their ice models with the standard viscous–plastic rheology of Hibler (1979). In the experiment described in section 2, we assume the air–ice drag coefficient is 1.3 × 10−3 and P* is taken as 1.5 × 104 N m−2.

To understand the impacts of air–ice drag coefficient on the simulated FWC, we reran the CIOM simulation for 2000–09 with the air–ice drag coefficient and P* set to 2.75 × 10−3 and 2.5 × 104 N m−2. The other conditions for this experiment are the same as those described in section 2. Compared to the FWC simulation results in Fig. 6a, increases in air–ice drag coefficient and P* can significantly increase the simulated estimates for the FWC (Fig. 15). Although our results still somewhat underestimate the observed FWC in the central Beaufort Sea (Fig. 6b), our simulation with an increased air–ice drag coefficient and P* significantly improved our estimates for the FWC.

Fig. 15.

Simulated FWC (m) annual average with increased air–ice drag coeffcient and P*. Plots in each panel as in Figs. 6a,b.

Fig. 15.

Simulated FWC (m) annual average with increased air–ice drag coeffcient and P*. Plots in each panel as in Figs. 6a,b.

We also obtain changes in the FWC in the Beaufort Sea by varying the vertical mixing, which alters the ocean’s stratification (Zhang and Steele 2007). In sensitivity experiments to understand the impacts of vertical mixing on the FWC simulation, we double the surface vertical mixing. Thus, we find that the increased vertical mixing increases the FWC in the Beaufort Sea from 2003 to 2008 (Fig. 16), which is consistent with the studies by Zhang and Steele (2007). For the range of values tested, these results suggest that FWC increases are more responsive to changes in the air–ice drag coefficient than in vertical mixing.

Fig. 16.

FWC (m) annual average resulting from increased surface mixing, plotted as in Figs. 6a,b.

Fig. 16.

FWC (m) annual average resulting from increased surface mixing, plotted as in Figs. 6a,b.

5. Conclusions

In this paper, we studied the interannual variation of FWC and SSH in the central Beaufort Sea using a coupled ice–ocean model (CIOM). Compared with the observations, the CIOM reasonably reproduces the ice extent, salinity, water temperature, and FWC in the central Beaufort Sea. Moreover, the CIOM also demonstrates a reasonable ability to reproduce the observed negative trend in ice extent as well as the interannual variations over recent years. In addition, the model simulates a minimum in salinity, and a FWC maximum in the Beaufort Sea, and a warm Atlantic water layer within the Arctic Ocean, which is similar to the PHC climatology.

The CIOM simulation captures the rapid increases of FWC and SSH during 2004–09. The simulated SSH increase is about 8 cm, while the FWC increase is about 2.5 m, with most of the increases occurring in the center of the Beaufort gyre. However, as the CIOM significantly underestimates the extent of the FWC increase during 2004–09, as shown in Fig. 6, the FWC increase in the real world is somewhat different. This study gives us some indications of the relative importance of each process in the FWC increase during 2004–09. The sensitivity studies suggest that the causes for underestimation of FWC are related to the smaller air–ice drag coefficient and weak vertical mixing. Based on model simulations, Zhang and Steele (2007) also suggest that the freshwater content in the central Beaufort Sea is closely related to the vertical mixing. Weak upper layer mixing usually results in an underestimation of FWC in the region. Fixed volume transport along the lateral boundaries and river runoff are not seen as the dominant processes.

The main sources for the FWC increase are associated with the increased Ekman pumping transport and enhanced ice melting. On average, the Ekman pumping transport contributes about 0.63 m yr−1 FWC to the Beaufort Sea during 2004–09, while the increase due to ice melting is about 0.3 m yr−1 FWC. A sensitivity study was performed that further shows the significant impacts of surface wind stress on the changes in FWC.

Finally, the enhanced surface wind during 2004–09 is related to the interannual variation of the 500-hPa Arctic polar vortex during the spring and summer. The second rotated EOF of geopotential height at 500 hPa shows a significant weakening during 2004–09—in particular, over the western Arctic, which increased the Beaufort Sea high (BSH) by enhancing the frequency of synoptic weather systems in the region. Furthermore, the anticyclonic wind associated with a strong BSH increased the freshwater transport into the western Arctic and thus resulted in increased FWC in the Beaufort Sea region.

Acknowledgments

We want to thank the anonymous reviewers for their insightful comments, which helped us improve our manuscript. Support for this research comes from the Federal International Polar Year (IPY) Office of Canada. JW also appreciates the support received from the NOAA Russian–American Long-Term Census of the Arctic (RUSALCA) IPY modeling project.

REFERENCES

REFERENCES
Aagaard
,
K.
, and
E. C.
Carmack
,
1989
:
The role of sea ice and fresh water in the Arctic circulation
.
J. Geophys. Res.
,
94
,
14 485
14 498
.
Blumberg
,
A. F.
, and
G. L.
Mellor
,
1987
:
A description of a three-dimensional coastal ocean circulation model
.
Three-Dimensional Coastal Ocean Models, N. S. Heaps, Ed., Coastal and Estuarine Sciences, Vol. 4, Amer. Geophys. Union, 1–16
.
Curry
,
R.
, and
C.
Mauritzen
,
2005
:
Dilution of the northern North Atlantic Ocean in recent decades
.
Science
,
308
,
1772
1774
,
doi:10.1126/science.1109477
.
Dickson
,
R. R.
,
J.
Meincke
,
S.-A.
Malmberg
, and
A. J.
Lee
,
1988
:
The “Great Salinity Anomaly” in the northern North Atlantic 1968-1982
.
Prog. Oceanogr.
,
20
,
103
151
.
Golubeva
,
E. N.
, and
G. A.
Platov
,
2007
:
On improving the simulation of Atlantic Water circulation in the Arctic Ocean
.
J. Geophys. Res.
,
112
,
C04S05
,
doi:10.1029/2006JC003734
.
Häkkinen
,
S.
, and
A.
Proshutinsky
,
2004
:
Freshwater content variability in the Arctic Ocean
.
J. Geophys. Res.
,
109
,
C03051
,
doi:10.1029/2003JC001940
.
Hibler
,
W. D.
, III
,
1979
:
A dynamic thermodynamic sea ice model
.
J. Phys. Oceanogr.
,
9
,
815
864
.
Hibler
,
W. D.
, III
,
1980
:
Modeling a variable thickness sea ice cover
.
Mon. Wea. Rev.
,
108
,
1943
1973
.
Hibler
,
W. D.
, III
, and
J. E.
Walsh
,
1982
:
On modeling seasonal and interannual fluctuations of Arctic sea ice
.
J. Phys. Oceanogr.
,
12
,
1514
1523
.
Køltzow
,
M.
,
2007
:
The effect of a new snow and sea ice albedo scheme on regional climate model simulations
.
J. Geophys. Res.
,
112
,
D07110
,
doi:10.1029/2006JD007693
.
Kreyscher
,
M.
,
M.
Harder
, and
P.
Lemke
,
1997
:
First results of the Sea Ice Model Intercomparison Project (SIMIP)
.
Ann. Glaciol.
,
25
,
8
11
.
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
,
doi:10.1029/2010GL044678
.
McPhee
,
M. G.
,
A.
Proshutinsky
,
J. H.
Morison
,
M.
Steele
, and
M. B.
Alkire
,
2009
:
Rapid change in freshwater content of the Arctic Ocean
.
Geophys. Res. Lett.
,
36
,
L10602
,
doi:10.1029/2009GL037525
.
Mellor
,
G. L.
, and
T.
Yamada
,
1982
:
Development of a turbulent closure model for geophysical fluid problems
.
Rev. Geophys.
,
20
,
851
875
.
Mellor
,
G. L.
, and
L. H.
Kantha
,
1989
:
An ice-ocean coupled model
.
J. Geophys. Res.
,
94
,
10 937
10 954
.
Mellor
,
G. L.
,
T.
Ezer
, and
L.-Y.
Oey
,
1994
:
The pressure gradient conundrum of sigma coordinate ocean models
.
J. Atmos. Oceanic Technol.
,
11
,
1126
1134
.
Prange
,
M.
,
2003
:
Influence of Arctic freshwater sources on the circulation in the arctic Mediterranean and the North Atlantic in a prognostic ocean/sea-ice model
.
Ph.D. thesis, University of Bremen, 252 pp
.
Prange
,
M.
, and
G.
Lohmann
,
2004
:
Variable freshwater input to the Arctic Ocean during the Holocene: Implications for large-scale ocean-sea ice dynamics as simulated by a circulation model
.
The Climate in Historical Times: Towards a Synthesis of Holocene Proxy Data and Climate Models, H. Fischer et al., Eds., GKSS School of Environmental Research, Springer-Verlag, 319–336
.
Prinsenberg
,
S.
, and
I. K.
Peterson
,
2002
:
Variations in air-ice drag coefficient due to ice surface roughness
.
Int. J. Offshore Polar Eng.
,
12
,
121
125
.
Proshutinsky
,
A.
, and
M. A.
Johnson
,
1997
:
Two circulation regimes of the wind-driven Arctic Ocean
.
J. Geophys. Res.
,
102
,
12 493
12 514
.
Proshutinsky
,
A.
,
R. H.
Bourke
,
F. A.
McLaughlin
,
2002
:
The role of the Beaufort gyre in Arctic climatic variability: Seasonal to decadal climate scales
.
Geophys. Res. Lett.
,
29
,
2100
,
doi:10.1029/2002GL015847
.
Proshutinsky
,
A.
, and
Coauthors
,
2009
:
Beaufort gyre freshwater reservoir: State and variability from observations
.
J. Geophys. Res.
,
114
,
C00A10
,
doi:10.1029/2008JC005104
.
Rayner
,
N. A.
,
D. E.
Parker
,
E. B.
Horton
,
C. K.
Folland
,
L. V.
Alexander
,
D. P.
Rowell
,
E. C.
Kent
, and
A.
Kaplan
,
2003
:
Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century
.
J. Geophys. Res.
,
108
,
4407
,
doi:10.1029/2002JD002670
.
Reed
,
R. K.
,
1977
:
On estimating insolation over the ocean
.
J. Phys. Oceanogr.
,
7
,
482
485
.
Serreze
,
M. C.
, and
A. P.
Barrett
,
2011
:
Characteristics of the Beaufort Sea high
.
J. Climate
,
24
,
159
182
.
Shine
,
K. P.
, and
R. G.
Crane
,
1984
:
The sensitivity of a one-dimensional thermodynamic sea ice model to changes in cloudiness
.
J. Geophys. Res.
,
89
,
10 615
10 622
.
Smith
,
S. D.
, and
F. W.
Dobson
,
1984
:
The heat budget at ocean weather station Bravo
.
Atmos.–Ocean
,
22
,
1
22
.
Steele
,
M.
, and
W.
Ermold
,
2007
:
Steric sea level change in the northern seas
.
J. Climate
,
20
,
403
417
.
Steele
,
M.
,
R.
Morley
, and
W.
Ermold
,
2001
:
PHC: A global ocean hydrography with a high-quality Arctic Ocean
.
J. Climate
,
14
,
2079
2087
.
Tang
,
C. L.
,
T.
Yao
,
W.
Perrie
,
B. M.
Detracey
,
B.
Toulany
,
E.
Dunlap
, and
Y.
Wu
,
2008
:
BIO ice-ocean and wave forecasting models and systems for eastern Canadian waters
.
Canadian Tech. Rep., Hydrography and Ocean Science 261, 65 pp
.
Thorndike
,
A. S.
,
D. A.
Rothrock
,
G. A.
Maykut
, and
R.
Colony
,
1975
:
The thickness distribution of sea ice
.
J. Geophys. Res.
,
80
,
4501
4513
.
Tremblay
,
L.-B.
, and
M.
Hakakian
,
2006
:
Estimating the sea ice compressive strength from satellite-derived sea ice drift and NCEP reanalysis data
.
J. Phys. Oceanogr.
,
36
,
2165
2172
.
Wang
,
J.
,
Q.
Liu
,
M.
Jin
,
M.
Ikeda
, and
F. J.
Saucier
,
2005
:
A coupled ice-ocean model in the pan-Arctic and North Atlantic Ocean: Simulation of seasonal cycles
.
J. Oceanogr.
,
61
,
213
233
.
Wang
,
J.
,
J.
Zhang
,
E.
Watanabe
,
M.
Ikeda
,
K.
Mizobata
,
J. E.
Walsh
,
X.
Bai
, and
B.
Wu
,
2009
:
Is the dipole anomaly a major driver to record lows in Arctic summer sea ice extent?
Geophys. Res. Lett.
,
36
,
L05706
,
doi:10.1029/2008GL036706
.
Yamamoto-Kawai
,
M.
,
F. A.
McLaughlin
,
E. C.
Carmack
,
S.
Nishino
,
K.
Shimada
, and
N.
Kurita
,
2009
:
Surface freshening of the Canada Basin, 2003–2007: River runoff versus sea ice meltwater
.
J. Geophys. Res.
,
114
,
C00A05
,
doi:10.1029/2008JC005000
.
Yao
,
T.
,
C. L.
Tang
, and
I. K.
Peterson
,
2000
:
Modeling the seasonal variation of sea ice in the Labrador Sea with a coupled multicategory ice model and the Princeton ocean model
.
J. Geophys. Res.
,
105
,
1153
1165
.
Zhang
,
J.
, and
M.
Steele
,
2007
:
Effect of vertical mixing on the Atlantic Water layer circulation in the Arctic Ocean
.
J. Geophys. Res.
,
112
,
C04S04
,
doi:10.1029/2006JC003732
.

Footnotes

*

Great Lakes Environmental Research Laboratory Contribution Number 1601.