Abstract

As the Arctic sea ice thins and ultimately disappears in a warming climate, its insulating power decreases. This causes the surface air temperature to approach the temperature of the relatively warm ocean water below the ice. The resulting increases in air temperature, water vapor, and cloudiness lead to an increase in the surface downwelling longwave radiation (DLR), which enables a further thinning of the ice. This positive ice–insulation feedback operates mainly in the autumn and winter. A climate change simulation with the Community Earth System Model shows that, averaged over the year, the increase in Arctic DLR is 3 times stronger than the increase in Arctic absorbed solar radiation at the surface. The warming of the surface air over the Arctic Ocean during fall and winter creates a strong thermal contrast with the colder surrounding continents. Sea level pressure falls over the Arctic Ocean, and the high-latitude circulation reorganizes into a shallow “winter monsoon.” The resulting increase in surface wind speed promotes stronger surface evaporation and higher humidity over portions of the Arctic Ocean, thus reinforcing the ice–insulation feedback.

1. Introduction

In recent decades, the Arctic surface air temperature has been rising nearly twice as fast as the global mean (e.g., Stroeve and Meier 2012; Serreze and Barry 2011; Screen and Simmonds 2010a). This “Arctic amplification” is a robust feature of climate simulations with enhanced CO2 (e.g., Manabe and Wetherald 1975, 1980; Hansen et al. 1984; Holland and Bitz 2003; Lu and Cai 2010). The Fifth Assessment Report of the Intergovernmental Panel on Climate Change projects that anthropogenic warming over the Arctic will continue as the Arctic sea ice decreases in extent and thickness (Bindoff et al. 2013).

The high albedo of the sea ice limits the amount of solar radiation absorbed by the surface. As Arctic snow and ice melt, the dark ocean absorbs more of the sun’s energy during the summer months (Fig. 1). This well-known positive ice–albedo feedback is an important cause of Arctic amplification (e.g., Manabe and Wetherald 1975; Hall 2004; Winton 2006; Serreze and Barry 2011). During winter, solar radiation is absent, and the main effect of the sea ice is to insulate the atmosphere from the relatively warm seawater below the ice. Decreases in ice thickness and extent reduce this insulating effect and so can lead to a warming of the air near the surface.

Fig. 1.

Schematic of the ice–albedo feedback, which is only active during the summer months.

Fig. 1.

Schematic of the ice–albedo feedback, which is only active during the summer months.

Hall (2004) found that the ice–albedo feedback accounts for only about half of the Arctic surface temperature increase associated with a CO2-induced warming; similar conclusions were reached by Graversen and Wang (2009) and Graversen et al. (2014). These authors inferred that changes in sea ice thickness and extent account for much of the remainder of the warming, consistent with earlier studies by Manabe and Stouffer (1980) and Robock (1980). Numerous studies have pointed to the importance of feedbacks involving water vapor (Ghatak and Miller 2013; Graversen and Wang 2009) and clouds (Winton 2006; Francis and Hunter 2007; Pithan and Mauritsen 2014; Schweiger et al. 2008; Vavrus 2004; Graversen and Wang 2009).

Leads in the ice expose the warm ocean water to the cold atmosphere and allow sensible heat and water vapor to enter the atmosphere, but, in the present climate, open water occupies only a small fraction of the Arctic Ocean during winter. As a result, longwave radiation is the main driver of the surface energy budget in winter. Recent work (e.g., Winton 2006; Graversen and Wang 2009) suggests that warming-induced increases in the surface downwelling longwave radiation (DLR) are even more important than the ice–albedo feedback.

Previous analyses of observations (e.g., Walsh and Chapman 1998; Francis and Hunter 2007; Gorodetskaya and Tremblay 2008; Lee et al. 2011), atmospheric reanalyses (e.g., Walsh and Chapman 1998; D.-S. Park et al. 2015, hereinafter DSP; H.-S. Park et al. 2015), and modeling studies (e.g., Zhang et al. 1996; Abbot et al. 2009; Graversen and Wang 2009) have shown the importance of increased DLR for Arctic amplification. Walsh and Chapman (1998) showed that, during the polar night, the surface air temperature over sea ice increases when the DLR increases. A recent observational study by DSP investigates the mechanisms that lead to increases in the DLR; later in this paper we compare our numerical results with their observational results.

Increases in the DLR can be partly due to increases in atmospheric water vapor, which in turn can be due to an increase of both local evaporation and transport from lower latitudes (e.g., Trenberth et al. 2005; Rinke et al. 2009; Jakobson and Vihma 2010; Kurita 2011; Di Biagio et al. 2012). As the sea ice retreats during spring and summer and slowly reforms during autumn, an enhanced latent heat flux from the ocean surface to the atmosphere leads to a moistening of the Arctic lower troposphere (Screen and Simmonds 2010b; Ghatak and Miller 2013). Kurita (2011) analyzed water isotopes as a proxy for the source region of water vapor and found a seasonally varying mixture of local and remote origins, with the local source dominating during late fall and early winter.

Simulations of climate change robustly project increases in the transport of water vapor into the Arctic, primarily during summer and autumn (Zhang et al. 2012; Bintanja and Selten 2014), largely because of the increased water vapor content of the warmer air (Held and Soden 2006; Bengtsson et al. 2011). Changes in the winds can also influence moisture transport into the Arctic, however, as discussed by Higgins and Cassano (2009), Skific et al. (2009a,b), and Skific and Francis (2013). Their results suggest that up to 25% of the future changes in moisture transport during summer into the Arctic can be attributed to changes in the circulation, rather than changes in atmospheric humidity. In the 2 × CO2 simulations of Vavrus et al. (2011a), increased Arctic water vapor results from both enhanced local evaporation and increased transport from lower latitudes.

The DLR can also increase as a result of changes in cloudiness. As is well known, clouds strongly influence the surface fluxes of both shortwave and longwave radiation, so changes in the clouds can influence the melting of snow and ice in the Arctic (Curry and Ebert 1992). An observational study by Kay and Gettelman (2009) suggests that, in the current climate, year-to-year variability in Arctic cloudiness during early fall is controlled by year-to-year variability in the large-scale atmospheric circulation and the position of the sea ice edge. Abbot and Tziperman (2008a,b), Abbot et al. (2009), Leibowicz et al. (2012), and Arnold et al. (2014) have proposed a positive wintertime feedback between convective clouds and Arctic sea ice loss. Walsh and Chapman (1998) and Gorodetskaya and Tremblay (2008) showed that clouds warm the surface by increasing the DLR, which can enhance sea ice melt and extend the melting season. Francis et al. (2005, p. 4) suggested that “the relationship between DLF (i.e., DLR) and ice edge position is also two way. Additional open water will likely be warmer than the ice it replaced, and thus be a stronger emitter of longwave radiation. Cloud bases would absorb this energy, warm, and emit more radiation toward the surface. This relationship is a positive feedback in the system.”

The purpose of the present study is to investigate feedbacks that favor Arctic amplification during the dark winter season.1 We use a transient climate change simulation to address the following questions:

  1. How do changes in the DLR interact with changes in the extent and thickness of the sea ice?

  2. How do changes in the large-scale atmospheric circulation interact with changes in the extent and thickness of the sea ice?

The remainder of this paper is organized as follows. Section 2 describes the climate model, experimental design, and reanalysis data used. Section 3 presents results from a simulation of future climate change, with an emphasis on local processes. Section 4 describes simulated changes in the circulation that are linked to Arctic amplification. The paper concludes with a summary of our findings.

2. Methods

a. CESM, version 1.1.1

To address the impacts of dark-season surface radiative processes in CO2-induced Arctic climate change, we performed a transient climate change simulation with version 1.1.1 of the Community Earth System Model (CESM), which uses CAM5 atmospheric physics. We chose the CESM for our study because it is a full-featured climate model that is freely available, and a great variety of climate simulation results produced with the CESM have been analyzed by many investigators. In particular, previous versions of the CESM have been extensively used for studies of Arctic processes and climate change and have been shown to perform quite well (e.g., Jahn et al. 2012; Vavrus et al. 2011b; Deser et al. 2010). A detailed description of the CESM is given by Hurrell et al. (2013). We used the finite-volume dynamical core with a 1.9° × 2.5° latitude–longitude grid and 30 levels. Microphysics was included using the available two-moment parameterization (Morrison and Gettelman 2008; Gettelman et al. 2008). The Rapid Radiative Transfer Model (Iacono et al. 2008) was used to calculate the radiative fluxes and heating rates. The Parallel Ocean Program, version 2 (POP2), ocean model and the Los Alamos Sea Ice Model, version 4 (CICE4), were run on the “gx1v6” grid, which has a nominal resolution of 1° (Smith et al. 2010). CICE4 includes elastic–viscous–plastic sea ice dynamics (Hunke and Dukowicz 1997) and energy-conserving thermodynamics (Bitz and Lipscomb 1999). It also includes the effects of multiple scattering of shortwave radiation, an explicit simulation of melt pond evolution, and the deposition and cycling of dust and black carbon (Hunke and Lipscomb 2008; Hurrell et al. 2013; Holland et al. 2012). CICE4 includes a simple melt pond parameterization that simulates the pond volume and area as functions of the surface meltwater flux (Hurrell et al. 2013). Aerosol deposition and cycling on sea ice are also included. These new capabilities allow for a more complete treatment of the surface albedo and shortwave radiative transfer in the ice and overlying snowpack (Holland et al. 2012).

The initial condition for our simulation was taken from an archived 500-yr spinup [CCSM4, 2° resolution preindustrial (PI) control, case b40.1850.track1.2deg.003]. For the first 25 simulated years, the CO2 concentration was held constant at its preindustrial value of 285 ppmv. This PI control was followed by 140 years of simulation, during which the CO2 concentration was increased by 1% yr−1 until it reached 4 times its preindustrial value (1139 ppmv). The CO2 concentration was then held fixed for an additional 170 simulated years. Unless otherwise stated, the analysis and results presented in this study with CESM are based on monthly averages over the last 25 years of the PI control and the 4 × CO2 simulations. Our analysis focuses on the Arctic poleward of 70°N during November–December–January (NDJ).

b. ERA-Interim

As discussed later, we have compared some of our model results with the European Centre for Medium-Range Weather Forecasts (ECMWF) interim reanalysis dataset (ERA-Interim; Dee et al. 2011), which has been extensively used to study the ongoing changes in the Arctic climate (e.g., Lindsay et al. 2014; Zygmuntowska et al. 2012; Serreze et al. 2012). ERA-Interim combines information from a multitude of observations (surface-based observations, upper-air soundings, and satellites) with information from a short forecast started from a previous analysis (Dee et al. 2011). This process takes into account the uncertainties associated with both the model and the various observations to create a best estimate of the historical state of the atmosphere. ERA-Interim uses four-dimensional variational data assimilation (4D-Var) and has a horizontal resolution of T255 spectral truncation (nominally 0.70°; Dee et al. 2011).

3. The ice–insulation feedback

Figure 2 presents the seasonally varying trends of the simulated Arctic means of selected fields. In these “seasonal trend plots,” the years of the simulation run from left to right along the horizontal axis, and the months of the year run from bottom to top along the vertical axis, starting with June and ending with May so that December is about in the middle. For convenience, the differences between the 4 × CO2 simulation and the PI control are plotted separately, as a kind of summary, on the right-hand side of each panel. Before plotting, each calendar month of data has been subjected to a 5-yr running mean to generate, for example, a 5-yr running mean of Januaries, and so on for each month of the year.

Fig. 2.

Arctic-mean seasonal trends based on simulations with CESM. The quantities shown are averaged from 70°N to the pole. The horizontal axis in each panel is time in years, and the vertical axis shows the season by month, starting with June at the bottom and ending with May at the top. In each panel, the first white vertical line shows the end of the 25-yr control and the beginning of the 1% yr−1 CO2 increase. The second vertical white line shows when 4 × CO2 has been reached.

Fig. 2.

Arctic-mean seasonal trends based on simulations with CESM. The quantities shown are averaged from 70°N to the pole. The horizontal axis in each panel is time in years, and the vertical axis shows the season by month, starting with June at the bottom and ending with May at the top. In each panel, the first white vertical line shows the end of the 25-yr control and the beginning of the 1% yr−1 CO2 increase. The second vertical white line shows when 4 × CO2 has been reached.

Figure 2a shows that, as expected from earlier work (e.g., Pithan and Mauritsen 2014; Serreze and Barry 2011; Screen and Simmonds 2010a; Holland and Bitz 2003), the simulated surface warming is largest in winter, reaching about 20 K in December. In contrast, the June surface air temperature increases by only about 5 K. Figure 2b provides an explanation for the strong seasonality of the surface temperature change, in terms of the insulating power of sea ice. During the early part of the simulation, the sea ice is thin and has gaps at the end of the summer but becomes thick and uniform by late autumn. The thick winter ice acts as an efficient thermal insulator that allows the near-surface air to be much colder than the freezing point of seawater, which is about 271 K. The summer sea ice extent and thickness decline rapidly during the first 50 years after CO2 begins to increase, and, by the time of CO2 quadrupling, the ice is extremely thin or absent altogether from July to December—half of the year. At 4 × CO2, the reduced sea ice extent and thickness imply a decrease in the insulating power of the ice. That is why the air above the Arctic Ocean has a temperature only slightly colder than 271 K, even in late autumn (Fig. 2a), when the area-averaged ice thickness is less than 20 cm (Fig. 2b). We return to this point later.

As the air warms, the total column water vapor increases, nearly doubling in winter (Fig. 2c). This is important because water vapor is a greenhouse gas, which absorbs and emits longwave radiation. Clouds are also longwave absorbers and emitters, and Fig. 2d illustrates that Arctic cloudiness increases during most seasons, especially during winter. The greatest increases in cloudiness are in the middle and upper troposphere (not shown) and are associated with an inflow of moisture from lower latitudes, consistent with results of Vavrus et al. (2011a).

We now examine the various components of the surface energy budget, averaged over the Arctic. Figures 2e and 2f show that the sensible and latent heat fluxes increase, especially during autumn and winter. These fluxes tend to warm and moisten the air and may also contribute to the increase in cloudiness. Figure 2g shows that the net longwave radiation at the surface (defined positive upward) also increases modestly during winter. The net longwave radiation is, of course, the relatively small difference between the strong upward and downward longwave radiation streams. The upwelling longwave radiation (ULR) increases because the surface itself is warmer, and the DLR increases because of warmer air temperatures, combined with increases in water vapor, cloudiness, and of course CO2. Interestingly, as the climate warms, the net surface longwave flux actually decreases during summer. The DLR and the ULR both increase during summer, but the DLR increases more, resulting in a decrease in the net longwave radiation. Figure 2h shows that the DLR increases by a very large 80 W m−2 during winter. This is several times larger than the increase in the surface latent heat flux. The spatial pattern of the increase in the wintertime surface DLR is shown in Fig. 3a. The increase is greatest over the Arctic Ocean, but it is also substantial over northern Europe and North America. Figure 3b shows that, during NDJ, about 40 W m−2 of the increase in DLR comes from the effects of clouds (computed as the difference between all-sky DLR and clear-sky DLR); the remainder is due to the clear-sky effects of temperature, water vapor, and CO2. In other words, the clouds account for less than half of the increase in the DLR during winter.

Fig. 3.

Polar stereographic map of change (4 × CO2 simulation minus PI control) in CESM wintertime (NDJ) (a) DLR, and (b) the contribution to (a) from clouds.

Fig. 3.

Polar stereographic map of change (4 × CO2 simulation minus PI control) in CESM wintertime (NDJ) (a) DLR, and (b) the contribution to (a) from clouds.

Of course, the ice–albedo effect also contributes to the Arctic warming. Figure 2i shows that the solar radiation absorbed by the surface increases by 60 W m−2 during summer. This is true even though summer cloudiness increases slightly, as shown in Fig. 2d. Note, however, that the absorbed solar radiation does not increase after about year 80, because by then the summer sea ice has disappeared.

The changes in the energy budget of the Arctic atmosphere are shown in Figs. 2j–l. By the end of the simulation, the net energy flux at the surface (defined positive downward) increases by 80 W m−2 during the summer months because of the drastically decreased surface albedo, but during winter the rate at which the surface loses energy increases by about 40 W m−2. This energy loss by the surface is, of course, an energy gain for the atmosphere.

At the top of the atmosphere (Fig. 2k) the lower surface albedo causes an increase in the net downward radiation in summer, but during winter there is little change. The combined effects of the changes in the net energy fluxes at the surface and the top of the atmosphere are shown in Fig. 2l. The figure shows that, throughout the simulation and at all times of year, the Arctic atmosphere is cooled by the combined effects of the diabatic fluxes at the surface and the top of the atmosphere. This, of course, implies that the Arctic receives a net inflow of energy from lower latitudes. As the climate warms, the net diabatic cooling of the atmosphere changes little in summer, but the diabatic cooling becomes weaker in winter by about 30 W m−2. As a result, the net atmospheric energy transport into the Arctic from lower latitudes also becomes weaker during winter in the warmer climate. The weaker meridional energy transport is consistent with the reduced pole-to-equator temperature gradient of the 4 × CO2 climate. In summary, the strong warming of the Arctic atmosphere is associated with weaker diabatic cooling, rather than increased energy transport into the Arctic by the atmospheric circulation.

To further explore the DLR increase shown in Fig. 2h, we define an effective atmospheric emissivity εs by

 
formula

where F↓sfc is the surface downwelling longwave radiation, σ is the Stefan–Boltzmann constant, and Ts is the surface air temperature. With this definition, the initial value for emissivity εs is 0.8. By linearizing the right-hand side of Eq. (1), we can separate the change in DLR into two parts:

 
formula

Here, we use a greek uppercase delta to denote differences between 4 × CO2 simulation and the PI control. The first term on the right-hand side of Eq. (2) is the contribution to ΔF↓sfc from the change in the emissivity, and the second term is the contribution from the change in Ts. Figures 4a and 4b show the spatial distributions of the two terms. Averaged over the polar cap (70°–90°N, and including all longitudes), the changes in the emissivity and temperature contribute 9.5 and 51.9 W m−2, respectively, to the increase in the wintertime DLR at 4 × CO2. The sum of these is 61.4 W m−2, 12.5 W m−2 less than the actual change in DLR of 73.9 W m−2. The discrepancy is due to the linearization used to obtain Eq. (2).

Fig. 4.

Polar stereographic maps of the wintertime (NDJ) contributions to ΔF↓sfc (W m−2) from the changes in the (a) emissivity and (b) temperature. The calculations are based on the linearization shown in Eq. (2). (c) The sum of (a) and (b). (d) The actual change in DLR (W m−2). The differences between (c) and (d) show the errors resulting from the linearization.

Fig. 4.

Polar stereographic maps of the wintertime (NDJ) contributions to ΔF↓sfc (W m−2) from the changes in the (a) emissivity and (b) temperature. The calculations are based on the linearization shown in Eq. (2). (c) The sum of (a) and (b). (d) The actual change in DLR (W m−2). The differences between (c) and (d) show the errors resulting from the linearization.

It is interesting that during the summer the increase in DLR is mostly due to an increase in the emissivity rather than the warming of the air. Again, the summer warming is relatively small.

As mentioned earlier, Fig. 2g shows that the cooling of the surface by the net surface longwave radiation increases slightly in the 4 × CO2 climate, relative to the PI control, because the strong increase in the DLR is slightly outweighed by an even stronger increase in the ULR. How should we think about cause and effect here? Suppose that an increase in the dark-season surface temperature, associated with a thinning of the ice, drove an increase in the ULR but that somehow the DLR did not change. The result would be a very rapid cooling of the surface (and thickening of the ice), which would be incompatible with the posited surface warming. This shows that the warmer surface temperature and decreased ice extent and thickness are made possible by the increased DLR. We cannot say that the warming and melting are caused by the increased DLR, but they could not happen without it. The thinning and/or disappearance of the ice leads directly to stronger ULR and warmer air temperatures. The warmer air and increased emissivity cause an increase in the DLR, which in turn makes it possible for the ice to become even thinner. In this way, the changes in ice thickness and DLR are mutually reinforcing and create a positive local thermodynamic feedback, which we call the ice–insulation feedback.

The DLR increase shown in Fig. 2h is substantial throughout the year and is strongest during winter. The increased solar radiation acts only during summer and has reached its maximum by about year 150 of our simulation, when the summer sea ice disappears. Figure 5 shows that, averaged over the entire year, the DLR increases by about 3 times as much as the absorbed solar radiation.

Fig. 5.

Time series of CESM-simulated annual mean change in surface DLR (red) and surface absorbed solar radiation (black) relative to the mean of the 25-yr control. These results are averaged over the polar cap, from 70° to 90°N.

Fig. 5.

Time series of CESM-simulated annual mean change in surface DLR (red) and surface absorbed solar radiation (black) relative to the mean of the 25-yr control. These results are averaged over the polar cap, from 70° to 90°N.

We close this section with Fig. 6, which shows in a very simple way that the zonally and annually averaged DLR increases much more strongly in the Arctic than anywhere else on Earth. This is partly because the Arctic warms more strongly than anywhere else on Earth. A second reason is that, in the Arctic, the water vapor increases from very small values in the PI climate. In lower latitudes, a substantial amount of water vapor is already present in the PI climate, so that adding even more water vapor has a relatively weak effect on the DLR.

Fig. 6.

Latitude–time plot of change (4 × CO2 simulation minus PI control) in CESM-simulated wintertime (NDJ) surface DLR (zonally averaged) relative to the mean of the 25-yr control.

Fig. 6.

Latitude–time plot of change (4 × CO2 simulation minus PI control) in CESM-simulated wintertime (NDJ) surface DLR (zonally averaged) relative to the mean of the 25-yr control.

4. The birth of a new monsoon

We now show that, during winter, the large-scale atmospheric circulation in the Arctic responds to the changing climate state in such a way as to reinforce the ice–insulation feedback discussed above. The circulation changes discussed below are a secondary effect. We do not claim that they cause or trigger the ice–insulation feedback described in section 3; if anything, the circulation changes (and the associated secondary feedback) are promoted by the ice–albedo feedback and the ice–insulation feedback.

The seasonal trend plot presented in Fig. 7a shows the evolution of the difference between the surface air temperature over the Arctic Ocean and the surface air temperature over the surrounding continents poleward of 60°N. We used the region poleward of 60°N (rather than 70°N) in order to incorporate more land points in the calculation. The figure shows that, at the start of the simulation, the ocean is colder than the land during spring and summer (March through September) and a few degrees warmer than the land during winter. By the end of the simulation, however, the air above the Arctic Ocean during December is more than 15 K warmer than the air above the surrounding land.

Fig. 7.

(a) As in Fig. 2, but for the CESM-simulated difference in surface air temperature between the Arctic Ocean (including areas covered by sea ice) and the surrounding land, from 60°N to the pole. Warmer colors indicate the ocean is warmer than the land, and cooler colors indicate the land is warmer than the ocean. (b) As in (a), but for the difference in surface pressure between the Arctic Ocean (including areas covered by sea ice) and the surrounding land, from 60°N to the pole. (c) The difference in surface air temperature between the Arctic Ocean (including areas covered by sea ice) and the surrounding land, from 60°N to the pole, as reported in ERA-Interim for the years 1979–2012. (d) As in (c), but for the difference in surface pressure between the Arctic Ocean (including areas covered by sea ice) and the surrounding land, from 60°N to the pole.

Fig. 7.

(a) As in Fig. 2, but for the CESM-simulated difference in surface air temperature between the Arctic Ocean (including areas covered by sea ice) and the surrounding land, from 60°N to the pole. Warmer colors indicate the ocean is warmer than the land, and cooler colors indicate the land is warmer than the ocean. (b) As in (a), but for the difference in surface pressure between the Arctic Ocean (including areas covered by sea ice) and the surrounding land, from 60°N to the pole. (c) The difference in surface air temperature between the Arctic Ocean (including areas covered by sea ice) and the surrounding land, from 60°N to the pole, as reported in ERA-Interim for the years 1979–2012. (d) As in (c), but for the difference in surface pressure between the Arctic Ocean (including areas covered by sea ice) and the surrounding land, from 60°N to the pole.

The strong thermal contrast between the warmer ocean and the surrounding colder land during winter leads to a decrease in sea level pressure over the Arctic Ocean (Fig. 7b) relative to the surrounding continents. This can be described as a “thermal low,” which is defined by the Glossary of Meteorology as “an area of low atmospheric pressure near the surface resulting from heating of the lower troposphere and the subsequent lifting of isobaric surfaces and divergence of air aloft” (American Meteorological Society 2015). Fig. 8a shows a map of the simulated NDJ sea level pressure at 4 × CO2, and Fig. 8b shows the difference in simulated sea level pressure between the 4 × CO2 simulation and the PI control. At 4 × CO2, the wintertime sea level pressure over the Arctic Ocean has decreased by several hectopascals relative to the PI control. As a result, the sea level pressure has become lower over the Arctic Ocean than over the surrounding continents.

Fig. 8.

(a) Map showing the 4 × CO2 CESM-simulated wintertime (NDJ) sea level pressure and (b) difference (4 × CO2 simulation minus PI control). Maps of the CESM-simulated wintertime (NDJ) 970-hPa winds (vectors) overlaid on the surface temperature (filled contours) for (c) PI control, (d) 4 × CO2 simulation, and (e) the difference.

Fig. 8.

(a) Map showing the 4 × CO2 CESM-simulated wintertime (NDJ) sea level pressure and (b) difference (4 × CO2 simulation minus PI control). Maps of the CESM-simulated wintertime (NDJ) 970-hPa winds (vectors) overlaid on the surface temperature (filled contours) for (c) PI control, (d) 4 × CO2 simulation, and (e) the difference.

Observations suggest that similar changes are occurring in the real world. Figure 7c shows the observed evolution, from 1979 to 2012, of the difference between the surface air temperature over the Arctic Ocean and the surrounding continents, from 60°N to the pole, as reported in ERA-Interim. During the northern summer, the air over the continents is warmer than the air over the surrounding ocean. During winter, however, the air over the ocean is warmer than the air over the land, and the figure shows that this difference has intensified over the 33-yr record, especially since 2000. Figure 7d shows the corresponding trends in the sea level pressure difference between the Arctic Ocean and the surrounding continents. After about 2000, there is a hint of decreasing sea level pressure over the Arctic Ocean during winter, relative to the surrounding continents. The statistical significance of the trends has been assessed using a two-tailed Student’s t test as in the study of Screen and Simmonds (2010b). Both are found to be statistically significant at the 95% level during winter. In short, the reanalysis data show trends that are similar to those found in our simulation.

Figures 8c–e, respectively, show the 970-hPa winds overlaid on surface temperature for PI control, 4 × CO2, and the difference between the 4 × CO2 simulation and the PI control. In the winter of the 4 × CO2 climate, the surface air temperature over the Arctic Ocean is considerably warmer than the temperature over the neighboring continents; the temperature gradients are much weaker in the PI control. The tremendous warming over the Arctic Ocean at 4 × CO2 has largely reversed the sign of the meridional temperature gradient near the borders of the Arctic Ocean. Figures 8c and 8d also show that, in the regions of the Barents, Kara, and Beaufort Seas, the low-level flow from land toward the Arctic Ocean has intensified at 4 × CO2. Despite the sea level pressure decreases in winter, the warming of the lower troposphere leads to geopotential height increases in the middle troposphere of the 4 × CO2 simulation (not shown). The dynamic response to Arctic warming thus has a baroclinic vertical structure, as previously reported in climate change simulations analyzed by Deser et al. (2010) and Sun et al. (2015).

Figure 9 shows the simulated total diabatic heating of the atmosphere at several levels in the lower troposphere for NDJ and for both the PI control and 4 × CO2 simulation. In the PI control (Fig. 9, left), the air over the Arctic Ocean is being diabatically cooled at all levels. This cooling is due to longwave radiation. In contrast, at 4 × CO2 (Fig. 9, right) the lower troposphere is weakly heated everywhere over the Arctic Ocean at the lowest levels. The heating is due to a combination of processes, including vertical diffusion of sensible heat and latent heat release. The heating continues up to the 887-hPa level over the Barents and Chukchi Seas. Cooling occurs above that level. During summer, diabatic cooling is found at all levels in the lower troposphere over the Arctic Ocean (not shown).

Fig. 9.

Maps of the CESM-simulated wintertime (NDJ) diabatic heating at several levels in the lower troposphere for both (left) the PI control and (right) 4 × CO2 simulation.

Fig. 9.

Maps of the CESM-simulated wintertime (NDJ) diabatic heating at several levels in the lower troposphere for both (left) the PI control and (right) 4 × CO2 simulation.

In the PI control, strong wintertime diabatic heating occurs over the North Atlantic and North Pacific Oceans, mainly below the Arctic Circle. This subarctic heating is associated with the strong temperature contrasts between the continents and oceans in those latitudes (e.g., between eastern North America and the western North Atlantic Ocean), which can be seen in Fig. 8. Figure 9 shows that the subarctic heating is considerably reduced in the 4 × CO2 climate. The reason is that the land–sea temperature contrasts are much weaker, as can be seen in Fig. 8.

In summary, at 4 × CO2 the Arctic circulation regime can be described as a weak and shallow “winter monsoon,” in which the low-level winds flow from the cold continents out over the much warmer Arctic Ocean. This monsoon circulation is a consequence of the Arctic warming, which is mainly driven by the combination of the ice–albedo feedback in summer and the ice–insulation feedback in winter. The winter monsoon circulation can also act to enhance the Arctic warming, however, as described below.

The changes in column water vapor shown in Fig. 2c motivate us to examine the vertically integrated Arctic water vapor budget, in which the two source terms are surface evaporation and moisture transport from lower latitudes. Figure 10 shows seasonal trend plots of the sources and sinks of precipitable water due to precipitation (Fig. 10a), evaporation (Fig. 10b), and moisture convergence (Fig. 10c). The moisture convergence has been diagnosed as a residual, taking into account the small tendency term (not shown). The Arctic-averaged moisture convergence and evaporation are both positive throughout the year and for the whole duration of the simulation. As the climate warms, moisture convergence increases substantially in late summer, but it hardly changes during the remainder of the year. In contrast, evaporation increases during late autumn and early winter.

Fig. 10.

As in Fig. 2, but for (a) total precipitation, (b) surface evaporation, and (c) moisture convergence.

Fig. 10.

As in Fig. 2, but for (a) total precipitation, (b) surface evaporation, and (c) moisture convergence.

Figure 11a shows a map of the increase in surface evaporation during NDJ. As expected from Fig. 10, evaporation has intensified over much of the region, especially over the Chukchi, Greenland, and Barents Seas. In the warmer climate, the enhanced wintertime evaporation is of course associated with the increase in open water, but comparison of Figs. 11a and 11b shows that is also promoted by an increase in the surface wind speed. The stronger winds are found over the Barents Sea and north of Eurasia. Comparison with Figs. 8c–e shows that these are regions of strong land–sea thermal contrast and enhanced low-level flow from land to sea. We conclude that the stronger surface winds associated with the winter monsoon circulation are promoting enhanced evaporation over the (relatively) warm water; similar evaporation maxima are associated with the winter monsoons of lower latitudes.

Fig. 11.

Polar stereographic map of change (4 × CO2 simulation minus PI control) in CESM wintertime (NDJ) for (a) evaporation and (b) surface wind speed.

Fig. 11.

Polar stereographic map of change (4 × CO2 simulation minus PI control) in CESM wintertime (NDJ) for (a) evaporation and (b) surface wind speed.

Recall from Fig. 2c that vertically integrated water vapor content of the Arctic atmosphere increases for all months of the year at 4 × CO2. The results shown in Fig. 10 suggest that, in our simulation, the moistening of the Arctic atmosphere in summer, at 4 × CO2, is due to increased moisture transport from lower latitudes but that the moistening in winter is due to increased surface evaporation. Figure 12 shows that the increased evaporation in winter is partly due to stronger surface winds, which are associated with the winter monsoon.

Fig. 12.

Schematic of the ice–insulation and winter monsoon feedbacks, which operate during winter.

Fig. 12.

Schematic of the ice–insulation and winter monsoon feedbacks, which operate during winter.

5. Discussion and concluding remarks

Figure 12 summarizes the results presented in sections 3 and 4 of this paper. The figure is intended to be complementary to Fig. 1, which separately illustrates the ice–albedo feedback. The local ice–insulation feedback, shown in red in the figure, favors warming over the Arctic Ocean in winter. This leads to a reversal of the low-level meridional temperature gradient near the boundaries of the Arctic Ocean. The Arctic winter monsoon is a dynamical response to this reversed temperature gradient. The monsoon circulation is weak and shallow and has only a qualitative resemblance to the powerful winter monsoons of the tropics, but it strongly contrasts with the preindustrial Arctic winter circulation.

DSP recently presented an observational study of the mechanisms that lead to decreases in Arctic sea ice concentration (SIC) during winter. They found that the chain of events begins with an intrusion of warm, humid, cloudy air into the Arctic from lower latitudes. The resulting increases in humidity, cloudiness, and temperature drive an increase in the DLR, which then leads to a decrease in the SIC and an increase in the sea surface temperature. The reduced SIC allows an increase in the surface sensible and latent heat fluxes. Based on these results, DSP concluded that reductions in SIC are primarily caused by intrusions of warm, humid air into the Arctic, associated with large-scale dynamical interactions between the Arctic and lower latitudes, and that the increased primarily humidity is caused by meridional moisture transport rather than stronger surface evaporation. Although the various observed fluctuations discussed by DSP occur on the time scale of days to weeks, they suggested that the observed multidecadal decline in SIC is associated with similar causal mechanisms involving changes in large-scale dynamics and the resulting changes in the transport of warm, humid air into the Arctic. This would imply that Arctic amplification is substantially driven by dynamical interactions with lower latitudes. H.-S. Park et al. (2015) suggest that such interactions can involve convective disturbances in the tropics.

The processes at work in our climate change simulation are different from those described by DSP. As shown in section 3, the simulated Arctic warming in winter is primarily due to a weakening of the diabatic cooling of the Arctic atmosphere; the energy transports from lower latitudes actually decrease during winter as the climate warms. We also showed, in section 4, that the simulated moistening of the Arctic atmosphere during winter is primarily due to an increase in surface evaporation, rather than an increase in poleward moisture transport. Our results do not necessarily contradict those of DSP, because we have analyzed simulated climate change on century time scales, while they studied observations of Arctic variability in the present climate.

Our study has been based on a transient CO2-warming simulation with the CESM, and the results are, of course, model dependent. We have performed a preliminary analysis of the instantaneous quadrupling experiments in the CMIP5 archive and find that many but not all of the CMIP5 models produce results similar to those reported here. Further research is needed to quantify and understand the model dependence of our results.

Acknowledgments

Our exchanges with the two reviewers and the editor have led to major improvements in the manuscript. We appreciate their help. This work has been supported by the National Science Foundation Science and Technology Center for Multi-Scale Modeling of Atmospheric Processes, managed by Colorado State University under Cooperative Agreement ATM-0425247. Computing resources were provided by NCAR’s Computational and Information Systems Laboratory (CISL).

REFERENCES

REFERENCES
Abbot
,
D. S.
, and
E.
Tziperman
,
2008a
:
A high-latitude convective cloud feedback and equable climates
.
Quart. J. Roy. Meteor. Soc.
,
134
,
165
185
, doi:.
Abbot
,
D. S.
, and
E.
Tziperman
,
2008b
:
Sea ice, high-latitude convection, and equable climates
.
Geophys. Res. Lett.
,
35
,
L03702
, doi:.
Abbot
,
D. S.
,
M.
Huber
,
G.
Bousquet
, and
C.
Walker
,
2009
:
High-CO2 cloud radiative forcing feedback over both land and ocean in a global climate model
.
Geophys. Res. Lett.
,
36
,
L05702
, doi:.
American Meteorological Society
,
2015
: Thermal low. Glossary of Meteorology. [Available online at http://glossary.ametsoc.org/wiki/Thermal_low.]
Arnold
,
N.
,
M.
Branson
,
M. A.
Burt
,
D. S.
Abbot
,
Z.
Kuang
,
D. A.
Randall
, and
E.
Tziperman
,
2014
:
Significant consequences of explicit representation of atmospheric convection at high CO2 concentration
.
Proc. Natl. Acad. Sci. USA
,
111
,
10 943
10 948
, doi:.
Bengtsson
,
L.
,
K. I.
Hodges
,
S.
Koumoutsaris
,
M.
Zahn
, and
N.
Keenlyside
,
2011
:
The changing atmospheric water cycle in polar regions in a warmer climate
.
Tellus
,
63A
,
907
920
, doi:.
Bindoff
,
N. L.
, and Coauthors
,
2013
:
Detection and attribution of climate change: From global to regional
. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 867–952.
Bintanja
,
R.
, and
F. M.
Selten
,
2014
:
Future increases in Arctic precipitation linked to local evaporation and sea ice retreat
.
Nature
,
509
,
479
482
, doi:.
Bitz
,
C. M.
, and
W. H.
Lipscomb
,
1999
:
An energy-conserving thermodynamic model of sea ice
.
J. Geophys. Res.
,
104
,
15 669
15 677
, doi:.
Curry
,
J. A.
, and
E. E.
Ebert
,
1992
:
Annual cycle of radiative fluxes over the Arctic Ocean: Sensitivity to cloud optical properties
.
J. Climate
,
5
,
1267
1280
, doi:.
Dee
,
D. P.
, and Coauthors
,
2011
:
The ERA-Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, doi:.
Deser
,
C.
,
R.
Tomas
,
M.
Alexander
, and
D.
Lawrence
,
2010
:
The seasonal atmospheric response to projected Arctic sea ice loss in the late twenty-first century
.
J. Climate
,
23
,
333
351
, doi:.
Di Biagio
,
C.
,
A.
di Sarra
,
P.
Eriksen
,
S. E.
Ascanius
,
G.
Muscari
, and
B.
Holben
,
2012
:
Effect of surface albedo, water vapour, and atmospheric aerosols on the cloud-free shortwave radiative budget in the Arctic
.
Climate Dyn.
,
39
,
953
969
, doi:.
Francis
,
J. A.
, and
E.
Hunter
,
2007
:
Changes in the fabric of the Arctic’s greenhouse blanket
.
Environ. Res. Lett.
,
2
,
045011
, doi:.
Francis
,
J. A.
,
J. R.
Key
, and
X.
Wang
,
2005
:
Clues to variability in Arctic minimum sea ice extent
.
Geophys. Res. Lett.
,
32
,
L21501
, doi:.
Gettelman
,
A.
,
H.
Morrison
, and
S. J.
Ghan
,
2008
:
A new two-moment bulk stratiform cloud microphysics scheme in the Community Atmosphere Model, version 3 (CAM3). Part II: Single-column and global results
.
J. Climate
,
21
,
3660
3679
, doi:.
Ghatak
,
D.
, and
J.
Miller
,
2013
:
Implications for Arctic amplification of changes in the strength of the water vapor feedback
.
J. Geophys. Res.
,
118
,
7569
7578
, doi:.
Gorodetskaya
,
I. V.
, and
L. B.
Tremblay
,
2008
: Arctic cloud properties and radiative forcing from observations and their role in sea ice decline predicted by the NCAR CCSM3 model during the 21st century. Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications, Geophys. Monogr., Vol. 180, Amer. Geophys. Union, 213–268.
Graversen
,
R. G.
, and
M.
Wang
,
2009
:
Arctic amplification in a coupled climate model with locked albedo
.
Climate Dyn.
,
33
,
629
643
, doi:.
Graversen
,
R. G.
,
P. G.
Langen
, and
T.
Mauritsen
,
2014
:
Arctic amplification in CCSM4: Contributions from the lapse rate and surface albedo feedbacks
.
J. Climate
,
27
,
4433
4450
, doi:.
Hall
,
A.
,
2004
:
Role of surface albedo feedback in climate
.
J. Climate
,
17
,
1550
1568
, doi:.
Hansen
,
J.
,
A.
Lacis
,
D.
Rind
,
G.
Russel
,
P.
Stone
,
I.
Fung
,
R.
Ruedy
, and
J.
Lerner
,
1984
: Climate sensitivity: Analysis of feedback mechanisms. Climate Processes and Climate Sensitivity, Geophys. Monogr., Vol. 29, Amer. Geophys. Union, 130–163.
Held
,
I. M.
, and
B. J.
Soden
,
2006
:
Robust responses of the hydrological cycle to global warming
.
J. Climate
,
19
,
5686
5699
, doi:.
Higgins
,
M. E.
, and
J. J.
Cassano
,
2009
:
Impacts of reduced sea ice on winter Arctic atmospheric circulation, precipitation, and temperature
.
J. Geophys. Res.
,
114
,
D16107
, doi:.
Holland
,
M. M.
, and
C. M.
Bitz
,
2003
:
Arctic amplification of climate change in coupled models
.
Climate Dyn.
,
21
,
221
232
, doi:.
Holland
,
M. M.
,
D. A.
Bailey
,
B. J.
Briegleb
,
B.
Light
, and
E.
Hunke
,
2012
:
Improved sea ice shortwave radiation physics in CCSM4: The impact of melt ponds and aerosols on Arctic sea ice
.
J. Climate
,
25
,
1413
1430
, doi:.
Hunke
,
E. C.
, and
J. K.
Dukowicz
,
1997
:
An elastic–viscous–plastic model for sea ice dynamics
.
J. Phys. Oceanogr.
,
27
,
1849
1867
, doi:.
Hunke
,
E. C.
, and
W. H.
Lipscomb
,
2008
: CICE: The Los Alamos Sea Ice Model, documentation and software, version 4.0. Los Alamos National Laboratory Tech. Rep. LACC-06-012, 76 pp.
Hurrell
,
J. W.
, and Coauthors
,
2013
:
The Community Earth System Model: A framework for collaborative research
.
Bull. Amer. Meteor. Soc.
,
94
,
1339
1360
, doi:.
Iacono
,
M. J.
,
J. S.
Delamere
,
E. J.
Mlawer
,
M. W.
Shephard
,
S. A.
Clough
, and
W. D.
Collins
,
2008
:
Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models
.
J. Geophys. Res.
,
113
,
D13103
, doi:.
Jahn
,
A.
, and Coauthors
,
2012
:
Late twentieth-century simulation of Arctic sea ice and ocean properties in CCSM4
.
J. Climate
,
25
,
1431
1452
, doi:.
Jakobson
,
E.
, and
T.
Vihma
,
2010
:
Atmospheric moisture budget in the Arctic based on the ERA-40 Reanalysis
.
Int. J. Climatol.
,
30
,
2175
2194
, doi:.
Kay
,
J. E.
, and
A.
Gettelman
,
2009
:
Cloud influence on and response to seasonal Arctic sea ice loss
.
J. Geophys. Res.
,
114
,
D18204
, doi:.
Kurita
,
N.
,
2011
:
Origin of Arctic water vapor during the ice-growth season
.
Geophys. Res. Lett.
,
38
,
L02709
, doi:.
Lee
,
S.
,
T.
Gong
,
N.
Johnson
,
S. B.
Feldstein
, and
D.
Pollard
,
2011
:
On the possible link between tropical convection and the Northern Hemisphere Arctic surface air temperature change between 1958 and 2001
.
J. Climate
,
24
,
4350
4367
, doi:.
Leibowicz
,
B. D.
,
D. S.
Abbot
,
K.
Emanuel
, and
E.
Tziperman
,
2012
:
Correlation between present-day model simulation of Arctic cloud radiative forcing and sea ice consistent with positive winter convective cloud feedback
.
J. Adv. Model. Earth Syst
,
4
,
M07002
, doi:.
Lindsay
,
R.
,
M.
Wensnahan
,
A.
Schweiger
, and
J.
Zhang
,
2014
:
Evaluation of seven different atmospheric reanalysis products in the Arctic
.
J. Climate
,
27
,
2588
2606
, doi:.
Lu
,
J.
, and
M.
Cai
,
2010
:
Quantifying contributions to polar warming amplification in an idealized coupled general circulation model
.
Climate Dyn.
,
34
,
669
687
, doi:.
Manabe
,
S.
, and
R. T.
Wetherald
,
1975
:
The effects of doubling the CO2 concentrations on the climate of a general circulation model
.
J. Atmos. Sci.
,
32
,
3
15
, doi:.
Manabe
,
S.
, and
R. J.
Stouffer
,
1980
:
Sensitivity of a global climate model to an increase of CO2 concentration in the atmosphere
.
J. Geophys. Res.
,
85
,
5529
5554
, doi:.
Manabe
,
S.
, and
R. T.
Wetherald
,
1980
:
On the distribution of climate change resulting from an increase in CO2 content of the atmosphere
.
J. Atmos. Sci.
,
37
,
99
118
, doi:.
Morrison
,
H.
, and
A. A.
Gettelman
,
2008
:
A new two-moment bulk stratiform cloud microphysics scheme in the NCAR Community Atmosphere Model, version 3 (CAM3). Part I: Description and numerical tests
.
J. Climate
,
21
,
3642
3659
, doi:.
Park
,
D.-S.
,
S.
Lee
, and
S. B.
Feldstein
,
2015
:
Attribution of the recent winter sea ice decline over the Atlantic sector of the Arctic Ocean
.
J. Climate
,
28
,
4027
4033
, doi:.
Park
,
H.-S.
,
S.
Lee
,
S.-W.
Son
,
S. B.
Feldstein
, and
Y.
Kosaka
,
2015
:
The impact of poleward moisture and sensible heat flux on Arctic winter sea ice variability
.
J. Climate
,
28
,
5030
5040
, doi:.
Pithan
,
F.
, and
T.
Mauritsen
,
2014
:
Arctic amplification dominated by temperature feedbacks in contemporary climate models
.
Nat. Geosci.
,
7
,
181
184
, doi:.
Rinke
,
A.
,
C.
Melsheimer
,
K.
Dethlof
, and
G.
Heygster
,
2009
:
Arctic total water vapor: Comparison of regional climate simulations with observations, and simulated decadal trends
.
J. Hydrometeor.
,
10
,
113
129
, doi:.
Robock
,
A.
,
1980
:
The seasonal cycle of snow cover, sea ice, and surface albedo
.
Mon. Wea. Rev.
,
108
,
267
285
, doi:.
Schweiger
,
A. J.
,
R. W.
Lindsay
,
S.
Vavrus
, and
J. A.
Francis
,
2008
:
Relationships between Arctic sea ice and clouds during autumn
.
J. Climate
,
21
,
4799
4810
, doi:.
Screen
,
J. A.
, and
I.
Simmonds
,
2010a
:
The central role of diminishing sea ice in recent Arctic temperature amplification
.
Nature
,
464
,
1334
1337
, doi:.
Screen
,
J. A.
, and
I.
Simmonds
,
2010b
:
Increasing fall–winter energy loss from the Arctic Ocean and its role in Arctic temperature amplification
.
Geophys. Res. Lett.
,
37
,
L16707
, doi:.
Serreze
,
M. C.
, and
R. G.
Barry
,
2011
:
Processes and impacts of Arctic amplification: A research synthesis
.
Global Planet. Change
,
77
,
85
96
, doi:.
Serreze
,
M. C.
,
A. P.
Barrett
, and
J.
Stroeve
,
2012
:
Recent changes in tropospheric water vapor over the Arctic as assessed from radiosondes and atmospheric reanalyses
.
J. Geophys. Res.
,
117
,
D10104
, doi:.
Skific
,
N.
, and
J. A.
Francis
,
2013
:
Drivers of projected change in Arctic moist static energy transport
.
J. Geophys. Res.
,
118
,
2748
2761
, doi:.
Skific
,
N.
,
J. A.
Francis
, and
J. J.
Cassano
,
2009a
:
Attribution of projected changes in atmospheric moisture transport in the Arctic: A self-organizing map perspective
.
J. Climate
,
22
,
4135
4153
, doi:.
Skific
,
N.
,
J. A.
Francis
, and
J. J.
Cassano
,
2009b
:
Attribution of seasonal and regional changes in Arctic moisture convergence
.
J. Climate
,
22
,
5115
5134
, doi:.
Smith
,
R. D.
, and Coauthors
,
2010
: The Parallel Ocean Program (POP) reference manual: Ocean component of the Community Climate System Model (CCSM) and Community Earth System Model (CESM). Los Alamos National Laboratory Tech. Rep. LAUR-10-01853, 141 pp. [Available online at http://www.cesm.ucar.edu/models/cesm1.0/pop2/doc/sci/POPRefManual.pdf.]
Stroeve
,
J. C.
, and
W.
Meier
,
2012
:
Arctic sea ice decline
. Greenhouse Gases—Emission, Measurement and Management. G. Liu, Eds., InTech, 441–462.
Sun
,
L.
,
C.
Deser
, and
R. A.
Tomas
,
2015
:
Mechanisms of stratospheric and tropospheric circulation response to projected Arctic sea ice loss
.
J. Climate
,
28
,
7824
7845
, doi:.
Trenberth
,
K. E.
,
J.
Fasullo
, and
L.
Smith
,
2005
:
Trends and variability in column-integrated atmospheric water vapour
.
Climate Dyn.
,
24
,
741
758
, doi:.
Vavrus
,
S.
,
2004
:
The impact of cloud feedbacks on Arctic climate under greenhouse forcing
.
J. Climate
,
17
,
603
615
, doi:.
Vavrus
,
S.
,
U.
Bhatt
, and
V.
Alexeev
,
2011a
:
Factors influencing simulated changes in future Arctic cloudiness
.
J. Climate
,
24
,
4817
4830
, doi:.
Vavrus
,
S.
,
M.
Holland
, and
D.
Bailey
,
2011b
:
Changes in Arctic clouds during intervals of rapid sea ice loss
.
Climate Dyn.
,
36
,
1475
1489
, doi:.
Walsh
,
J. E.
, and
W. L.
Chapman
,
1998
:
Arctic cloud–radiation–temperature associations in observational data and atmospheric reanalyses
.
J. Climate
,
11
,
3030
3045
, doi:.
Winton
,
M.
,
2006
:
Amplified Arctic climate change: What does surface albedo feedback have to do with it?
Geophys. Res. Lett.
,
33
,
L03701
, doi:.
Zhang
,
T.
,
K.
Stamnes
, and
S. A.
Bowling
,
1996
:
Impact of clouds on surface radiative fluxes and snowmelt in the Arctic and subarctic
.
J. Climate
,
9
,
2110
2123
, doi:.
Zhang
,
X.
,
J.
He
,
J.
Zhang
,
I.
Polyakov
,
R.
Gerdes
,
J.
Inoue
, and
P.
Wu
,
2012
:
Enhanced polarward moisture transport and amplified northern high-latitude wetting trend
.
Nat. Climate Change
,
3
,
47
51
, doi:.
Zygmuntowska
,
M.
,
T.
Mauritsen
,
J.
Quaas
, and
L.
Kaleschke
,
2012
:
Arctic clouds and surface radiation—A critical comparison of satellite retrievals and the ERA-Interim reanalysis
.
Atmos. Chem. Phys.
,
12
,
6667
6677
, doi:.

Footnotes

1

We use the term “dark season” to refer to the months when little or no solar radiation reaches the Arctic.