A record-breaking low ice cover over the Great Lakes during winter 2011/2012: combined effects of a strong positive NAO and La Niña

Abstract

A record-breaking low ice cover occurred in the North American Great Lakes during winter 2011/2012, in conjunction with a strong positive Arctic Oscillation/North Atlantic Oscillation (+AO/NAO) and a La Niña event. Large-scale atmosphere circulation in the Pacific/North America (PNA) region reflected a combined signal of La Niña and +NAO. Surface heat flux analysis shows that sensible heat flux contributed most to the net surface heat flux anomaly. Surface air temperature is the dominant factor governing the interannual variability of Great Lakes ice cover. Neither La Niña nor +NAO alone can be responsible for the extreme warmth; the typical mid-latitude response to La Niña events is a negative PNA pattern, which does not have a significant impact on Great Lakes winter climate; the positive phase of NAO is usually associated with moderate warming. When the two occurred simultaneously, the combined effects of La Niña and +NAO resulted in a negative East Pacific pattern with a negative center over Alaska/Western Canada, a positive center in the eastern North Pacific (north of Hawaii), and an enhanced positive center over the eastern and southern United States. The overall pattern prohibited the movement of the Arctic air mass into mid-latitudes and enhanced southerly flow and warm advection from the Gulf of Mexico over the eastern United States and Great Lakes region, leading to the record-breaking low ice cover. It is another climatic pattern that can induce extreme warming in the Great Lakes region in addition to strong El Niño events. A very similar event occurred in the winter of 1999/2000. This extreme warm winter and spring in 2012 had significant impacts on the physical environment, as well as counterintuitive effects on phytoplankton abundance.

Appendix: Sea ice thermodynamic equations

Typically, an ice cover will contain a variety of ice thickness. To approximately parameterize this variable thickness ice cover, two thickness levels are used in the sea ice model: thick and thin. Two quantities are used to describe an ice cover in any grid cell–equivalent thickness h and concentration A (0–1), which is defined as the fraction covered by thick ice. The rest of the cell is covered by thin ice, which is always taken to be of zero thickness (i.e. open water or lead) (Hibler 1979).

Within the two-level approximation, ice thickness and concentration growth is given by (Hibler 1979)

$$S_{h} \equiv \partial h/\partial t = f_{g} \left( {h/A} \right)A + \left( {1 - A} \right)f_{g} (0)$$

$$S_{a} \equiv \partial A/\partial t = \left( {f_{g} \left( 0 \right)/H_{0} } \right)\left( {1 - A} \right)\delta_{1} + \left( {A/2h} \right)S_{h} \delta_{2}$$

$$\delta_{1} = 1 if\;f_{g} \left( 0 \right) > 0,\delta_{1} = 0 if\,f_{g} \left( 0 \right) < 0$$

$$\delta_{2} = 1 if\;S_{h} < 0,\delta_{2} = 0\,if\;S_{h} < 0$$

with fg(h) the growth rate of ice of thickness h, and H ₀ a fixed demarcation thickness between thin and thick ice (0.5 m is often used in Arctic). The growth rate is determined by the net heat losses over ice or water

$$f_{g} (h) = - Q_{net} /\rho_{i} L_{v}$$

with Q _net the net surface heat flux over ice or water at the freezing temperature, ρ_i = 900 kgm⁻³ the density of ice, L_v = 2.4 × 10⁶ Jkg⁻¹ is the latent heat of fusion of ice.

1.1 Heat budget on the upper ice surface

The net heat flux on the ice surface is

$$Q_{ai} = H_{sw} + H_{lw} + H_{s} + H_{l}$$

where short wave radiation H _sw , net long wave radiation H _lw, the sensible heat flux H _s , the latent heat flux H _l , net long wave radiation, and conductive flux F _c are calculated by the formulations:

$$H_{sw} = I_{0} (1 - \alpha_{i} )$$

$$H_{lw} = - \varepsilon \delta \left[ {\left( {1 - k_{c} C} \right)\left( {0.254 - 4.95 \times 10^{ - 5} e_{a} } \right)T_{a} + 4(T_{i} - T_{a} )} \right]T_{a}^{3}$$

$$H_{s} = \rho_{a} C_{s} C_{p} \left| {V_{a} } \right|\left( {T_{a} - T_{i} } \right)$$

$$H_{l} = \rho_{a} C_{l} l_{e} \left| {V_{a} } \right|\left( {q_{a} - q_{i} } \right)$$

where C _p (=1,005 J kg⁻¹K⁻¹) is the specific heat of air and le (=2.5 × 10⁶ J/kg) is the latent heat of evaporation. T _a and T _i are the surface air temperature and the surface temperature of ice, respectively. q _a and q _i are the specific humidity of the surface air and ice surface, respectively. The heat transfer coefficient C _s and C _l , which depend on wind speed, are set based on Large and Pond (1981; 1982). For example, at a given wind of 5 m/s, the Cs is 1.06 × 10⁻³ (0.6 × 10⁻³) for an unstable (stable) condition, and C _l is 1.1 × 10⁻³. ε (=0.95) is the emissivity of sea surface relative to a black body. σ (=5.67 × 10⁻⁸) is Stefan-Boltzmann constant. k _c (=0.7) is the cloud factor, and C is the cloud fraction. ρ _a (=1.3 kg m⁻³) and ρ _w (=1.025 × 10³ kg m⁻³) is the air and the seawater density, respectively.

Inside the ice, heat conduction flux Fc exists because of a nonuniform temperature distribution:

$$F_{c} = - k_{i} (T_{i} - T_{f} )/h_{i}$$

where k _i  = 2.04 J (kg K)⁻¹ is the ice thermal conductivity, T _f is the ice bottom temperature, Ti is the sea ice surface temperature, and hi is the sea ice thickness.

The sea ice surface temperature Ti is an unknown, and is calculated using the following heat equilibrium equation at the air-ice interface:

$$Q_{ai} - F_{c} = 0$$

Solving this equation, we can obtain Ti by an iteration method, and, in turn, we can calculate the heat fluxes over the upper ice surface.

1.2 Heat budget over lead and water

Sensible and latent heat flux over lead and water are calculated using the bulk aerodynamic method suggested by Large and Pond (1982). The drag coefficient, Stanton number and Dalton number are functions of height and stability and commonly evaluated in the equivalent neutral case at 10 m. For brevity’s sake, the calculation process is not copied here. The net long wave radiation is calculated as the same method as over ice.