Large-scale Sea Ice Models Research Articles

Simulating sea ice drift in the Southern Ocean incorporating real wind data using the LSFEM

AbstractThe viscous‐plastic sea ice model of [2] describes the motion of sea ice on large scales. The numerical model of the sea ice drift considers velocities and stresses and is coupled with the field quantities ice thickness and ice concentration, which are modeled by transient advection equations. Here, the viscosity in sense of a non‐Newtonian fluid depends on velocity gradients, as well as on ice concentration and ice thickness. This leads to a strong nonlinearity of the constitutive model.In [8], we discussed the hurdles of large scale sea ice models and possible solutions of an implementation within a least‐squares finite element method (LSFEM). Previous studies on sea ice model implementations have shown that the LSFEM is a promising solution to the numerically difficult problem, cf. [9] and [7]. In this work, another essential aspect of sea ice research comes into play: embedding real data into the numerical solution is an important field of research in sea ice and ocean modeling. Here, we show a 2D boundary value problem for simulating sea ice drift in the Southern Ocean incorporating real wind data from the ERA‐5 project [1].

Sensitivity of Arctic sea ice to melt pond processes and atmospheric forcing: A model study

Melt ponds are pools of meltwater forming principally on Arctic sea ice during the melt season. The albedo of melt ponds is a key component of the surface energy balance. For this reason, various melt pond schemes have been developed for climate models. These schemes require assumptions on the physical processes governing melt ponds as well as a knowledge of the atmospheric state, which are not perfectly known. In this study, we investigate the effects of the sources of uncertainty from the prescribed atmospheric surface state, the melt pond scheme definition and the refreezing formulation of melt ponds on the simulated Arctic sea ice and melt pond properties with the NEMO-LIM3 ocean–sea ice general circulation model. We find that the simulated melt pond state is largely controlled by the freezing point of melt ponds. The representation of melt ponds is in better agreement with observations when using the freezing point of −0.15°C compared to the value of −2.00°C, in our model set-up. All the simulations feature positive trends in melt pond area fraction over the past decades. However, only 3 out of 8 simulations have significant positive trends in melt pond volume per sea ice area. This suggests an influence of the sea ice state for melt ponds over the last 30 years. Overall, we find that the simulated sea ice state, and in particular sea ice volume, is more affected by changes in the prescribed atmospheric forcing than by changes in the prescribed melt pond scheme or refreezing formulation. Including explicit melt pond schemes in large-scale sea-ice models offer the possibility to improve the representation of the surface energy balance in climate general circulation models. Our results underline that, in parallel to these efforts in model developments, improved estimates of surface atmospheric conditions will be required to achieve more realistic sea ice states.

A Computational Fluid Dynamics Model for the Small-Scale Dynamics of Wave, Ice Floe and Interstitial Grease Ice Interaction

The marginal ice zone is a highly dynamical region where sea ice and ocean waves interact. Large-scale sea ice models only compute domain-averaged responses. As the majority of the marginal ice zone consists of mobile ice floes surrounded by grease ice, finer-scale modelling is needed to resolve variations of its mechanical properties, wave-induced pressure gradients and drag forces acting on the ice floes. A novel computational fluid dynamics approach is presented that considers the heterogeneous sea ice material composition and accounts for the wave-ice interaction dynamics. Results show, after comparing three realistic sea ice layouts with similar concentration and floe diameter, that the discrepancy between the domain-averaged temporal stress and strain rate evolutions increases for decreasing wave period. Furthermore, strain rate and viscosity are mostly affected by the variability of ice floe shape and diameter.

Thermodynamics of slush and snow–ice formation in the Antarctic sea-ice zone

Snow over Antarctic sea ice is often flooded by brine or seawater, particularly in spring, forming slush and snow ice. Here, we evaluate the representation of the thermodynamics of slush and snow–ice formation in large-scale sea-ice models, using laboratory experiments (NaCl solutions poured into grated ice in an isolated container). Scaling analysis highlights latent heat as the main term of the energy budget. The temperature of the new sea ice immediately after flooding is found very close to the saltwater freezing point, whereas its bulk salinity is typically >20g/kg. Large-scale sea-ice models faithfully represent such physics, yet the uncertainty on the origin of flooding saltwater impacts the calculated new ice temperature, because of the different salinities of seawater and brine. The laboratory experiments also suggest a potential limitation to the existing physical representations of flooding: for brine fractions >60%, ice crystals start floating upon saltwater. Natural sea-ice observations suggest that the isolated system assumption holds for a few hours at most, after which rapid heat and salt exchanges mostly destroy the initial flooding signature on temperature and salinity. A small footprint on ice salinity remains however, natural snow ice is found 3–5g/kg more saline than other forms of sea ice.

Lagrangian modelling of frazil ice in the ocean

AbstractA new modelling framework using Lagrangian particle tracking has been developed to assess dynamic and thermodynamic effects of underwater frazil ice. This frazil-ice model treats a Lagrangian particle as a bulk cluster of many frazil crystals, and calculates the thermodynamic growth of each particle and the corresponding budget of latent heat and fresh water. The effective density and viscosity of sea water depend on the mass fraction of underwater frazil ice, and hence affect ocean convection. An idealized experiment using our model successfully reproduces the formation of underwater frazil ice and its transition to grease ice at the surface. Because underwater frazil ice does not reduce the atmosphere/ocean heat exchange, surface heat flux and net sea-ice production in the experiment with frazil ice are relatively high compared with the experiment where surface cooling directly leads to columnar growth of a solid ice cover which effectively insulates the heat flux. These results suggest that large-scale sea-ice models which do not take account of the effects of frazil ice might underestimate atmosphere/ocean heat exchange, particularly at times of active new ice formation.

Benefits from representing snow properties and related processes in coupled ocean–sea ice models

Several large-scale sea ice simulations are performed over the last three decades using a coupled ocean–sea ice model under the same experimental setup but partly modifying the representation of snow physics in the model. The inter-simulation spread analysis yields that the simulated multi-year ice is sensitive to such changes while the seasonal sea ice, is rather dominantly driven by the external oceanic and atmospheric forcings. In the context of a thinning Arctic sea ice cover, those findings suggest that including snow processes in large-scale sea ice models is beneficial, if not necessary, to predict the timing of the Arctic multi-year ice disappearance, whereas the operational forecasting of first-year ice extent using fully coupled models will likely require improvement to the oceanic and atmospheric components themselves.

A parallel Jacobian-free Newton–Krylov solver for a coupled sea ice-ocean model

The most common representation of sea ice dynamics in climate models assumes that sea ice is a quasi-continuous non-normal fluid with a viscous-plastic rheology. This rheology leads to non-linear sea ice momentum equations that are notoriously difficult to solve. Recently a Jacobian-free Newton–Krylov (JFNK) solver was shown to solve the equations accurately at moderate costs. This solver is extended for massive parallel architectures and vector computers and implemented in a coupled sea ice–ocean general circulation model for climate studies. Numerical performance is discussed along with numerical difficulties in realistic applications with up to 1920 CPUs. The parallel JFNK-solverʼs scalability competes with traditional solvers although the collective communication overhead starts to show a little earlier. When accuracy of the solution is required (i.e. reduction of the residual norm of the momentum equations of more that one or two orders of magnitude) the JFNK-solver is unrivalled in efficiency. The new implementation opens up the opportunity to explore physical mechanisms in the context of large scale sea ice models and climate models and to clearly differentiate these physical effects from numerical artifacts.

A new snow thermodynamic scheme for large-scale sea-ice models

AbstractThis paper assesses the capabilities of a new one-dimensional snow scheme developed for the thermodynamic component of the Louvain-la-Neuve sea-Ice Model (LIM). the model is validated at Point Barrow, Alaska, and at Ice Station Polarstern (ISPOL) in the western Weddell Sea, Southern Ocean. the new snow thermodynamic scheme leads to better snow internal temperature profiles, with a set-up-dependent increase in the correlation between simulated and observed temperature profiles. on average over all runs, these correlations are 27% better with the six-layer configuration. the model’s ability to reproduce observed temperatures improves with the number of snow layers, but stabilizes after a threshold layer number is reached. the lowest and highest values for this threshold are 3 (at Point Barrow) and 6 (at ISPOL), respectively. Overall, the improvement of the model’s ability to simulate sea-ice thickness is not as significant as for snow temperature, probably because of the rather crude representation of the snow stratigraphy in the model.

Assessment of radiation forcing data sets for large-scale sea ice models in the Southern Ocean

Little is known about errors in the atmospheric forcings of large-scale sea ice-ocean models around Antarctica. These forcings involve atmospheric reanalyses, typically those from the National Center for Environmental Prediction and National Center from Atmospheric Research (NCEP-NCAR), climatologies, and empirical parameterizations of atmosphere-ice heat and radiation fluxes. In the present paper, we evaluate the atmospheric forcing fields of sea ice models in the Southern Ocean using meteorological and radiation observations from two drifting station experiments over Antarctic sea ice. These are Sea Ice Mass Balance in the Antarctic (SIMBA, Bellingshausen Sea, October 2007) and ISPOL (Ice Station POLarstern, Weddell Sea, December 2004). For the comparison, it is assumed that those point measurements are representative of the whole model grid cell they were collected in. Analysis suggests that the NCEP-NCAR reanalyses have relatively low biases for variables that are assimilated by the system (temperature, winds and humidity) and are less accurate for those which are not (cloud fraction and radiation fluxes). The main deficiencies are significant day-to-day errors in air temperature (root-mean-square error 1.4–3.8 °C) and a 0.2–0.6 g/kg mean overestimation in NCEP-NCAR specific humidity. In addition, associated with an underestimation of cloud fraction, NCEP-NCAR shortwave radiation features a large positive bias (43–109 W/m 2), partly compensated by a 20–45 W/m 2 negative bias in longwave radiation. Those biases can be drastically reduced by using empirical formulae of radiation fluxes and climatologies of relative humidity and cloud cover. However, this procedure leads to a loss of day-to-day and interannual variability in the radiation fields. We provide technical recommendations on how the radiation forcing should be handled to reduce sea ice model forcing errors. The various errors in forcing fields found here should not hide the great value of atmospheric reanalyses for the simulation of the ice-ocean system.

On the assimilation of ice velocity and concentration data into large-scale sea ice models

Abstract. Data assimilation into sea ice models designed for climate studies has started about 15 years ago. In most of the studies conducted so far, it is assumed that the improvement brought by the assimilation is straightforward. However, some studies suggest this might not be true. In order to elucidate this question and to find an appropriate way to further assimilate sea ice concentration and velocity observations into a global sea ice-ocean model, we analyze here results from a number of twin experiments (i.e. experiments in which the assimilated data are model outputs) carried out with a simplified model of the Arctic sea ice pack. Our objective is to determine to what degree the assimilation of ice velocity and/or concentration data improves the global performance of the model and, more specifically, reduces the error in the computed ice thickness. A simple optimal interpolation scheme is used, and outputs from a control run and from perturbed experiments without and with data assimilation are thoroughly compared. Our results indicate that, under certain conditions depending on the assimilation weights and the type of model error, the assimilation of ice velocity data enhances the model performance. The assimilation of ice concentration data can also help in improving the model behavior, but it has to be handled with care because of the strong connection between ice concentration and ice thickness. This study is first step towards real data assimilation into NEMO-LIM, a global sea ice-ocean model.

Sensitivity of the Antarctic sea ice to the thermal conductivity of snow

The sensitivity of a global, coarse‐resolution ice‐ocean model to a decrease in the thermal conductivity of the snow overlaying Antarctic sea ice (ksant) is investigated. This study was motivated by recent observations made in the eastern sector of the Southern Ocean and in the Bellingshausen, Amundsen, and Ross Seas, which suggest that the value of ksant usually used in large‐scale sea ice models is about a factor two too high. When ksant is reduced by half in our model, the average thickness of the Antarctic ice pack decreases by 10 cm (10%) and the geographical distributions of snow and ice thicknesses become more realistic. A strengthening of the Southern Ocean stratification and a concomitant weakening of the Antarctic Bottom Water meridional overturning are also noticed. This ultimately impacts on the sensible heat flux from the ocean to the ice and thus on the rate of thermodynamic ice growth.

The effect of sea-ice parameterizations on the simulation of the Arctic ice pack

A one-dimensional (1-D), thermodynamic sea-ice model with parameterized ice dynamics is coupled to a mixed-layer ocean model and driven with prescribed atmospheric forcings for the central Arctic. The model is used to calculate the sensitivity of the ice pack to various parameterizations that have traditionally been neglected or considered only implicitly in large-scale sea-ice models. The model includes melt ponds, leads (with summertime stratification), an ice-export term, a stability-dependent air–sea heat-exchange coefficient, a prognostic ocean–ice heat exchange, a crude ice-thickness distribution, and a sophisticated albedo parameterization.The ice pack is sensitive to the partitioning of solar energy between lateral melting and mixed-layer warming, with the most realistic simulations occurring when the heat is nearly evenly divided between these two processes. Conversely, ice thickness and coverage are fairly insensitive to the amount of lateral mixing within the upper ocean, vertical mixing within leads, and to the partitioning of mixed-layer heat content between warming the water and melting the ice bottom. The ice concentration during summer is strongly dependent on the assumed ice-thickness distribution: the amount of open water during summer is less than half the size of the empirically based distribution used here, compared with one in which ice floes are distributed uniformly across a range of thicknesses.

On modeling sea-ice fracture and flow in numerical investigations of climate

Sea ice on the large scale is characterized by leads and ridges that typically have a given orientation. Because of various flaws, we would expect that the ice will form oriented leads and ice-thickness characteristics that control the heat and moisture fluxes into the atmosphere. Prediction of these oriented leads, ridges and slip lines is relevant to understanding the role of ice mechanics in global climate change as they can play a significant role in the ice-thickness distribution.In this paper we develop a model for the dynamical treatment of leads and oriented flaws in large-scale sea-ice models. Two particular isotropic realizations of this model relevant to climate studies are examined: (a) an isotropic composite with oriented leads in all directions imbedded in thick ice, and (b) a simple "strain hardening" isotropic model where only oriented leads having the potential to open rapidly are allowed. Under applied stress both models yield preferential deformation along a symmetric pair of intersecting leads or ridges with the intersection angles dependent on the confinement stress. The "uniform-orientation" model results in a yield curve that approximates a sine lens, while the "strain hardening" model has a teardrop-like yield curve. How the resulting fracture-based yield curves and non-normal flow rules may be cast in a form usable in numerical investigations of climate is discussed.

Effects of the snow cover on Antarctic sea ice and potential modulation of its response to climate change

Based on presented field data, it is shown that snow contributes roughly 8% to the total mass of ice in the Weddell Sea. Snow depth averages 0.16 m on first-year ice (average thickness 0.75 m) and 0.53 m on second-year ice (average thickness 1.70 m). Due to snow loading, sea ice is depressed below water level and flooded by sea water. As a result of flooding, snow ice forms through congelation of sea water and brine in a matrix of meteoric ice (i.e. snow). Sea-ice growth has been simulated with a one-dimensional model, treating the evolution of salinity, porosity and thermal properties of the ice. Simulations demonstrate that in the presence of a snow cover, ice growth is significantly reduced. Brine volumes increase by a factor of 1.5–2, affecting properties such as ice strength. Snow-ice formation depends on the evolution of freeboard and ice permeability. Effects of accumulation-rate changes have been assessed, for the Weddell Sea with a large-scale sea-ice model accounting for snow-ice formation. Results for different scenarios are presented and compared with field data and one-dimensional simulations. The role of snow in modulating the response of Antarctic sea ice to climate change is discussed.

Effects of the snow cover on Antarctic sea ice and potential modulation of its response to climate change

Based on presented field data, it is shown that snow contributes roughly 8% to the total mass of ice in the Weddell Sea. Snow depth averages 0.16 m on first-year ice (average thickness 0.75 m) and 0.53 m on second-year ice (average thickness 1.70 m). Due to snow loading, sea ice is depressed below water level and flooded by sea water. As a result of flooding, snow ice forms through congelation of sea water and brine in a matrix of meteoric ice (i.e. snow). Sea-ice growth has been simulated with a one-dimensional model, treating the evolution of salinity, porosity and thermal properties of the ice. Simulations demonstrate that in the presence of a snow cover, ice growth is significantly reduced. Brine volumes increase by a factor of 1.5–2, affecting properties such as ice strength. Snow-ice formation depends on the evolution of freeboard and ice permeability. Effects of accumulation-rate changes have been assessed, for the Weddell Sea with a large-scale sea-ice model accounting for snow-ice formation. Results for different scenarios are presented and compared with field data and one-dimensional simulations. The role of snow in modulating the response of Antarctic sea ice to climate change is discussed.

On the momentum forcing of a large-scale sea-ice model

A large-scale sea-ice - oceanic mixed-layer model for the Southern Ocean is forced with daily atmospheric fields from operational numerical weather prediction analyses. The strength of the atmospheric forcing is modified considering atmospheric surface-layer physics, which is itself directly dependent on the instantaneous sea-ice condition provided by the sea-ice model. In earlier applications, the atmospheric drag on sea ice was computed from the local momentum transfer over ice. In the present study, this is replaced by a large-scale momentum flux, which is characterized by a large-scale stability function and a large-scale roughness length. The large-scale roughness length depends on the local skin drags and on the form drag, where the latter is given as a function of the ice-plus-snow freeboard and the ice concentration, both provided by the sea-ice model. The thermodynamic part of the calculation is given by the local fluxes, which depend on the local stability of the atmospheric surface layer. This, physically more reasonable, description of the largescale dynamic forcing generally leads to an increase of the momentum transfer via an increase of the roughness length and a decrease of the stability in the atmospheric surface layer. Finally, this yields improved model results, especially in terms of a more dynamic pattern of the ice-thickness distribution.

A dynamic and thermodynamic sea ice model for subpolar regions

A mesoscale dynamic thermodynamic sea ice model is presented. The model differs from the large scale sea ice model [Hibler, 1979; Parkinson and Washington, 1979] in both the dynamic and thermodynamic parts. In the dynamic calculation, a constitutive relation based on ice floe collision is used for the determination of the internal ice force in the Marginal Ice Zone (MIZ), while in the thermodynamic calculation, a three‐level approach for the ice thickness distribution is adopted instead of the common two‐level approach [Hibler, 1979]. The simulation results from the mesoscale sea ice model show that the three‐level approach is more sensitive to geophysical forcing than the two‐level approach. Thus it is demonstrated that the mesoscale sea ice model has the potential of a more accurate description of the dynamics and the thermodynamics than the previous models in the simulation of sea ice processes in subpolar regions. Giving the exact initial conditions and the atmospheric and oceanic forcings, this model can be used as a sea ice forecasting model in subpolar regions.

The Effect Of Rheology On Seasonal Sea-Ice Simulations

On the seasonal time scales relevant to numerical investigations of climate, the rheology used in large-scale sea ice models significantly affects the ice thickness build-up and ice velocity fields. Plastic rheologies with a normal flow rule have been used to-date in seasonal dynamic thermodynamic simulations. These rheologies have proved useful in simulating discontinuous slip near the coast while still supplying relatively robust velocity fields in the central Arctic Basin. However, as indicated by limited numerical sensitivity studies with different types of elliptical yield curves, the amount of shear strength significantly affects the ice build-up and can possibly cause a stoppage of the ice outflow through Fram Strait. In addition to the shear strength issue, there is also the possibility that non-normal flow rule rheologies, such as the Mohr Coulomb failure criterion used in soil mechanics, may cause somewhat different types of flow patterns, especially in the Fram Strait region. However, to date no seasonal simulations with such non-normal flow rule rheologies have been carried out. In order to investigate the role of different rheologies on the large-scale flow patterns in the Arctic Basin, a more general numerical scheme than that of Hibler (1979) is developed, which allows the simulation of the dynamic thermodynamic behavior of sea ice with a wide variety of different non linear rheologies. Using this numerical scheme, comparative simulations are carried out to seasonal equilibrium with several variations of the Mohr Coulomb rheology and compared to the more standard Elliptical yield curve results. In particular, the main control Mohr Coulomb case is a capped rheology in which the shear strength is taken to be proportional to the compressive stress. In this capped case only shear flow is allowed until a maximum allowable compressive stress is reached. This cap strength is parameterized to be a function of the ice thickness and compactness. For comparison, a simulation with a very large cap strength is also carried out, and an experiment with a similar compressive cap but much lower shear strength. Overall the results are analyzed to determine the sensitivity of the ice build-up to flow rule and shear strength magnitude. In addition special attention is given to the character of the flow and stoppage (if any) through Fram Strait.

The Effect Of Rheology On Seasonal Sea-Ice Simulations

On the seasonal time scales relevant to numerical investigations of climate, the rheology used in large-scale sea ice models significantly affects the ice thickness build-up and ice velocity fields. Plastic rheologies with a normal flow rule have been used to-date in seasonal dynamic thermodynamic simulations. These rheologies have proved useful in simulating discontinuous slip near the coast while still supplying relatively robust velocity fields in the central Arctic Basin. However, as indicated by limited numerical sensitivity studies with different types of elliptical yield curves, the amount of shear strength significantly affects the ice build-up and can possibly cause a stoppage of the ice outflow through Fram Strait. In addition to the shear strength issue, there is also the possibility that non-normal flow rule rheologies, such as the Mohr Coulomb failure criterion used in soil mechanics, may cause somewhat different types of flow patterns, especially in the Fram Strait region. However, to date no seasonal simulations with such non-normal flow rule rheologies have been carried out.In order to investigate the role of different rheologies on the large-scale flow patterns in the Arctic Basin, a more general numerical scheme than that of Hibler (1979) is developed, which allows the simulation of the dynamic thermodynamic behavior of sea ice with a wide variety of different non linear rheologies. Using this numerical scheme, comparative simulations are carried out to seasonal equilibrium with several variations of the Mohr Coulomb rheology and compared to the more standard Elliptical yield curve results. In particular, the main control Mohr Coulomb case is a capped rheology in which the shear strength is taken to be proportional to the compressive stress. In this capped case only shear flow is allowed until a maximum allowable compressive stress is reached. This cap strength is parameterized to be a function of the ice thickness and compactness. For comparison, a simulation with a very large cap strength is also carried out, and an experiment with a similar compressive cap but much lower shear strength. Overall the results are analyzed to determine the sensitivity of the ice build-up to flow rule and shear strength magnitude. In addition special attention is given to the character of the flow and stoppage (if any) through Fram Strait.