WP1b
Towards a Comprehensive 3-D Dynamic-Thermodynamic Ice Thickness Distribution Model
T.E. Arbetter, J.A. Curry and M.M. Holland
Program in Atmospheric and Oceanic Sciences, University of Colorado, Boulder, Colorado, USA
The intent of our work is to develop a comprehensive three-dimensional dynamic- thermodynamic sea ice thickness distribution model. Such a model would provide needed insight into the role of sea ice in the planet's climate and in climate change scenarios. As new and better parameterizations of physical processes become available, they can be incorporated into the model. Currently, we are incorporating better thermodynamics into a two-dimensional dynamic-thermodynamic model, including features like a distribution of sea ice thicknesses, a spectral albedo dependent on surface type and solar zenith angle, spatially and temporally varying snowfall, melt ponds, and a brine pocket parameterization.
Future work will add multiple thermodynamic layers in the sea ice to the model. The desire to include all of the best possible physics is countered by limits on computer resources. A model that is computationally expensive to run would not be suitable for inclusion in a coupled atmosphere-ocean GCM. Sensitivity studies using both the 2-D model and a one- dimensional sea ice thickness distribution model developed here help to indicate which parameterizations have the most impact on the simulation results. These results are useful in guiding the model's development.
wp1c
A Thermodynamic Sea Ice Model and the Effects of Kinematics on Sea Ice Redistribution
D. Salas y MÈlia
Centre National de Recherches MÈtÈorologiques, Toulouse, FRANCE
The development of a thermodynamic sea ice model was initiated in the framework of the "Ice State" European project at NERSC (Nansen Environmental and Remote Sensing Center, Bergen, Norway) and was completed at CNRM (Centre National de Recherches MÈtÈorologiques, Toulouse, France).
The main characteristics of the model are the following: the number of sea ice types can be chosen, as well as the number of layers when solving the vertical heat conduction equation in the ice and snow layer; a new scheme for the evolution in time of snow physical characteristics was used, and in the absence of snow, the ice slab is assumed to absorb part of the short wave incoming radiation, due to the thermodynamic effect of brine pockets. Finally, a statistical scheme giving an estimate of ice floe opening and closing was implemented and coupled to the thermodynamics; the whole model could reproduce theoretically explained features of sea ice thickness distribution functions.
WP1D
On the simulation of Antarctic sea ice in an AGCM
Xingren Wu, W.F. Budd, Ian Simmonds and Ian Allison
Antarctic CRC and Australian Antarctic
Division, Hobart, Tasmania, Australia
Antarctic CRC, Hobart, Tasmania, Australia
School of Earth Sciences, University of Melbourne, Parkville,
Victoria, Australia
Antarctic CRC and Australian Antarctic Division, Hobart,
Tasmania, Australia
A dynamic-thermodynamic sea ice model is developed and coupled with the Melbourne University atmospheric general circulation model to simulate the seasonal cycle of the Antarctic sea ice distribution. The model is efficient, rapid to compute and useful for a range of climate studies. The thermodynamic part of the sea ice model is similar to that developed by Parkinson and Washington (1979), which is based on energy conservation at the top surface of the ice/snow and the ice/water interface as well as the open water area to determine the ice formation, accretion and ablation. A lead parameterisation is introduced with an effective partitioning scheme for freezing between and under the ice floes (Wu et al., 1996). The dynamics contain a simplified ice rheology that resists compression, its calculation determines the motion of ice, which is forced primarily with the atmospheric wind, taking account of ice resistance and rafting at high ice concentration. The simulated sea ice distributions for concentration and thickness compares reasonably well with observations over most regions. The seasonal cycle of ice extent is well simulated in phase as well as in magnitude. When the distribution of prescribed mean ocean currents is included, the simulation of ice thickness is improved in the Weddell Sea. These results indicate the relative importance of ice advection in response to separate or combined wind and ocean drift forcing in the determination of the sea ice distribution.
WP1E
saline and fresh icing of hydrotechnical objects and structures
Anatoly R. Karev and L.G. Kachurin
Republic Hydrometeorolgical Institute of Macedonia, Skupi w/n, Skopje, 91000, MACEDONIA
A new approach to the solution of test body icing within a supercooled aerosol environment has been developed. The novelty consists of use of the dynamics of supercooled water skin motion on the icing surface as a relevant factor for the separation a few completely different icing regimes. Therefore, each of them has been determined by the regime of supercooled water skin motion. The dynamics of skin motion and the theory of heat- and mass-exchange are considered as two principal interdependent key factors, having an essential influence upon the intensity of icing process, its regime, the density of growing ice layer. The reaction of icing regime at the fluctuations of entrance parameters and the discrepancy in heat and mass balances are secondary important factors for the regimes classification.
The results obtained have been compared with the experimental wind tunnel data. The model of icing on different hydrotechnical objects has been constructed with use of the correlative connection between the results obtained from the theoretical modelling and the observations in natural sea conditions. Thus, the icing forecast in natural conditions has been established on the use of forecasting meteorological elements, defining the states of atmosphere and ocean as the entrance parameters in the obtained theoretical icing model.
This new approach in solving the problem of complex interaction among three phases (aerosol-liquid-solid body) has a wide applicability in solving of similar crystallization tasks in closely related physical fields.
WP1f
The use of a single column ice/ocean model and field data for the improvement of model parameterizations
Marika M Holland, Judith A Curry and Julie Schramm
Program in Atmospheric and Oceanic Sciences, University of Colorado, Boulder, USA
A single-cell ice thickness distribution model has been configured to track the evolution of an ice floe on a horizontal scale of 20-200 km (depending on the scale of the external forcing), corresponding to the aggregate scale of the ice thickness distribution. The ice thickness distribution model is coupled to an ocean mixed-layer model and an atmospheric radiative transfer model, so that the interfacial heat fluxes can be determined interactively. This model contains multiple vertical levels, a melt pond parameterization, a complex surface albedo parameterization and spectral radiative transfer. A mechanical redistribution function is included and sea ice is resolved by a number of classes characterized by various properties, including thickness, age and surface characteristics.
This model is useful for field programs such as SHEBA because it can be forced at the model boundaries using field data and run in data assimilation mode, assimilating observations such as parameters for individual ice types, and surface data obtained from aircraft. By utilizing the available observations in the context of an internally consistent ice thickness distribution model, the model can fill in observational gaps and provide internally consistent determinations of various diagnostics including the ice mass balance, snow mass balance, sea ice heat budget, upper ocean heat and salt budget, and disposition of shortwave radiation in sea ice and the upper ocean.
Based on sensitivity studies, various deficiencies and uncertainties in the model have been identified. These include the formation of new ice in leads, melt pond characteristics, and ridged ice heat exchange, among others. Observational data obtained during field programs such as SHEBA will be used to improve the model representation of these physical processes. This will allow for the formulation of parameterizations which are appropriate for incorporation into regional and global climate models.
WP1G
The role of snow ice formation in the seasonal cycle of the antartic sea ice cover
M.A. Morales Maqueda and T. Fichefet
Department of Mathematics, Keele University,
UK
Institute d’Astronomie et de Geophysique G. Lemaitre,
Universite Catholique de Louvain, BELGIUM
Snow ice is the ice that forms at the snow-ice interface when the snow layer becomes thick enough to depress the ice below the water level. Water from leads then infiltrates the submerged snow and freezes there, creating snow ice. Snow ice can therefore be regarded as consisting of two components: (1) the submerged snow that transforms into ice (meteoric ice) and (2) the frozen infiltrated water (seawater ice). The net effect of snow ice formation is to increase the ice thickness. Here we use a large-scale thermodynamic dynamic sea-ice model to assess the role played by this process in shaping the seasonal cycle of the sea-ice cover in the Southern Ocean. The snow ice parameterization distinguishes between the meteoric and seawater components of snow ice and takes into account that the heat conductivity of snow ice is generally smaller than that of black ice (ice formed by direct congellation of seawater). Sensitivity experiments with this model show that the net contribution of snow ice to the total volume of Antarctic sea ice greatly depends on the relative amount of meteoric and seawater ice in snow ice. Snow ice formation increases when meteoric (seawater) ice content decreases (increases). Observations indicate that Antarctic snow ice is made up of around 41% of meteoric ice and 59% of seawater ice. After entering these proportions in the snow ice parameterization, the model predicts that the snow ice contribution to the total Antarctic ice volume is slightly larger than 10%. The current uncertainties in the Antarctic snowfall rates and snow transport by wind have suggested further experiments in which the snow input is uniformly increased or decreased by a certain amount. The impact of these perturbations on snow ice formation is examined.
WP1h
A Regional High Resolution Coupled Ocean/Sea-Ice Model
Simon Marsland and J–rg-Olaf Wolff
Institute of Antarctic and Southern Ocean
Studies, University of Tasmania, Hobart, AUSTRALIA
Antarctic Cooperative Research Centre, University of Tasmania,
Hobart, AUSTRALIA
The seasonal sea-ice cover that advances and retreats around the coastline of Antarctica each year has a significant impact on the global climate system, through its moderation of the exchanges of heat, freshwater, and momentum between the ocean and atmosphere systems. The ice exists in a delicate balance with the underlying ocean waters. Interactions and feedbacks involving ice formation, brine rejection, mixed layer stability, deep convection, vertical mixing, oceanic heat flux, and ice melt all play a part in determining the evolution of ice thickness and compactness. Consideration of sea-ice dynamics adds to the complexity of the system. Near surface convergence and divergence leads to sources and sinks in the freshwater budget. The role of ice transport associated with the underlying mesoscale oceanic fluctuations is only poorly understood.
We are currently testing a coupled ocean/sea-ice model at mesoscale eddy resolution in a 1200 km long by 1200 km wide re-entrant (periodic) channel along the Antarctic continental shelf. The ocean model is based on the Hamburg Ocean Primitive Equation model (HOPE). A modified Hibler-type sea-ice model is coupled to the HOPE model and allows prognostic calculation of sea-ice thickness, concentration and velocity. This is the first time a coupled ocean/sea-ice model will be used in the Southern Ocean at eddy resolution.
The aims of this study are to be able to numerically reproduce major features of the ocean/sea-ice system, and then to use the model as a tool in determining properties of this system such as the spatial-temporal variability in oceanic heat flux, its effect on thermodynamic ice growth, and the influence of ice and mesoscale ocean dynamics on the budgets of heat, salt and momentum at the ocean surface. It is expected the model will also be useful for making estimates of eddy contribution to poleward heatflux in the Southern ocean, and identifying deepwater formation mechanisms and rates.
WP1I
assimilation of ice motion vectors into a 2-d sea ice model
Walter N. Meier and James Maslanik
Program in Atmospheric and Oceanic Sciences, University of Colorado, Boulder, CO, USA
Assimilation of remote sensing data into sea ice models shows promise in providing insight into the determination of sea ice thickness and transport in the Arctic Ocean. Daily varying grids of ice motion vectors are assimilated via a weighted blending method into a two-dimensional, Hibler-based sea ice model. The ice motions are calculated from SSM/I and AVHRR remotely-sensed data and in situ buoy data and are optimally interpolated to the Polar Pathfinder grid. Six years of daily motion fields are assimilated into the model. Coincident daily NCEP reanalysis parameters are used as forcing fields for the model. Effects of systematic and random errors in the assimilated motion vectors on model outputs are investigated. Blending weights are varied to determine calculations. Ice thickness and transport for each of the six years are calculated for model runs with and without assimilation. Differences in seasonal and interannual variability of the ice thickness and transport are investigated. Spatial differences across the Arctic Basin between the assimilated runs and standard, model-only runs are also noted.
WP1J
validation of kinematics in a sea ice model
Matti Lepparanta and Jari Haapala
Department of Geophysics, University of Helsinki, Helsinki, FINLAND
Numerical models for sea ice thickness distribution and velocity are used for ice dynamics research and ice forecasting. In the modelling work ERS-1 SAR is an excellent tool, in particular providing spatial ice velocity fields. ERS-1 SAR produced ice kinematics (3-day displacements) are examined here with a Hibler type viscous-plastic sea ice model. A considerable stiffening of the ice pack was observed due to the change in the character of ice deformation under compression from rafting to ridging as the minimum ice thickness increased from 10 to 30 cm. The coastal alignment was strong in the ice motion and the coastal boundary layer width was 20-30km. An analysis of the SAR data with an ice dynamics model showed that the observed overall ice velocity field could be produced with a plastic ice rheology. The compressive strength of the ice (in 10-km scale) became 25,000 N/m2 for ridging and negligible for rafting of very thin ice. The 3 day repeat cycle is good for updating an ice model but for detailed ice dynamics investigations a data frequency of one cycle per day or higher would be preferable.
WP1K
high resolution simulation of arctic cryospheric anomalies
A.H. Lynch and J.A. Maslanik
CIRES, Campus Box 216, University of Colarodo, Boulder, CO, USA
Recent global climate modelling results highlight the Arctic as a region of particular importance in global climate change. Numerous observational studies have demonstrated the sensitivity of sea ice and snow to changes in regional climate and larger scale modes of atmospheric circulation. However, there remain strong biases in GCM simulations, despite dramatic improvements in recent years. The goal of the current work is to identify the mechanisms responsible for recently observed trends and anomalies in the Arctic cryosphere, in an effort to elucidate the most important mechanisms involved and hence determine the crucial processes missing from GCM simulations of the Arctic climate under conditions of change. Initial simulations of regions where trends and anomalies have been identified have been performed using the Arctic region climate system model, and will be described in this presentation.
WP1l
The seasonal cycle of Antarctic sea ice in a high resolution climate model
Siobhan. P. O'Farrell
CSIRO Division of Atmospheric Research, Aspendale, AUSTRALIA
The coupled ocean-ice-atmosphere climate model used at CSIRO for climate change simulation, which includes a full dynamic thermodynamic ice model with cavitating fluid rheology, underestimates the seasonal cycle of ice extent in the Southern Ocean. This is partly due to poor resolution, since the atmospheric model operates at a spectral resolution of R21 with the ocean and ice components on a matching grid (5.6 degree longitude by 3.2 degree latitude).
Also during the model development greater emphasis was given to the Arctic sea ice simulation which was highly sensitive to warming on the neighbouring continent and many of the parameters chosen for the ice model (e.g. lateral melt rates, mixed layer depths, surface albedos) may not be optimal for the Antarctic sea ice. A further problem is the steepness of the topography off East Antarctica which leads to a Gibbs phenomena (spectral overshoot) in the sea ice zone. Though the air temperatures are corrected back to sea level they are too cold even in summer to cause surface melt and hence the ice retreat is insufficient.
On the other side of the continent the build up of ice in the Weddell sea along the Antarctic Peninsula is not correctly modelled. The ocean model at R21 resolution does not resolve the Weddell Gyre and in the atmospheric model the Antarctic peninsula does not provide sufficient blocking of the flow to drive a realistic circulation which could lead to thicker ice along the coast.
By setting a requirement that the model retain ice in the Weddell sea throughout the year, there is too much ice left in other sectors in summer. Namely, if the ocean to ice heat flux is increased to thin the ice next to East Antarctic coast then the Weddell sea will become ice free. Hence, a compromise is made to allow excess ice to remain in late summer/early autumn around Antarctica in the CSIRO coupled model simulations which is then taken into account in interpreting the climate change results. However, for the next climate model generation, improvements in the seasonal cycle are required. Alterations to the physical processes driving the ice need to be assessed not only for improving the climatology of an atmospheric/sea ice model under current climate conditions, but for the implications of these changes in the calculation of flux corrections in the ice covered regions of coupled models, and also the sensitivity of these parameterizations to changes in climate conditions.
This paper outlines the improvements in the Antarctic sea ice seasonal cycle that have been achieved using a R42 horizontal resolution model (twice the earlier simulations), where the ocean currents and atmospheric winds are more realistic. Other changes which have been implemented include increasing the temperature dependent component of the ocean-ice heat flux as well as the background threshold value; altering the mixed layer depth in the ocean, the rates of lateral melt, the wind turning angle, and including a reduced albedo for water-logged ice in the melt season.