C41C-0206 0800h
Porosity-Permeability-Salinity Relationships in First-Year Arctic Sea Ice
Despite the importance of sea-ice permeability in controlling mass and heat fluxes through the ice cover, few measurements of ice permeability (as a function of porosity) have been completed to date. Here, we report on results of ice permeability measurements based on in-situ borehole slug tests. Data were obtained from different depth levels of homogeneous Arctic first-year landfast ice near Barrow, Alaska throughout the ice seasons of 1999 to 2001. Thermistor arrays frozen into the ice and analysis of ice cores and large slabs cut through the entire ice thickness provided anicllary data on sea ice thermal, salinity and microstructural evolution. Permeabilities ranged between 3 x 10$^{-13}$ m$^{2}$ (brine volume fraction 0.033) and 4 x 10 $^{-10}$ m$^{2}$ (brine volume fraction 0.24), with a sharp increase towards the base of the ice cover. Permeability-porosity relationships appear to fall into two regimes: (1) a two orders of magnitude increase in permeability for the 0.03 to 0.09 brine volume fraction interval, and (2) a one order of magnitude increase for brine volume fractions $>$0.09. However, interpretation of these regimes is hampered by significant scatter in the data due to measurement errors and natural variability. Examination of pore micro-/macrostructure and time series of vertical salinity profiles can provide additional insights into the co-evolution of ice permeability and porosity.
C41C-0207 0800h
Thermal dependence of brine salinity in the surface layer of snow-covered sea ice under the variable conditions
We observed snow covered sea ice in Barrow, Alaska in February 2004 to investigate thermal dependence of brine salinity in the surface layer. The observed data was compared with simulated data, which was estimated by existing thermal models (Nakawo and Sinha, 1981; Ono, 1968). The brine salinity of the surface layer is very important to analyze microwave images, because its dielectric properties determine emissivity and penetration depth. In combination with the observation, we took some sea ice cores, measuring their length, temperature and salinity. After the observation, we obtained vertical temperature and salinity profiles. We also measured air temperature, snow surface temperature, snow/ice interface temperature, water temperature and snow depth and density. These data were used for the simulations. Regarding the simulation, at first, we estimated snow/ice interface temperature using the equation of Nakawo and Sinha (1981). This equation is formed by 6 parameters, which are the thermal conductivity of snow and sea ice, snow depth, sea ice thickness, air temperature and melting point of sea ice. And then, we estimated brine salinity using the results of Nakawo and Sinha and the equation of Ono (1968). This equation for obtaining brine salinity is calculated as a function of sea ice temperature. When we simulate brine salinity, only four data points are needed, because melting point and thermal conductivity of sea ice can be assumed as -1.8$\deg$C and 2 W m$^{-1}$ K$^{-1}$, respectively. As for the thermal conductivity of snow, we can obtain from the equation of Devaux (1933). This equation requires information on snow density. Namely, we need air temperature, snow depth, snow density and thickness of sea ice when we simulate brine salinity. This means that field work and calculation relevant to simulation becomes more easily. Comparisons between observation and simulation have indicated good correlation. This result suggests the applicability of the simple simulation method. Based on these positive results, we calculated the relation between snow/ice interface temperature and snow depth as a function of air temperature. Specifically, we varied air temperature between -10$\deg$C, -20$\deg$C and -30$\deg$C. For a sea ice thickness of 1.25m and 2m, and an air temperature of -10$\deg$C, snow/ice interface temperature is almost same. On the other hand, when sea ice thickness was 30cm, snow/ice interface temperature became almost same at each air temperature if snow depth was over 50cm. And also, mean snow/ice interface temperature was higher than for the other two cases. We also calculated brine salinity in the surface layer of snow covered sea ice as a function of air temperature using the simulated results of snow/ice interface temperature. The variable range of air temperature is same as case of simulation for the snow/ice interface temperature. There was proportionality relation between sea ice thickness and brine salinity in the surface of sea ice, but relation between snow/ice interface temperature and brine salinity indicated inverse proportion. As indicated above, we compared observed data with simulated results to investigate the thermal dependence of brine salinity in the surface layer of snow covered sea ice. As a result, observed data and simulated results have indicated good correlation. We can estimate the dielectric constant of snow covered sea ice in the surface layer by simulating brine salinity under the various conditions. Hereby, we can expect an improvement of the accuracy of observation by satellite microwave remote sensing.
C41C-0208 0800h
A Contribution to the Sea-Ice Model Intercomparison Project 2 (SIMIP2) Thermodynamic Sea-Ice Models Assessment : Impact of Vertical Resolution, Vertical Salinity Profile, and Salinity- and Temperature-Dependent Thermal Properties
Results from a sophisticated thermodynamic sea-ice model are presented. The model has the following peculiarities. First, the sea-ice pack is represented by several layers of snow on top of several layers of ice, and different ways of specifying the number and thickness of vertical layers are possible. Second, the heat-diffusion equation and its boundary conditions explicitly take into account the thermal damping effect of the brine pockets trapped into the ice through temperature- and salinity-dependent thermal properties, but density is nevertheless a constant. The main experiment consists in an almost-one-year integration of the model following the SIMIP2 (Sea-Ice Model Intercomparision Project) specifications, i.e., a Lagrangian multiyear undeformed ice floe forced by observed hourly atmospheric and monthly oceanic data fields. Model results are compared to weekly observations of ice and snow thicknesses. Snow precipitations have to be increased by 50% for the snow thickness to be correctly reproduced, supporting gauge problems already reported in the past. Ice thickness shows good overall agreement, especially in winter, and an underestimation of around 10 cm of surface melting in summer. Sensitivity experiments show that the effect of salinity and temperature through thermal properties is significant. The non-linear vertical salinity profile can be replaced by a linear profile with same mean salinity. Finally, the simplest resolution which conserves the main features of the results gathers one layer in the snow and between 5 and 10 layers in the ice.
C41C-0209 0800h
Evaluation of Efficiency and Impact of a Hierarchy of Parameterizations of Temporal Evolution of Sea-Ice Salinity Vertical Profile in One-Dimensional Thermodynamic Models
The results and impact of three different parameterizations of temporal evolution of the vertical salinity in a one-dimensional energy-conserving sea-ice model are assessed. The first and simplest parameterization (1) gives a time-independent bulk salinity in terms of ice thickness and age of the ice. The second one (2) simulates the evolution of the bulk salinity according to brine-drainage processes that are chosen empirically, assuming a vertically constant salinity for first-year (FY) ice and a linear profile of salinity for multi-year (MY) ice. The last one (3), the most complex, physically computes the temporal evolution of the non-linear vertical salinity profile (based on Cox and Weeks). The energy conserving sea-ice thermodynamic model has one layer of snow on top of several layers of sea ice, with thermal properties depending on salinity and temperature in order to take into account the thermal damping effect of brine pockets on heat transfer in the ice. The model is run with each of the three parameterizations, with NCEP-NCAR atmospheric forcing coming from two regions of the Arctic area. These two regions respectively correspond to FY and MY ice types. Salinity bulk values and profiles are compared with available observations, and impact on the thermodynamic characteristics of the ice is assessed. The analysis show that 1) parameterization (1) cannot reproduce intense desalination in summer and leads to an excessive ice thickness through overestimated surface thermal inertia in summer; 2) bulk salinities simulated by (2) are in agreement with observations, with rapid desalination for FY ice due to intense initial drainage, and a MY ice seasonal cycle, and; (3) if the scope of a study does not go beyond ice thickness and temperature profile in the ice , then, (3) does not improve the simulation compared to (2).
C41C-0210 0800h
A Simple Object-Oriented Framework For Making Sea-Ice Models ``Plug-and-Play''
Historically, sea-ice models have developed incrementally and in a decentralized manner, with changes and improvements authored by research groups all over the world. Models tend to be ``one-of-a-kind,'' limiting their modularity. Current efforts in making models more modular have focused upon establishing comprehensive programming standards. Comprehensive interface standards, however, tend to be quite complex. Given most climate scientists are not computer scientists, there is a need for approaches that are simple and intuitive, while also being powerful enough to include many different levels of model sophistication. In this presentation we describe efforts in using the Python language to develop such a framework for sea-ice models. Python is an open-source, object-oriented, interpreted language that is also able to interface tightly with compiled libraries. This results in programs that are easier to develop, more maintainable, and less fragile, while simultaneously reaping the performance benefits of compiled code. The framework is applied to Semtner's (1976) simple thermodynamic model as well as the NCAR CSIM (Schramm et al.\ 2004). We describe the framework, its advantages, and how the incorporation of such ideas in future sea-ice models can result in a suite of ``plug-and-play'' models.
http://www.johnny-lin.com/presn.html
C41C-0211 0800h
The Stability of Landfast Sea Ice: What Makes it so Fast?
Arctic landfast ice is shaped by a range of complex dynamic and thermodynamic sea ice processes that do not exist in the deep Arctic Ocean. The state of the landfast ice cover at any time depends on the history of conditions earlier in the year. Although thermodynamic processes largely determine the growth of stationary ice, the thickness of level landfast ice at the end of the season depends upon the duration for which the ice cover has been stationary. Here, we present and combine the results of three methods of observing the landfast sea ice in Alaska across different scales and addressing the mechanisms that hold it fast to the coast and determine the duration of its quiescent growth period. At the regional scale, Radarsat Synthetic Aperture Radar (SAR) imagery has been used to delineate the seaward landfast ice edge (SLIE) 31 times through the 2001-02 ice season along the northern Alaskan coast. At this scale, the SLIEs resolve themselves into nodes of stability separating regions of greater variability. At the local scale, a land-based side-looking X-band marine radar has been employed to scan landfast ice within approximately 5 km of Barrow, Alaska every 5 minutes. This high temporal resolution captures ice dynamics during the creation of grounded ridges. These ridges can be built in less than an hour and remain for the entire season, before breaking up and drifting away as swiftly as they arrived. At the highest spatial resolution, a Differential Global Positioning System (DGPS) and an electromagnetic induction device have been combined to measure sea ice elevation and thickness nearly continuously along transects within the land-based radar footprint from the beach at Barrow to the SLIE. By combining these data we investigate the relationship between the location of stable nodes along the SLIE and the local bathymetry and coastal configuration. We also examine the processes responsible for creating the nodes in terms of ice motion and deformation. Finally, we look at the possible role that such nodes can play in stabilizing the surrounding landfast ice.
C41C-0212 0800h
Effects of Boundary Layer Shear on the Morphology of Sea-Ice
Shear flow in the sub ice boundary layer has been shown theoretically to induce a Bernoulli pressure variation which drives a morphological instability of the ice/liquid interface. We investigate this process through an analogue to sea-ice system; the controlled solidification of an ammonium chloride solution in a laboratory flume. We observe the growth of "sea-ice" in the presence of a range of geophysically realistic flow speeds and find a threshold speed above which a spatiotemporal variation of the phase fraction of the layer appears. Upon removal of the external flow, the material returns to a uniform state. We describe the systematics of the instability and its importance to the formation of sea ice and associated salt fluxes.
C41C-0213 0800h
Internal Melt of Ridge Keels
Ridged ice forms a large fraction of the total volume of pack ice in the Arctic Ocean. Therefore, a realistic parameterization of the physical processes maintaining ice ridges is essential for accurate modelling of pack-ice evolution. However, the deterioration of ice ridges by melting is poorly understood. Observations of ridged sea ice during the melt season reveal that the thinning of ridges is much more rapid than the melt of surrounding level ice. We suggest that this enhanced melt is associated with the porous structure of the ridges, which allows warm oceanic water to penetrate the ridge permitting melting over a very large surface area relative to volume. To explore the effect of internal melt on ridges, a simple model of porous flow through the keel and the accompanying heat transfer has been developed. Using observations of pack-ice draft in the Beaufort Sea, a draft-dependent melt rate for a population of ice ridges can been calculated based on upper ocean temperature, relative ice-ocean velocity and geometrical parameters of idealized ridges. Predicted internal melt rates may be an order of magnitude greater than that of level ice and are found to dominate the loss of ice volume through ablation. The estimated rate of ridge deterioration using this model is comparable to the rate of ablation derived from observations of ice draft ablation by moored ice-profiling sonar. A model for ice-draft redistribution has been configured to simulate local conditions in the Beaufort Sea where data on ice draft are available. Running the model with and without internal melt has demonstrated that the porous flow through ridged ice and accompanying heat transfer can account for the enhanced rates of melt that are observed. Simulations that do not incorporate the internal melting of keels greatly overestimate the volume of the thick ice. Internal melt of ridged ice is thus an important element in the continual re-distribution of pack-ice volume. Important factors include the temperature of under-ice boundary layer and the size and packing of the blocks that form ridge keels. Clearly, accurate simulation of ridge melt processes will require more detailed information about the ice cover and ocean boundary layer than has been previously thought necessary.
C41C-0214 0800h
Heat Flux Comparison Using Buoy- and SAR-derived Motion Products From ISW 1992
Sea-ice kinematics relevant to surface fluxes using ERS-1 SAR images coincident with buoys in the western Weddell Sea in Austral Autumn of 1992 is examined. Using a toy model, effects of aliasing in surface flux determination are tested. Results show variability associated with storms, ocean tides, inertial oscillations, and other high frequency forcing affects integrated sea-ice growth rates along this shelf/slope location. Integrated salt and new ice production rates computed from buoys are found to be two times larger than those using ERS-1 SAR motion products. Cognizant of the limitations in satellite image pairs separated by time, we report on differences in salt and ice production rates, it follows directly that the differences in salt and ice production rates result primarily from inadequate temporal resolution of shorter than daily (sub-daily) heat flux variability and sea-ice divergence. Comparison with other studies shows the problem is widespread thereby impacting the modeling of sea-ice mass balance and variability. These small-scale processes have significant ramifications to larger scales and the global thermohaline circulation.
C41C-0215 0800h
Comparison of the Ice Production and Thicknesses in the Chukchi Sea Polynyas Derived From AMSR-E and SSM/I, and its Implication for Other Regions
For January-March 2003, we use ScanSAR, SSM/I and AMSR to examine the behavior of two polynyas, one large and one small, that occur along the Alaskan Chukchi coast. The large polynya, called Chukchi, forms between Cape Lisburne and Point Barrow; the smaller polynya, called Lisburne, forms between Point Hope and Cape Lisburne. Within these polynyas, the thin ice thickness distributions are derived daily from the ratio of the SMM/I 37 GHz and AMSR 36 GHz brightness temperatures; the heat fluxes from a combination of the thicknesses with meteorological data. Because the AMSR channels have a daily and seasonally dependent drift in calibration, the AMSR 36 GHz channels must first be corrected against the SSM/I 37 GHz channels. Given that the AMSR 36 GHz resolution is 12.5 km, compared with 25 km for the SSM/I 37 GHz channel, the AMSR has two advantages, a better landmask and an improved resolution. For days with active polynyas, comparison of the passive microwave data sets against ScanSAR imagery shows that AMSR provides a better definition of the polynyas. Examination of the daily and cumulative heat loss time series shows that for the Chukchi polynya, the AMSR yields a greater total heat loss than the SSM/I, while for the Lisburne polynya, the AMSR permits measurement of its productivity and resolution of its size even when the polynya size is of order of a single SSM/I pixel. Finally, in a brief example, we will show that in the southern hemisphere, the combination of ScanSAR and passive microwave permits delineation of the ice shelf boundaries and thus allows for a more precise investigation of the Ross Sea polynyas than in previous work.
http://polar.ocean.washington.edu/
C41C-0216 0800h
Gamma Function Parameterization of Sea-Ice Thickness Distribution
Antarctic ship observations collected from 1980 to the present are used to test the parameterization of sea-ice thickness distribution using a composite of Incomplete Gamma Function modes. The data are partitioned seasonally and regionally into five sectors around Antarctica. Using a maximum likelihood method, the input data used to construct a probability distribution are sorted to isolate thickness modes with the thinnest incorporating open water fraction. Each mode is mathematically described using the Incomplete Gamma Function with three parameters of shape, scale, and shift. These mathematical variables are related to the shape of the probability for each mode including mode thickness value, thickness range, probability distribution value, and mean thickness. The goal is to provide a simple parametric description of regional thickness distribution that can be imported into numerical models in a concise compact description with a physical relationship relating the mathematical parametric values to the data structure.
C41C-0217 0800h
The Seasonal Evolution of Sea Ice Floe Size Distribution
The Arctic sea ice cover undergoes large changes over an annual cycle. In winter and spring the ice cover consists of large, snow-covered plate-like ice floes, with very little open water. By the end of summer the snow cover is gone and the large floes have broken into a complex mosaic of smaller, rounded floes surrounded by a lace of open water. This evolution strongly affects the distribution of the solar radiation deposited in the ice-ocean system and consequently the heat budget of the ice cover. In particular, increased floe perimeter can result in enhanced lateral melting. We attempt to quantify the floe evolution process through the concepts of a floe size distribution and a breaking function. A time series of aerial photographic surveys made during the SHEBA field experiment was analyzed to determine evolution of the floe size distribution from spring through summer. Based on earlier studies, we assumed the floe size distribution could be represented by a two-parameter power law F(D) = a D$^{-b}$, where D is the floe diameter and a and b are the two parameters. The size of small floes could be measured directly from the images. For larger floes they were truncated at the edge of the photographs, measurements of total ice area and perimeter were used to estimate two parameters of the size distribution. As summer progressed, there was a decrease in the magnitude of the exponent as the size distribution shifted towards smaller floes. We are developing simple parameterizations of the evolution of the floe size distribution that can be incorporated into large-scale sea ice models.