Inside this Issue

• Near-Future Ocean Modeling
• Art in Science
• Designing New Materials
• Summer Interns
• ARSC Cadet Program
• Expanding our Horizons
• Cray SX-6
• Modeling the Arctic Ice
• ARSC Training Workshops

Modeling the Arctic Ice

Story by Jenn Wagaman

Computational scientists must always balance the need for more spatial definition with the processing capabilities of current technology. The accuracy of predictive models often increases as a model’s detail increases, but higher resolution requires more powerful computers in order to complete the computation in a reasonable amount of time. As supercomputers become faster and able to handle larger datasets, scientists are able to push towards more accurate prediction models. However, resolution is not the only factor in reliable prediction. Researchers also need accurate models that properly reflect natural phenomena and which are supported by solid underlying physics.

Recent advances in ocean and ice modeling in the Arctic have suggested that major changes in ice/ocean circulation have occurred over the past two decades. In order to make reliable predictions about future environmental changes, in-depth understanding of more recent variability in the region is imperative. Dr. Wieslaw Maslowski of the U.S. Naval Postgraduate School uses ARSC resources to run models of sea ice characteristics that contribute to the effort to model the arctic region accurately enough to make reliable predictions. Maslowski’s recent work has led him and his team to conclude that existing global climate model predictions have large errors due to both insufficient model resolution and “missing” physics.

Computational studies of environmental change suggest that the Arctic is highly sensitive to greenhouse warming, largely because of positive feedback associated with thinning and retreat of the insulating, highly-reflective sea ice cover. The problem is that sea ice models used in such studies typically have been run at fairly coarse resolution, using crude parameterizations of the thermodynamic and dynamic processes that determine the ice thickness and extent. Maslowski and his team are looking for ways to improve sea ice simulations by combining the use of recent improvements to sea ice thermodynamics and higher resolution in the models. They are taking advantage of supercomputing technology to improve the fidelity of Arctic Ocean simulations and thereby increase the reliability of environmental change predictions.

The researchers developed a coupled ice-ocean model configured at a nine-kilometer and forty five-level spherical grid for the sea-ice-covered northern hemisphere. The model was integrated for over seventy years using European Centre for Medium-Range Weather Forecast (ECMWF) reanalysis data and operational products. In addition, a multi-category ice thickness distribution model was developed using resources at the U.S. Army Engineer Research and Development Center and integrated in order to study improvements resulting from more realistic thermodynamics. The results of these two models were compared with observational data to make determinations about how the changes are improving the models.

To more accurately represent sea ice response to wind-stress forcing, the new model also includes a modified parameterization of ice strength that depends on the amount and thickness of thin ice. Several studies have shown that ice strength, heat exchange with the atmosphere, and other properties affect ice thickness distribution, especially the amount of thin, first-year ice. As a result, the new model also gives improved ice deformation and drift fields.

Results from the two nine-kilometer resolution and one 18-kilometer resolution regional models were analyzed to determine the improvements resulting from the ice dynamics/thermodynamics and resolution changes. To differentiate the effects of the new ice model from those with increased spatial resolution, the researchers first compared sea-ice drift in two coupled ice-ocean models of the Arctic Ocean, both using the “old” ice model but at differing resolutions. The coarser model showed that the mean sea-ice drift changes its circulation over a ten-year period. The difference between the two mean ice velocity fields illustrates a basin-wide cyclonic trend. The high-resolution model results suggest that the general, large-scale ice circulation and the decadal shift between the two means are similar to results from the older model. This analysis showed that increased spatial resolution indirectly improved the model results through the coupling of a more realistic ocean model with the sea ice. Although the old sea ice model had a reasonable ice thickness distribution, it was lacking in accuracy of velocity and deformation fields. The improvements of the thermodynamics of the new sea-ice model, including the multiple ice thickness category approach to represent the ice thickness distribution, allow the ice strength to depend on the amount and thickness of the ice, instead of only the mean ice thickness.

With the help of these improvements, the new sea ice model produced more accurate ice deformation and drift fields. The long narrow features in ice divergence and shear fields resembled those observed in synthetic aperture radar (SAR) data, except that the features’ average width was over-estimated, possibly due to insufficient horizontal resolution.

Based on the results of the model runs, the researchers concluded that if the strength of the sea ice primarily depends on the amount and thickness of thin ice, then the new multi-category ice model is much more accurate than the old, two-category ice model in representing sea-ice deformation fields. The coupling of the new sea-ice model with the three-dimensional ocean model will allow a more in-depth look at changes in ice/ocean circulation and characteristics that have occurred over the past two decades.

Accurate models of the arctic environment are important to the U.S. Navy for a variety of reasons, including the possibility of establishing a Northern Sea Route from the Pacific to the Atlantic as an alternative to the Panama Canal. Such a scenario could drastically change the role of the Arctic Ocean in oceanic travel. In addition, increased commercial and defense activity in a partially ice-free Arctic would set new requirements for the U.S. Navy and Coast Guard, as well as change the way of life for Native communities across the Arctic. The ability to carry out these kinds of activities requires an understanding of the scenarios of future environmental change in the Arctic.

Return to Challenges index.