(part I: NPS overview by A. Semtner, W. Maslowski, and Y. Zhang)
Naval Postgraduate School
The Arctic and Global Modeling Group at the Naval Postgraduate School (NPS) is highly leveraged by funding beyond its base of NPS support. The Group is now funded by continuing grants from NSF Arctic System Science, NSF WOCE, NSF WOCE/Analysis, NASA's TOPEX/POSEIDON Extended Mission, and the DOE Climate Change Prediction Program. The hallmark of the Group's research is the development and validation of high- resolution coupled models of the Global or Arctic Ocean and sea ice, using observed high-frequency atmospheric forcing, plus satellite and in-situ data to evaluate the simulations and to guide improvements. The models are continually being reformulated in both physical and computational terms, often through associations with researchers at Los Alamos National Laboratory (LANL) and the National Center for Atmospheric Research (NCAR). Among physical improvements are a new elastic-viscous-plastic (EVP) ice rheology, as an alternative to the Hibler viscous-plastic rheology, and a K-profile parameterization (KPP) of the oceanic boundary layer. Better treatments of convection and transport are also being explored. The computational improvements include the use of a free surface, unsmoothed bathymetry, removal of the polar grid singularity, and reprogramming of both the ice and ocean models to use parallel machines efficiently, including machines that have distributed-shared-memory (DSM) architectures and use message-passing-interface (MPI) communication among processors.
Of the six persons in the NPS Group, Dr. Wieslaw Maslowski is the lead scientist for modeling the Arctic Ocean on parallel machines, and Dr. Yuxia Zhang is the lead scientist for sea-ice modeling and validation. In addition, Dr. Julie McClean works with Maslowski on global modeling using DSM machines; and Dr. Robin Tokmakian conducts WOCE modeling research and TOPEX data assimilation. McClean, Tokmakian, and Zhang collectively produce the forcing datasets, recently using the 1979-93 reanalyses of the European Centre for Medium-range Weather Forecasts and 1994-97 ECMWF operational analyses. Mr. Peter Braccio is the only programmer supported by the existing grants. Dr. Albert Semtner is overall leader of the modeling effort; and this includes coordination of NPS research with ocean and ice modeling efforts at LANL and NCAR.
Presently, two modeling thrusts are underway that can lead to an Arctic ocean-ice model suitable for operational Navy ice prediction. The first is an Arctic / subpolar North Atlantic ice-ocean model with 18-km grid spacing and 30 levels -- this version is being refined to have 9-km grid spacing and 40 levels, and it can be extended to the North Pacific as well. The second is a fully global ocean model with high resolution in the Arctic (averaging about 12 km with 32 levels, with projected grid refinement to 6 km in the Arctic and 40 levels) -- this will be coupled to EVP sea ice on the same grid, and it can be truncated at some northern latitude for Arctic-only work. Both coupled models are oriented toward the massively parallel T3E and emerging DSM architectures. An important consideration in choosing a version is the amount of machine time anticipated to run a forecast eventually in an operational setting (although we can plan to use DoD high-performance computing resources for validating both models in hindcasting mode). For this reason, the first model may be better.
Many contributions to model development will occur at no cost to ONR. The NPS Group is anticipating that a NOAA-supported postdoc will be assigned jointly to the Army Cold Regions Laboratory and NPS in order to improve representations of ice thermodynamics and multiple thickness categories for both the Arctic and the Southern Ocean. This is in view of the importance of improving thermodynamics as well as the interaction between dynamics and thermodynamics. Also, Navy students pursuing Masters degrees are now working on aspects of Arctic modeling and validation under the supervision of Maslowski, Zhang, and Semtner. One of those students is being detailed to FNMOC in July of this year. Finally, CDR Rost Parsons will become involved in Arctic research and resume the modeling studies that constituted much of his PhD research, upon his return to NPS as a Military Oceanography Instructor.
Naval Research Laboratory, Stennis
The present operational version of the U.S. Navy's sea ice forecasting system, the Polar Ice Prediction system (PIPS 2.0) is a coupled ice-ocean model used to forecast ice motion, thickness and concentration over a 120 hour period. The coupled model, which is the basis for the PIPS 2.0 system, consists of the Hibler viscous-plastic ice model coupled to the Cox ocean model. PIPS 2.0, developed in the early 1990's became operational in the mid-1990's. Recent advances in our basic understanding of sea ice as well as its interaction with the atmosphere and the ocean have lead to improved formulations for the processes that govern these interactions.
The goal of this proposal is to improve the U.S. Navy's existing sea ice forecasting capability by incorporating advances developed within the basic reach community during the past five years into the Navy's sea ice forecasting system.
It is proposed that these improvements be incorporated and tested in the existing coupled ice-ocean model developed at the Naval Postgraduate school by Dr. Albert J. Semtner's ice-ocean modeling team. The NPS team will incorporate/test these improvements in their existing Arctic/subpolar couple ice ocean model. This presently runs at higher horizontal (9km versus 25 km) resolution and higher vertical (40 versus 15 levels) resolution than the existing PIPS 2.0 model. The model also presently incorporates a more sophisticated version of the Cox-Bryan ocean model developed by Dr. Semtner. In particular the improvements to be tested in this proposed work are improvements to the ice rheology (anisotropic rheologies), and an improved treatment of fluxes at the air-ice- ocean interface. These improvement would be made in coordination/ collaboration with the 6.1 sea ice-ocean community.
These improvements will be both coded and tested and by the NPS modelling group. Once the new code exists, The NPS group, in conjunction with Dr. Ruth Preller's group at the Naval Research Laboratory will complete the 6.2 model development and testing. This will include developing the model to meet the operational needs of the user, the National Ice Center (NIC), and as such will be done in coordination with the NIC. In addition, both the NPS and NRL will work together to make sure that the numerical code is appropriate for the operational systems at the Fleet Numerical Meteorology and Oceanography Center. This implies that the code is optimized for scalable, MPI machines. The new model will then be transitioned to NRL for final 6.4 testing and validation as well as documentation. NRL will transition the new code into the PIPS 2.0 system as an upgrade to the existing Navy capability.