Unstructured Grid Antarctic Weather Forecast System

Weather forecasting in Antarctica is essential for US Antarctic Program (USAP) operations. However, forecasting in this region is extremely difficult both because the observational and modeling tools available are inadequate and because our understanding of the science behind the meteorology of Antarctica is incomplete. The weather of Antarctica is dominated by three factors: (1) the polar high and the baroclinic waves that circumnavigate the pole; (2) the terrain forcing that arises from the interaction of the complex topography and dynamically changing surface features with the larger scale flow; and (3) the katabatic processes that cool the near surface atmosphere at the higher surface elevations which leads to flow acceleration downslope. In order to forecast in this environment, it is essential that the models used resolve all of the important temporal and spatial scales of motion and include all of the important physical processes. Improving the state of weather forecasting for USAP operations requires both improving the models and transitioning this capability to operations. Improving operational forecasting involves a system of systems, each of which must be considered in order to improve forecast skill: (1) data acquisition; (2) data ingest, quality control, and assimilation; (3) physical modeling; (4) visualization; and (5) results dissemination. For USAP operations, the basic lack of in-situ (only 32 surface observation and 8 rawinsonde stations exist) and remotely sensed (only partial satellite coverage exists and only during selected time periods of the day) data and data bandwidth are over-riding constraints that must be considered. Several groups are proposing new observation systems (ATOVS, COSMIC) that will add much needed data to the system, but this new data must be ingested, quality controlled, and assimilated into the forecasting systems. At a recent NSF sponsored Antarctica Weather Forecasting Workshop (Bromwich, private communication) several themes emerged. The themes included: • A need for more observational (in-situ and remotely sensed) data; • A need for higher spatial resolution weather forecast models; • A need to improve the physical parameterizations in forecast models; • A need for field data to develop the improved parameterizations of surface properties and for model evaluation; • Increased interaction between the research and operational communities; and • Improved model guidance for the operational forecasters. Most of the existing operational weather forecasting systems lack the ability to adapt to the underlying terrain as is required to simulate the highly scale interactive environment that exists in Antarctica. The Operational Multiscale Environment model with Grid Adaptivity (OMEGA) is a high resolution, high fidelity, operational weather forecasting system (Bacon et al., 2000) based on an adaptive unstructured grid. OMEGA is a non-hydrostatic multiscale forecast model that has been used to forecast extreme or severe meteorological events from global scale to local scale as well as point and large area dispersion phenomena. SAIC has been exploring the utility of OMEGA for forecasting in Antarctica.