OMEGA

High Resolution Weather Forecasting
For 50 years, predicting the weather has been one of the grand challenges of numerical simulation. This problem is incredibly difficult for many reasons: it involves scales of motion that vary by many orders of magnitude in space and time, it involves complex physical processes only some of which we completely understand, we lack basic data on the surface of the Earth that affects the interaction of the atmosphere with the surface, and we lack data on the atmosphere itself. In spite of all of these issues, weather forecasting has steadily advanced for five decades.

For over 15 years SAIC has been developing an advanced numerical weather prediction system. The Operational Multiscale Environment model with Grid Adaptivity (OMEGA) is based upon an adaptive unstructured triangular mesh that facilitates the placement of additional spatial resolution anywhere required. The basic philosophy behind OMEGA development has been the creation of an operational tool that can support both weather forecasting and predicting the dispersion of hazardous materials such as from a nuclear reactor accident like the Chernobyl disaster, widely known as the worst nuclear power plant accident in history.

The flexibility of the OMEGA grid structure allows the creation of grids that have high resolution at diverse specific locations (see below). This allows an improvement in forecast accuracy by resolving both the physical processes in the region of interest at higher resolution and the surface properties that impact the weather (e.g., elevation, land use, land/water fraction, vegetation, albedo, soil type) at higher resolution. It also allows OMEGA to put the computational power where it is needed – for example forecasting specific areas such as airports or tracking a storm such as a hurricane.

OMEGA grid placing high resolution around the 25 busiest airports in the US.

Hurricanes represent one of the most severe forms of weather. Their tremendous impact derives from the intensity of the wind speed, the magnitude of the precipitation, and the duration of the storm. In spite of the tremendous affect of these storms, however, the current operational accuracy of forecast systems has been considerably less than that desired. In 1998, the city of New Orleans was evacuated because of a forecast for Hurricane George that turned out to be in error due to a last minute shift in the storm track to the east.

Had information from SAIC's OMEGA been available at the time, different steps might have been taken. A historical forecast produced by OMEGA for Hurricane George shows the OMEGA surface pressure field (color and contours) along with the observed storm location. The 24 hour and 72 hour forecasted positions are shown below.

OMEGA forecasted pressure for Hurricane George (1998) valid on September 26 (left) and September 28 (right). The actual storm track is shown as dark blue dots.

SAIC continues to improve and develop OMEGA for a host of government and commercial applications. We believe that it can significantly improve operations in the transportation, energy, oil and gas, and emergency response sectors of the marketplace.