SAIC Algorithm Improves Weather Forecasting
When a hurricane forms, protecting life and property depends in part on accurate storm track forecasting. SAIC has responded to this need by developing OMEGA, a high-resolution atmospheric modeling system.
The OMEGA system represents a significant departure from traditional weather prediction methods because it melds modern grid generation technologies from the computational fluid dynamics realm with the state-of-the-art numerical weather prediction techniques (which use mathematical models to predict weather).
Enabling more accurate forecasts
OMEGA dynamically adapts its computational grid during a simulation to provide high resolution in regions where storms can potentially develop, and it resolves important features of atmospheric circulation and cloud dynamics as they appear.
In fact, OMEGA can provide continuously variable resolution from 100 kilometers over the oceans to less than 10 kilometers over other regions of interest. This flexibility helps enable more accurate forecasts in a computationally efficient manner.
The hydrodynamic solver is key
A central element of OMEGA is its hydrodynamic solver. SAIC researchers applied the solver — multidimensional positive definite advection transport algorithm (MPDATA) — to the unstructured triangular grids that OMEGA generates.
The MPDATA improves the efficiency of OMEGA's real-time atmospheric flow simulations, which require a lot of computer processing time for calculations related to planetary boundary layer physics, atmospheric radiation heat transfer, and cloud microphysics.
In their STFC Award-winning paper, SAIC authors Nash'at Ahmad, David Bacon, Mary Hall, and Ananthakrishna Sarma implemented MPDATA on an unstructured grid and demonstrated its accuracy and efficiency using analytic and idealized test cases.
The paper, "Application of multidimensional positive definite advection transport algorithm (MPDATA) to environmental modeling on adaptive unstructured grids," was published in the on-line December 2005 edition of the International Journal for Numerical Methods in Fluids.





