Toward More Reliable Weather Forecasts

Weather predictions may become more reliable with the models and software under development at Portland State University and Virginia Polytechnic Institute and State University. Team leaders Dacian Daescu and Adrian Sandu expect the tools to improve forecasting estimates because they take into account the most recent atmospheric data.

Improving the predictive capabilities of atmospheric models is an increasingly important research area. Worldwide models are essential, providing science-based input to the developers of environmental policies.

The combined information extracted from mathematical models and observational data provided by NASA's Earth Observing System present opportunities to better understand issues related to atmospheric modeling, weather and air-quality forecasts and climate change.

Modeling of atmospheric dynamics and integration of observational data into large-scale atmospheric models are among the most challenging and computationally intensive problems in applied science. This research focuses on developing a theoretical framework and the computational tools needed to optimize information from models and observations.

More importantly, these tools aim to identify the model and data components most affected by uncertainties. Optimizing these components could reduce forecast uncertainties.