Fusion of Hurricane Models and Observations: Developing the Technology to Improve the Forecasts

Recognizing an urgent need for more accurate hurricane forecasts, the National Oceanic and Atmospheric Administration (NOAA) recently established the multi-agency 10-year Hurricane Forecast Improvement Project (HFIP). The two critical pathways to hurricane forecast improvement are: validation and improvement of hurricane models through the use of satellite data; development and implementation of advanced techniques for assimilation of satellite observations inside the hurricane precipitating core. Despite the significant amount of satellite observations today, they are still underutilized in hurricane research and operations. This talk will describe our efforts in developing new technology to bring models and observations into a common information system. We will begin by briefly describing two previous very successful NASA-funded projects, the JPL Tropical Cyclone Information System -TCIS - (http://tropicalcyclone.jpl.nasa.gov and http://grip.jpl.nasa.gov) and the NASA Earth Observing System Simulator Suite (NEOS3). These two efforts resulted in building the critical components for our current work, aimed at providing fusion of hurricane models and observations with the goal to improve hurricane forecast. The talk will outline the three areas of on-going research: - the coupling of the instrument simulator with operational hurricane forecast models and incorporation of simulated satellite observables into the existing database of satellite and air-borne observations (TCIS). As part of this integration we will develop tools for model-observations fusion (e.g. data mining to determine when and what satellite observations are available inside the model domain; model sub-sampling in accordance with the time and space coverage of the satellite/airborne overpasses) - the development of a set of analysis tools that will enable users to calculate joint statistics, produce composites, compare modeled and observed quantities, and apply advanced strategies to assimilate remote sensing observations into meso-scale models. - the development of data immersion techniques to enable real-time interaction with the models and visualization of highly complex systems. We will build upon the approach we have developed to visualize a comprehensive set of satellite observations (see http://grip.jpl.nasa.gov). Under this effort we will develop new approaches to include the visualization of the time-series 3D model data. Finally, we will provide examples of model validation and improvement studies that will be facilitated by the new tools that we are now developing. These new tools will provide the missing components that are needed to fully realize the potential of NASA's satellite and airborne observations to validate and improve hurricane forecasts, demonstrating to the public the high value of NASA's satellite data in monitoring and accurately predicting extreme weather events with high societal impact.