SBIR-STTR Award

We propose to develop and deliver a generalized 3DVAR data assimilation module, in an opensource and non-proprietary programming language, which is compatible with both ROMS and FVCOM and easily incorporated into the existing NOAA operational forecast system (OFS). Our innovation is to develop a data assimilation roadmap contrasting 3DVAR with other advanced techniques such as multi-scale 3DVAR, 4DVAR or Local Ensemble Transform Kalman Filter (LETKF) in terms of accuracy and computational efficiency. We will leverage our experience and expertise in developing 3DVAR for a ROMS-based real-time forecast system for the California coastal ocean. 3DVAR has an ability to propagate observational information in both the horizontal and vertical directions while still keeping the computational overhead at a manageable level (e.g., 2X the forward model run time as compared to 20X or more for (4DVAR). Working closely with NOAA scientists, we will identify requirements for data assimilation, implement 3DVAR into the model selected by NOAA, and demonstrate the ability of 3DVAR to 1) incorporate various observational data sets into the existing NOAA OFS, 2) run efficiently from the computationally perspective with a user friendly interface, and 3) improve over unassimilated simulations.

We propose to continue the development of a generalized 3DVAR data assimilation module, in an open-source and non-proprietary programming language, which is compatible with FVCOM and incorporated into the Lake Erie Operational Forecast System (LEOFS). We will leverage our experience and expertise in using 3DVAR for a ROMS-based real-time forecast system for the California coastal ocean. 3DVAR has an ability to propagate observational information in both the horizontal and vertical directions while still keeping the computational overhead at a manageable level (e.g., 2X the forward model run time as compared to 20X or more for 4DVAR). Working closely with NOAA scientists, we will perform a hindcast experiment and demonstrate the ability of 3DVAR to 1) incorporate both in situ and satellite observational data sets into the existing LEOFS, 2) run efficiently from a computational perspective, and 3) improve over non-assimilating simulations. Our innovation is to include an Ensemble Kalman Filter (EnKF) module and develop a data assimilation roadmap towards a hybrid ensemble-variational (EnVar) data assimilation system SUMMARY OF

Anticipated Results:
Successful completion of the proposed work will demonstrate the positive impact of 3DVAR assimilation of both in situ and satellite observations on the performance of the Lake Erie Operational Forecast System (LEOFS). Through 3DVAR data assimilation, the impact of observational data on the operational forecast system will be quantified. Specific delivery includes an advanced 3DVAR data assimilation module, in an open-source and non-proprietary programming language, which is compatible with FVCOM and incorporated into the LEOFS. The EnKF data assimilation system will also be developed for the LEOFS with a goal to formulate a data assimilation roadmap that will guide future improvement of the 3DVAR data assimilation. The outlook for the commercial need and market penetration for the ocean data assimilation technology is very promising. Since this technology fits into our company's mission statement, we will develop a more detailed business plan to expand the sales and marketing strategy for commercialization.