Ensemble Kalman Filter Data Assimilation Development for Lake Erie

Session: 36. - Improving Model Predictions through Coupled System and Data Assimilation

Juliette Daily, Rochester Institute of Technology , [email protected]
Matthew Hoffman, Rochester Inst. of Technology, School of Mathematical Sci., [email protected]

Abstract

Data assimilation is the process of combining observational data with computational models to improve state estimation and forecasting skill in geophysical systems. NOAA has recently upgraded their Lake Erie Operational Forecast System (LEOFS) to a high resolution, state-of-the-art model—the Finite Volume Community Ocean Model (FVCOM)—but the system currently contains no data assimilation. We explore the potential for improvement of nowcasts and forecasts of the LEOFS through data assimilation using the Local Ensemble Transform Kalman Filter (LETFK). The LETKF has been applied to many oceanic and atmospheric systems, has been tested at operational forecasting centers, and is capable of three and four dimensional data assimilation. Here we present preliminary results on coupling the LETKF with the FVCOM model from the Lake Erie Operational Forecast System and performing Observational System Simulation Experiments (OSSEs) using synthetic temperature data. In an OSSE, observations are simulated by taking results from a model run and adding error. Because there is a known truth, the data assimilation system can be tested and the potential for improvement to the estimation of temperatures and currents can be benchmarked.

1. Keyword
computer models

2. Keyword
Lake Erie

3. Keyword
modeling

4. Additional Keyword
data assimilation

5. Additional Keyword
ensemble Kalman filter