Lake Erie Modeling and Data Assimilation to Improve Operational Forecast

Session: Improving Model Predictions Through Coupled System and Data Assimilation (2)

Yi Chao, Remote Sensing Solutions, Inc., [email protected]
Philip Chu, NOAA/GLERL, [email protected]
Eric Anderson, NOAA/GLERL, [email protected]

Abstract

Operational ocean prediction is still in its infancy. As NOAA is developing a comprehensive ocean observing system, there is an increasing need to develop a companion ocean forecasting system similar as numerical weather prediction.  Data assimilation is a critical piece of this ocean forecasting puzzle.  This talk will describe a 3DVAR data assimilation system implemented for the Lake Erie FVCOM operational forecasting model.  The three-dimensional variational (3DVAR) data assimilation method was selected because of its computational efficiency as compared to other advanced methods such as 4DVAR or Kalman Filter.  The Lake Erie FVCOM model simulations during 2005 and 2017 were performed with 3DVAR data assimilation of various observational data sets (e.g., satellite SST and SSH, tide gauge water level, NDBC buoy surface temperature, as well as mooring temperature profiles).  The performance of model simulations with and without 3DVAR data assimilation will be presented.  Future work to test this 3DVAR data assimilation in real-time Lake Erie operational forecast will be proposed.