Development of a 3DVAR Data Assimilation System for Lake Erie Operational Forecast System

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

Yi Chao, Remote Sensing Solutions, Inc., [email protected]
John Farrara, Remote Sensing Solutions, Inc., [email protected]
Carrie Zhang, UCLA, [email protected]

Abstract

A generalized 3-dimensional variational (3DVAR) data assimilation scheme has been implemented into a Lake Erie FVCOM model with a goal to improve the performance of the NOAA Operational Forecasting Systems for the Great Lakes.  The warm bias as identified by the NOAA scientists is significantly reduced by assimilation of the sea surface temperature (SST) data measured by the AVHRR infrared satellites and NDBC buoys.  Water level measurements from both the altimetric satellites and tide gauge stations are also assimilated by 3DVAR into the FVCOM model and shown to significantly reduce the error in subsurface temperature simulations.  The 2005 reanalysis is obtained by assimilating both SST and water level data as well as 50% of the mooring vertical profiles of temperature.  The other 50% mooring data are reserved as the independent data to evaluate the impact of the 3DVAR data assimilation on the FVCOM simulation.  It is demonstrated that each of the three data sets (i.e., SST, water level, and mooring vertical profiles of temperature) are adding incremental values in correcting the warm bias near the surface and placing the mixed layer at the right depth.

1. Keyword
modeling

2. Keyword
hydrodynamic model

3. Keyword
Lake Erie