P1.12 Toward producing a state-of-the-art reanalysis over the Chukchi/Beaufort Seas via the WRF-Var data assimilation system: assimilation of in situ and satellite-retrieved surface and upper air observations

Monday, 2 May 2011
Rooftop Ballroom (15th Floor) (Omni Parker House )
Fuhong Liu, North Carolina A&T State University, Greensboro, NC; and J. R. Krieger and J. Zhang

Offshore oil development in the Chukchi/Beaufort Seas requires an improved understanding of the surface wind field, an important variable for driving ocean currents and dispersion of potential oil spills. Thus, a study sponsored by BOEMRE has been established to investigate the mesoscale and climatological features of the surface wind field throughout the study area. To this end, producing a very high resolution, long-term regional reanalysis is the top priority for this study.

The Weather Research and Forecasting (WRF) model and its variational data assimilation system WRF-Var were used to produce this reanalysis simulation. We first investigated how varying the model background errors (BE) affect the assimilation performance. A set of observational datasets, including surface observations and radiosondes, as well as satellite retrievals, were then assimilated into the model to investigate their impacts on Chukchi/Beaufort Seas mesoscale modeling. All the assimilation tests were conducted in a continuous simulation from Jul 1 – Aug 30, 2009, forced by the ERA-Interim reanalysis. The observational data were assimilated every 6 hours and the hourly model outputs were verified against all available observations. The results indicate that:

1. The built-in background error (BE) isn't suitable for use in our study area; when using the default BE, the assimilation is noticeably degraded.

2. The customized background error improves the results, with the BE generated from a 1-year simulation performing slightly better than that produced with a 2-month simulation.

3. The largest positive impacts are seen through the assimilation of surface and radiosonde observations.

4. Assimilating QuikSCAT SeaWinds improves the simulation of the surface wind field over the ocean.

5. Assimilating MODIS profiles also produces positive impacts for the upper air temperature and wind fields, while the assimilation of COSMIC profiles only slightly improves the upper air temperature, probably due to limited data availability.

6. Assimilating MODIS moisture profiles has a positive impact; conversely, the assimilation of COSMIC moisture profiles is clearly detrimental.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner