Evaluation of WRFDA 4D-VAR through Month-long Run and Case Study

WRFDA (The Weather Research and Forecasting Data Assimilation) 4D-VAR has been developed since 2004 by NCAR/MMM Data Assimilation Group. It is in the public domain and is freely available for community use. It is designed to be a flexible, state-of-the-art atmospheric data assimilation system that is portable and efficient on available parallel computing platforms. 4D-VAR system keeps being updated and optimized year by year.

This study will focus on evaluating the performance of the newest released WRFDA 4D-VAR version in April 2010, which combined the new development of optimization for adjoint model. One-month cool-start (no cycling) experiments was conducted on AFWA T8 domain with 45km resolution, which covers the Gulf of Mexico and north Atlantic area during August and September in 2007. The observation data used in the experiment include SYNOP, SOUND, METAR, QSCAT, BUOY, SHIP, PILOT, AIREP, and SATLLITE WIND. For comparison, the paralleled 3D-VAR experiment was employed monthly too with the same experimental design as 4D-VAR. We examined not only the monthly performance of 4D-VAR, but also the impact on hurricane track and intensity forecast. Preliminary result has been accessed from the month-long run. 4D-VAR can ingest more high time frequency surface data since the more accurate background provided by 4D-VAR, which was forecasted by non-linear model whine the time window. The 4D-VAR presented an impressive improvement both in the initial and 48 hours forecast wind field. For the next step, we will check if 4D-VAR is able to improve the hurricane track and intensity forecast. The result will be presented at the conference.