BMB-UCAR WRF 3D-Var project

Summary for 15 October29 December 2006

Reporting time (for this report): 29 December 2006

 

 

Task 5. Meteorological performance

5.1. BE

 

Shuiyong re-calculated the BE for 27 and 9km domains using the accumulated pre-operational forecast samples in a period of the August 2006 with the setup of WRFSI+wrfvar (BE interpolation) +WRF(v2.1). Both of the single observation and AWS+GPS experiments were performed. But the humidity assimilation needs further work.

 

5.2. WRFDFI

 

This task is finished. Please refer to MinÕs summary.

 

5.3. Comparison

 

The comparison between the pre-operational WRF system and the operational MM5 system for the forecasts for June, July and August, 2006 has been conducted by Jiqin. ItÕs found that MM5 outperforms WRF for surface wind and temperature, WRF outperforms MM5 for surface humidity, and both have similar forecast performances for upperair variables.

This task needs continuous work and regular report will be submitted in the next year.

 

Task 6. Monitoring of pre-operational runs and case studies

 

6.1. Verification

 

This task is finished.

 

6.2. OSE

 

BMB real-time GTS data collecting system is still under construction. The main problem is that weÕre in lack of robust decoder software. This task is still ongoing.

 

6.3. Cycling

 

A pre-operational WRF-based cycling system is ready.  WRFVAR analysis and forecasts are cycled on both 27km and 9km domains with 6-hour interval. The cycles restart at 1200UTC per day.  48-hour forecasts are performed at 00 and 12 UTC, while just 6-hour forecasts are performed at 06 and 18UTC. The lateral boundary conditions are provided by GFS forecasts.

A testing period from 12UTC 25 July to 00UTC 01 August is chosen for performances comparison between the WRF-based cycle and MM5 operational runs. A detailed report about the comparison will be submitted next time.

 

6.4. Tuning

 

WPS v204 + WRF v2048 has been updated to WPS v209 + WRF v2056 with WeiÕs help. Shuiyong conducted a series of experiments to test the sensitivity to different schemes: vertical levels; surface physics; landuse table; shortwave radiation scheme. ItÕs revealed that new version reduces cold bias and improves precipitation forecast.

 

6.5. MM5DFI

 

This task is finished.

 

Next report time: 14 January, 2007!!