WRF-DFI project meeting – 20060808

Hans, Min, Ju-Won

Lilly room, 2:00-4:00pm, Tuesday 8 Aug 2006

1.Work review

1.1. Single Point Variables

Ju-Won coded the output of time series of surface pressure, 2 m temperature, Qvapor, and vertical velocity on a single model point every hour. However, we need to have the output every time step.

1.2. Verification with observations

Ju-Won got the verification results about DFL test. DFL results (RMSE, Bias, the used observation numbers) are very similar to CYC (No DFL+U3VR) results for most parameters, with a few exceptions that are slightly better or worse. Precipitation scores are better in the all of them, bias, threat score, accuracy, probability of detection, false alarm ration, and equitable threat score.

1.3. DFI

Min showed the backward integration results. They all look reasonable. Ju-Won also confirmed that 10 km backward integration runs without problem.

Min (also Hans and Ju-Won) had several tests for DFI in cycling mode, but the model run always crashes in the second cycle. Min started to look into the compatibility issue between VAR and WRF.

Hans suggested that min have to show MU value every time step on a single point. In that case, the value at initial time should be same with analysis one, and the difference between the initial time value and the last time one should be same with N/2.

Hans suggested having an update boundary after DFI.

2.Work plan for two weeks

Hans: check MU on every time step with Min; help Min with update boundary; will be on vacation during next week.

Min: check MU on every time step MU and remove the shift time step problem; boundary update; continue to work on the cycle related problem.

Ju-Won: write down a single point value for every time step, especially for the center of the cyclone; run a DFI cycle experiment.

3.AOB

Hans will take a vacation starting August 14. We should get all theoretical related issues solved before his vacation.

We will have the next project meeting on: Tuesday 22 Aug, 2 pm, Lilly room.