linking WRF input files

[chunchih]

Hi,

I am using NAM analysis to force my real-case WRF simulation. From the NCDC model archives (https://www.ncei.noaa.gov/data/north-american-mesoscale-model/access/historical/analysis/200706/20070622/), I downloaded all the .grb files for the day of interest. The NAM analysis files are named as:

'namanl_218_20070622_0000_000.grb'
'namanl_218_20070622_0000_003.grb'
'namanl_218_20070622_0000_006.grb'
'namanl_218_20070622_0600_000.grb'
'namanl_218_20070622_0600_003.grb'
'namanl_218_20070622_0600_006.grb'

I believe '20070622' is the date of the analysis file, '0000' and '0600' indicate the analysis times (00 and 06 UTC) of the date, and the last three numbers in the file names indicate forecast output hour since the analysis time (i.e, '003' is the forecast at 3-hr since the analysis time and '006' is the forecast at 6-hr since the analysis time).

My question is, are the forecast-hour files (the ones with the last three numbers in the file names = '003' or '006') needed to properly drive my WRF simulation? Or it's just the analysis files (the ones with the last three numbers in the file names being '000') are needed to be linked via running ./link_grib.sh?

Thanks!

Chun-Chih

 

[Ming Chen]

HI, Chun-Chih,
Your understanding of '20070622', '0000' and '0600' is correct. To answer your question, I believe that the analysis files at 6-hour intervals are sufficient to drive WRF. Of course you can also use the forecast files, which provides forcing at 3-hour intervals.

 

[chunchih]

Hi Ming,

Thanks for your answer! For now, I set my interval_seconds to 21600 (six hours), so I believe my simulation is driven only by the 6-hourly analysis files.

If I set interval_seconds to 10800 (3 hours) instead, then the NAM forecast files would be used to create the met_em files?

 

[Ming Chen]

Chun-Chih,
You are right.If you want to run WRF driven by 3-hour NAM data, then you need to ungrib NAM forecast files.