I am not clear on the purpose of the "Appended" met data files on the ARL server. If I understand things correctly, I would not need to use anything other than the regular model output files in order to ensure that I am taking advantage of the best data. For example - say it is 10AM EDT and I want to do a 24-hour dispersion forecast for a release that began at 8PM EDT the previous evening. My understanding is that if I download the met file from the NAM run initialized at 00Z and use it to do the model run, I will still be getting the benefit of observed data after 00Z, as well as the benefit of the 06Z model run for the time period after 06Z. In other words, the version of the 00Z model run file that I download at 10AM EDT would have been updated to include data that became available later than 00Z. Is this correct? Or would I want to be using appended data?
John Kent
Met data file types
-
- Posts: 451
- Joined: November 9th, 2012, 4:23 pm
- Registered HYSPLIT User: Yes
Re: Met data file types
The "Appended" met data files are concatenated (appended) +0 +3-hour data from forecast files from successive forecasts. Since the forecasts are run every 6 hours, this creates a pseudo-archive. The last time period in the appended file is the time period immediately before the corresponding forecast file starts. For example for today's 00z GFS cycle, the appended file begins at 00z June 28 (two days ago) and end at 21z June 29; the forecast file begins at 00z June 30.
The original purpose, when coupled with the corresponding forecast file, was to start a backward trajectory within the period of the forecast file, and have the trajectory continue into the time period of the appended file.
For your example - There is only one "version" of the 00z forecast file. There is no update. At 10 am EDT, the 06z would be the current forecast. There are two ways you could do this run using the forecast and appended data:
(1) 06z appended file, 06z forecast file - This will use 00z, 03z from the appended file, then the forecast file
(2) 00z forecast, 06z forecast - This will use 00z, 03z from the 00z forecast (since it is not in the 06z forecast), then use the 06z forecast, (since it has the shorter forecast hour than the 00z forecast (at the 12z time period, it is a 6-h forecast in the 06z file, but a 12-h forecast in the 00z file).
The original purpose, when coupled with the corresponding forecast file, was to start a backward trajectory within the period of the forecast file, and have the trajectory continue into the time period of the appended file.
For your example - There is only one "version" of the 00z forecast file. There is no update. At 10 am EDT, the 06z would be the current forecast. There are two ways you could do this run using the forecast and appended data:
(1) 06z appended file, 06z forecast file - This will use 00z, 03z from the appended file, then the forecast file
(2) 00z forecast, 06z forecast - This will use 00z, 03z from the 00z forecast (since it is not in the 06z forecast), then use the 06z forecast, (since it has the shorter forecast hour than the 00z forecast (at the 12z time period, it is a 6-h forecast in the 06z file, but a 12-h forecast in the 00z file).
Re: Met data file types
Thanks for the reply. OK, so I need to use two met files to do a dispersion run that begins in the past and continues into the future. Is there a way to do that using the web version of HYSPLIT? I don't know of a way to specify more than one met file with the web version.
-
- Posts: 451
- Joined: November 9th, 2012, 4:23 pm
- Registered HYSPLIT User: Yes
Re: Met data file types
The web automatically has one forecast file and then several of the appended files going back in time. Try your run, then look at the file CONTROL. That lists the meteorology files.