What is the HRRR file naming convention?

[blaylockbk]

Answered this in an email...

HRRR files are named as follows

hrrr.t<hour>z.wrf<field>f<forecast>.grib2

• <hour> is the model initialization hour, [00-23]
• <field> is the fields type, sfc, prs, or nat for variables on surface level, pressure levels, and native levels files
• <forecast> is the forecast hour, [00-18] or [00-48] depending on the hour

Example 1

For hrrr.t00z.sfcf06.grib2, t00 is the model run at 0000 UTC, and F06 would be the six-hour forecast, and is valid at 0600 UTC.
At F06, there are two precipitation fields,

• accumulation from F00-F06 (6 hours)
• accumulation from F05-F06 (1 hour)

Example 2

For hrrr.t03z.sfcf12.grib2, t03 is the model run at 0300 UTC, and F12 would be the twelve-hour forecast, and is valid at 1500 UTC.
At F12, there are two precipitation fields,

• accumulation from F00-F12 (12 hours)
• accumulation from F11-F12 (1 hour)

From the GRIB fields, you can eke out both the model initialization time and the valid time (the time the data represents).

Another thing to keep straight is that the F00 zero-hour forecast is called the model analysis or initial state after all the observation data have been assimilated. There is no useful accumulated precipitation data in the model analysis (precipitation accumulated from 00-00 is NaN).