US058GMET-GR1mdl.0058_0240_00000F0RL2008082700_0100_010132-000000wnd_vcmp
What does 'F0RL' represent ?

F0 =    forecast ops    (fcst_ops)
RL =    Real time

Incidentally, if your searching for these strings, the F0 is "F" "zero", not
"F" "capital o".  Don't ask me how this got started, but my guess is that it
was an initial misunderstanding in the beginning.  Maybe someone wrote FO,
and whoever implemented it in the original system probably interpreted it as
F0 and it has lived forever.  In some fonts, visually the O and 0 look
exactly the same.

The first two characters refer to the "data set" name for the field in ISIS.
There are "forecast ops", "analysis ops", "forecast validation", "forecast
beta", etc.  It is an agreed-upon name to identify a particular data set
that is coming from the models. This helps keep the sources of the fields
uniquely identified so that you don't confuse operational fields with other
type of fields that may have been generated for a specific purpose.  In
fact, sometimes fields with different data sets might be written into the
same ISIS directory.  If the distinction wasn't made in the fields (i.e., in
the filename), then one dataset type would clobber another dataset type if
the field was available as both types.

The second two characters refer to the model runtime.  There are several
runtimes, for example

        PL      (preliminary)   (output products are based on an early run
of the model (and partial obs data set)
        P2      (pre-two)               (this is another early run
designator. (partial obs data set)
        RL      (realtime)              (output is based on all available
obs data for the watch cycle)
                                         for NOGAPS, the model start about 3
hours into the watch.
                                        These are always designated for the
00Z and 12Z watch cycles
        OF      (offtime)               (output is from a model run made
during the 06Z and 18Z watch cycles)
        NL      (null)          (catch-all, or unspecified model runtime)

These are also used to distinguish the fields according to which model
runtime provided the data.  For instance, in ISIS, the NOGAPS output fields
for the entire watch cycle used to be put into a single directory, and would
include 00z and 06z realtime and offtime fields.  (I think they have changed
the scheme to put the fields in separate runtime directories these days).  

Now, for models other than NOGAPS, such as COAMPS in the pre-CAGIPS era, the
runtime specified the NOGAPS model runtime that provided the boundary
conditions to COAMPS.  In order to distribute the COAMPS (and other regional
models) processing over the watch cycle for the various areas, a couple of
COAMPS areas might be based on the NOGAPS output for P2, others based on
NOGAPS output for PL, etc.  However, in CAGIPS grib products, the runtime
distinction for COAMPS, and possibly some other regional models, has
disappeared and they are all labeled as "NL".  This is not true for CAGIPS
NOGAPS products; the distinction is still retained in the CAGIPS NOGAPS grib
product names, as your example shows.