[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20001130: MDXX problems in McIDAS-X 7.6x PROPLT (cont.)



>From: "Patrick S. Market" <address@hidden>
>Organization: University of Missouri-Columbia
>Keywords: 200011141520.eAEFKZO20569 McIDAS-X 7.6x PROPLT

Pat,

>Hi.  Just a follow up.  I decided to start from scratch.  I wrote Rosie at
>SSEC and got the profiler files again, and ran 'cdftomd' on them to make
>a new MDXX file.  I now seem to get the same results you were getting before,
>and no data at all for LTHM7 or WNCI2 (of course the two I need most *sigh*).

OK.  I know that this is not the end of the story given your second note
which I have appended below.

>Running MDL 7780 ALL KEYS=DAY TIME IDA IDB DIR SPD Z=0 MODE=LOW TIME=12    
>
>gave results similar to what you got.

That is good to know.

>No more miracles...

Except for your next note :-o

>From address@hidden Thu Nov 30 16:08:42 2000
>
>I spoke too soon.  This is too odd.
>
>The same problems have cropped up again.  MDL  appears to show no data in the
>file for, say, WNCI2, but PROPLT is more than willing to plot for it.

I _really_ do not understand this at all.

>At first, it wouldn't, and PROFSECT would polt data for PLTC2.

This is what I expected.

>After
>about 10 minutes of experimentation, PROPLT now polts data for stations
>that I don't think there's data for, and PROFSECT now won't plot data
>for stations that there should be data for.

Hmm... exactly what was the experimentation that you did?  I don't know
of anything that should force a program to start plotting data which
doesn't exist other than rewriting the data in the file (this, PROPLT
definitely does not do).

>I don't know if my McIDAS is buggy, if I have an old decoder, bad data
>files or what.

Me, neither!  Right now, I would have to say that the "or what" option
is the most likely.

>I am about to give up on this, though.  Of that I'm certain.

Too bad.  This has all of the markings of a good mystery story.

>This is turning out to be way too much work for too small of a payoff on
>a storm for which an analysis of profiler data just isn't crucial.

I understand and appreciate what must be a high frustration level
(really, I do!).

>Any suggestions?

The thing is that it is very easy to look at the original netCDF
files and determine if there is, in fact, any data for the stations you
are looking to plot (use ncdump).  If there isn't, then the initial
runs of PROPLT and PROFSECT were telling you exactly what you didn't
want to hear; that there was no data in the files.  If there is, then
there would have to be a bug in:

MDL
PROPLT
PROFSECT

I shudder to think that this is even a possibility!

What happens when you do a planer plot of the profiler data?  To do
this you would use the PROFP command (non-ADDE) or the PROFDISP command
(ADDE).  I just used the PROFSECT command here on the original data
file that you made available to me, MDXX4450, and I see no wind plots
for any stations in IL:

PROFDISP WINDB 4000 USA  18 1997100 DAT=MYDATA/PTSRCS.4450
PROFDISP WINDB 8000 OLAY 18 1997100 DAT=MYDATA/PTSRCS.4450

Give me a login as the user you are running McIDAS as to the exact
system that you are using and let me play around for an hour or so.  If
there is anything funny going on, I should be able to spot it easier
than trying to imagine multiple sequences of events that could lead to
the behaviour that you are seeing.  I could also take a look at the
netCDF files you got from SSEC to see if there really is any data for
WNCI2.

Tom