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

20020910: File counts (cont.)

>From: Kelvin Droegemeier <address@hidden>
>Organization: OU/CAPS
>Keywords: 200209101349.g8ADnQ118052 NEXRAD Level III volume scans

Hi Kelvin,

>This is **exceedingly** helpful to us because of the "loss-less" nature
>of the transfer and the fact that the file counts are *exactly* what we
>need.

Excellent.  I am glad to hear this.

>A few things, however.  First, we've not yet processed the
>file counts for the days you provide but will try to do so quickly.

OK.

>Second,
>is there any chance of getting this information further back in time?
>I know you archive only the latest 15 days of data, but do you somehow
>save the counts while not saving the actual files?

Unfortunately, no.  We have not been keeping this kind of statistic on
NEXRAD Level III product receipt.

>Third, do your
>counts run from 00 UTC to 2359 UTC, i.e., a GMT-day rather than
>a standard 24-hour day?

All counts were for a UTC day: 0 - 23:59:59 UTC.

>This is an important issue to make sure we're comparing apples with apples.

Yes, exactly.

>Finally, if in the future we could
>use your counts as a benchmark for what the radar actually collects,
>would it be a big pain to strip off this information, say for each 15-day
>block of time?

I just wrote a little shell script that will extract the count information
for each NEXRAD for each day's worth of data we have on disk.  I'll play 
around with this script to extract the counts for "yesterday's" data.
I will then setup cron to run the script at, say 12 Z.  Finally, I will
see if I can make the results available on a web page.

>I'm beside myself trying to get accurate volume scan
>counts and it appears as though you have a true and stable answer...

Well, what I can say is that the file counts we have are a good
representation of the N0R products we receive over NOAAPORT feed.  How
representative this is of the number of volume scans a NEXRAD makes is
open to debate.  One of the reasons I say this is that _if_ the maximum
rate for volume scans for a NEXRAD is 5/min, then there should be no
more than 288 of any type of Level III product produced at PUP attached
to that NEXRAD.  If you look down through the list I sent you, you will
see that there are a couple/three NEXRADs that report _more_ than 288
N0R products in a day.

Also, a couple of us have never seen a Level III product from a radar
that is in Maintenance Mode.  It is possible that those NEXRADs still
send out Level II volume scans, so if they do, the number of scans you
see will exceed -- possibly significantly -- the number of Level III
products produced for that NEXRAD.

Also, there are times when the Level III products do not get sent out
in NOAAPORT even though they are available.  This happens when the
queue of products for the NWSTG channel backs up.  It seems like the
older Level III products get dumped so that the more current ones can
get through, but this is simply speculation on my part.

Finally, the Level III products are pretty much designed to come out
in regular intervals: every 5 minutes, every 6 minutes, or every
10 minutes.  If a NEXRAD can complete a volume scan in a fraction of
the interval for Level III product production

So, if you think about things, there are a number of perfectly logical
explanations for how there could be more volume scans from a radar than
numbers of Level III products.  What would not make sense is having
more Level III products than volume scans.  How does a PUP make a Level
III product when there is no data to make it from!?

>Unidata to the rescue again -- big surprise!!

Yea, we're good ;-)

>You folks always seem to have your act together

We really do try...

>-- Linda must really get angry otherwise, right?  ;-)

Since Linda is from Minnesota, she can't get too angry.  It is just not
in her nature :-)

>Thanks Tom!

No problem.  If/when I get the counts setup on a web page, I will let
you know.

So, who is "Sooner Weather Girl"?

Tom