On the issue of product latency Karen asked:
Out of curiousity, is it possible you recieved the files earlier, and
that you re-received them later?
So I wrote a specific action to file RTMA products by valid time, PARM,
and receipt time.. then created a text file of all associated file names
covering a roughly 24 hour period... I then imported that file into
excel for analysis and created pivot tables to see if I could better
understand what was going on.
Here notice the left column is how I actually filed the data.. with the
valid time of the data, then FHR, PARM, and then the pqact
"%Y%m%d%H.grib2" for time that the data was received via my dishes.
From this I created a pivot table to show the count of products
received, over each valid hour.. This was to confirm if Karen's
suspicions were correct. The results look as follows:
I colored the outliers in red so that they can be easily seen.. So
remember this is how many products were received with that VALID time,
or the time the data was valid for. Thus if more than the normal of 20
or so products were received, this would mean multiple reissues, showing
that Karen was indeed correct... The times that the reissues occurred
was:
Where for 15z a total of 83 RTMA products were transmitted over the
dish, 73 for 16z, 56 for 18z and so forth... So we know that much is
true.. Cudos Karen! :)
The question remains though, why I received errors due to product delay?
What I have done for now is to write a double redundancy check, so that
if something is amiss with the data it will autofetch from the ncep
servers.
If anyone would like to view or play with the data I used to analyze
data receipt, the Excel spreadsheet, including the pivot table, is
available here (Excel 2016):
http://modelweather.com/files/grids/rtma/
Also included is a raw text file of the filenames containing the VALID
time and receipt time, in case you do not have excel :)
cheers,
--patrick
---------------------------------------------------
Patrick L. Francis
Vice President of Research & Development
AerisWeather.com
wxprofessor@xxxxxxxxx
http://facebook.com/wxprofessor
--------------------------------------------------
--
