[decoders] nominal time

• To: decoders@xxxxxxxxxxxxxxxx
• Subject: [decoders] nominal time
• From: "Matt Alonso" <matt.alonso@xxxxxxxxx>
• Date: Tue, 14 Oct 2008 12:17:29 -0400

Hi All,

I apologize if this is a common question but I could not seem to find any 
answers as of yet.  Please bear with me as I am very new to dealing with this 
data.

I'm using the perl netcdf synoptic decoder v5.0.0 (syn2nc) to convert data 
delivered via LDM v6.7.0 and in my pqact.conf file I am only (supposed) to be 
pulling the 6-hourly files (SM....).  I'm ohoping that someone can explain what 
exactly time_nominal represents?  My initial thought was that it would be the 
UTC timestamp that corresponds to the Synoptic reporting time.   So in the file 
created for the 12Z ob on 20081014 (Surface_Synoptic_20081014_1200.nc) I would 
have expected all the timestamps to be 1223985600 which I get from the 
following command:

date -u --date="20081014 12:00" +%s

However, an ncdump of that file shows that the time_nominal value can vary.  
Looking at the time_nominal for WMO#96315 shows a value of 1223965800 which 
equates to 06:30 so I am obviously missing something.  I also noticed that 
seemingly spurious observations take place at intervals outside of the 6-hours 
I would expect my pqact.conf to be retrieving.  For example the file 
Surface_Synoptic_20081013_1000.nc was created which only contains an 
observation for WMO#91824:
SIPS20 NZKL 130900
AAXX 13104
91824 46/// ......

In the header of the raw synop it seems to indicate that it is valid for 0900 8/13 but 
isn't the 1310 indicating that it is a 10UTC ob?  The time_nominal in the converted file 
returns 09:40 on 8/13.  What is the deciding factor in which file a given synop 
observation is placed in?  Would people be willing to suggest any "rules" they 
use to discard erroneous data in situations like this?

Again I apologize if this is just my lack of understanding when it comes to 
Synop syntax.  Thanks in advance for the help.

Cheers,
Matt