DQRID : D050215.5
Start DateStart TimeEnd DateEnd Time Data Quality Metric
11/11/2003000002/05/20040000Suspect
more
Subject:
NSA/MPL/C1 - MPL clock drift
DataStreams:nsamplC1.a1, nsamplpsC1.a0
Description:
The MPL PC clock was noted to be off (slow) by more than an hour over a several day 
period.  It has previously been noted that the time synchronization program "D-4" fails to 
update the PC clock under certain circumstances.  As a result, the program "D-4" was replaced 
by "AboutTime" on 2/5/2004.  

The specifics of this problem are not well understood.  It turns out that the instrument 
has TWO clocks on the instrument computer display - one being displayed by the MPL 
software and the other being the PC clock displayed in the lower right-hand corner of the 
screen. The MPL software clock was the one that is noted on daily rounds and was never off by 
any amount. Apparently the only clock that was off was the computer clock.

What is clear is that between the time the MPL computer was installed on 11-11-2003 and 
the CM log on 1-06-2004 the MPL clock exhibited a discrepancy of over 1 hour and 20 
minutes.  What is not clear is whether this clock discrepancy was also applied to the data time 
stamps.  Furthermore, seven days later the MPL clock exhibited only a 1.77 second drift.  
Not exceptional at all, and not in line with the previous discrepancy of more than 4800 
seconds over a span of 57 days.

Attempts to identify gradual correctible clock drift in the MPL data were not successful.  
As a worst-case scenario, the MPL data may be inappropriately time-stamped by as much a 
1 hour 20 minutes and 41 seconds slow relative to UTC between the dates of 2003-11-11 and 
2004-01-06.  After this time until 2004-02-05 time synchronization was maintained within 
a few seconds a day with local observer assistance.
Suggestions: 
The data user is advised to use the MPL cloud products and backscatter profile with 

caution.  Correlation between low cloud base height detected by the MPL and by the 
ceilometer may be helpful to identify errors in time-stamp.  Also it may be possible to 
associate MPL detected clouds with increased brightness temperatures from the AERI or IRT.
Measurements:nsamplpsC1.a0:
  • time
  • time_offset
  • base_time
more
nsamplC1.a1:
  • base_time
  • time_offset
  • time
more

Close this window