joe-pitt
commented
2 months ago I have changed my mind about this. While Peter's additional timestamp adjustment does reproduce the old start of sampling times (calculated by the acrg code from the .C files), at least for MHD and TAC this doesn't appear to be the start of sampling. Here are some things I found digging into this:
TAC Medusa:
- Samples from 100 seconds to 1300 seconds on the stripchart
- Chrommatogram timestamped at about 30 mins on the stripchart
- Peter's .nc files are adjusted backwards from this by 20 mins
- So to get to the start of sampling we need to go back another 9 mins
- For example, the run 240821.1845 started at 18:15:00, started sampling at 18:16:40, finished sampling at 18:36:40, and finished the run at 19:20:00
MHD Medusa:
- Samples from 100 seconds to 1300 seconds on the stripchart
- Chrommatogram timestamped at about 36 mins on the stripchart
- Peter's .nc files are adjusted backwards from this by 20 mins
- So to get to the start of sampling we need to go back another 15 mins
- For example, the run 240821.2007 started at about 19:31, started sampling at about 19:32, finished sampling at about 19:52, and finished the run at 20:36
Peter Salameh appears to have implemented his own adjustment to the start of sampling (I guess in some recent GCWerks update). We should check this more rigorously, but if it is true in all cases then we should remove the part of agage-archive that does this adjustment