Availability incorrect for certain 3 channel Texan case(s)

[hrotman-pic]

Experiment with cases of this: 12-003, network Y3.2011. Observed during PH5 to TileDB experiment analysis review.

Three channel Texan experiments have data loaded as usual, and channel number 1 originally assigned in the DAS tables. If the Texan was connected to a vertical sensor, this is unchanged in the final PH5. If the Texan was connected to a horizontal sensor component, this is then changed to 2 or 3. The Array_t entries for channel number are 1,2,3 as appropriate. The utility fix_3chan_texan (assumed to be what was used for this experiment) makes modifications to the DAS tables. This experiment has data that started recording before the deploy time. The vertical channel appears to calculate availability correctly based on the deploy time in Array_t, but the horizontal channels have availability that starts later. This later start is simultaneous with the start of a new recording window in the DAS table and channel number in the DAS table changing from 1 to 2 or 3. The same effect is seen for the

PH5WS availability for an affected station:
#Network Station Location Channel Quality SampleRate Earliest Latest
Y3 3100 -- EP1 D 100.0 2011-09-01T19:00:00.000000Z 2011-09-09T22:15:18.710000Z
Y3 3100 -- EP2 D 100.0 2011-09-01T19:00:00.000000Z 2011-09-04T20:00:00.000000Z
Y3 3100 -- EPZ D 100.0 2011-09-01T18:24:00.000000Z 2011-09-09T22:15:14.710000Z

Deploy and pickup for the station (extraneous fields removed for readability)
Y3|3100||EP1|100.0|2011-09-01T18:24:00.0000|2011-09-09T23:59:00.0000
Y3|3100||EP2|100.0|2011-09-01T18:24:00.0000|2011-09-09T23:59:00.0000
Y3|3100||EPZ|100.0|2011-09-01T18:24:00.0000|2011-09-09T23:59:00.0000

Server-side mseed extraction matches PH5WS availability.

Other experiments affected: unknown. Texan recording windows vary by experiment so potential effective data loss also varies; 12-003 used 1-hour recording windows, so up to almost an hour of data could be missing even though it is intended to be described.

Solution: To be determined. The utility fix_3chan_texan is not deprecated in that it works as of last use, but Texans are no longer in the EPIC pool and the utility has not seen development. As far as I am aware it has been used at EPIC fewer than 5 times in the last 5 years. Obtaining correct availability and complete data extraction for this case may require an answer outside of PH5.

[hrotman-pic]

Here is the start of the DAS table for one horizontal channel in the case, showing the change from the original channel number 1 to channel number 2.
tex14085_partial_stn3100_EP1.csv