Downloading waveform data for some stations provides a chopped-up stream

[CedricDM]

Hi,

Recently, I noticed that the returned Stream object for some stations contains inconsistently chopped-up data.

For example, when doing:

In [5]: st = client.get_waveforms(network='NZ', station='SYZ', starttime = UTCDateTime('2024-02-27'), endtime = UTCDateTime('2024-02-28'), location='*', channel='HH?')

I get the following result:

In [7]: for tr in st:
...: print(tr)
...:
NZ.SYZ.10.HHE | 2024-02-27T00:00:00.000000Z - 2024-02-27T00:02:15.990000Z | 100.0 Hz, 13600 samples
NZ.SYZ.10.HHE | 2024-02-27T08:02:16.000000Z - 2024-02-27T09:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHE | 2024-02-27T10:02:16.000000Z - 2024-02-27T11:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHE | 2024-02-27T12:02:16.000000Z - 2024-02-27T13:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHE | 2024-02-27T14:02:16.000000Z - 2024-02-27T16:02:15.990000Z | 100.0 Hz, 720000 samples
NZ.SYZ.10.HHE | 2024-02-27T17:02:16.000000Z - 2024-02-27T19:02:15.990000Z | 100.0 Hz, 720000 samples
NZ.SYZ.10.HHE | 2024-02-27T22:02:16.000000Z - 2024-02-28T00:00:00.000000Z | 100.0 Hz, 706401 samples
NZ.SYZ.10.HHN | 2024-02-27T00:00:00.000000Z - 2024-02-27T00:00:03.720000Z | 100.0 Hz, 373 samples
NZ.SYZ.10.HHN | 2024-02-27T00:02:16.000000Z - 2024-02-27T02:02:15.990000Z | 100.0 Hz, 720000 samples
NZ.SYZ.10.HHN | 2024-02-27T04:02:16.000000Z - 2024-02-27T05:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHN | 2024-02-27T07:02:16.000000Z - 2024-02-27T09:02:15.990000Z | 100.0 Hz, 720000 samples
NZ.SYZ.10.HHN | 2024-02-27T14:02:16.000000Z - 2024-02-27T15:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHN | 2024-02-27T17:02:16.000000Z - 2024-02-27T20:02:15.990000Z | 100.0 Hz, 1080000 samples
NZ.SYZ.10.HHZ | 2024-02-27T00:00:00.000000Z - 2024-02-27T00:00:01.560000Z | 100.0 Hz, 157 samples
NZ.SYZ.10.HHZ | 2024-02-27T01:02:16.000000Z - 2024-02-27T04:02:15.990000Z | 100.0 Hz, 1080000 samples
NZ.SYZ.10.HHZ | 2024-02-27T06:02:16.000000Z - 2024-02-27T07:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHZ | 2024-02-27T11:02:16.000000Z - 2024-02-27T12:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHZ | 2024-02-27T13:02:16.000000Z - 2024-02-27T14:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHZ | 2024-02-27T17:02:16.000000Z - 2024-02-27T18:02:15.990000Z | 100.0 Hz, 360000 samples
NZ.SYZ.10.HHZ | 2024-02-27T19:02:16.000000Z - 2024-02-27T22:02:15.990000Z | 100.0 Hz, 1080000 samples
NZ.SYZ.10.HHZ | 2024-02-27T23:02:16.000000Z - 2024-02-28T00:00:00.000000Z | 100.0 Hz, 346401 samples

As you can see, all three channels have been chopped up into traces containing data for one or more hours exactly, in between which exactly one or more hours of data is missing. However, the most striking thing is that this is not consistent over all three channels, i.e. channels are not chopped up at the same start- or endtime. Therefore, I do not think this is an issue with the station itself.

There is a similar problem for the 15th of March 2024 for station DCZ.

Could this be an issue with the data upload to FDSN?

@calum-chamberlain

[calum-chamberlain]

It looks like the gaps also exist in the AWS bucket version of these data.

[elidana]

kia ora @CedricDM , thanks for reaching out on this.

At a first pass, it seems that both stations were having problems and had some of their equipment replaced on those days.
At SYZ there was a faulty digitizer and an issue with the data transport, and a similar problem was happening at DCZ.

The team is investigating in more detail (indeed is odd how waveforms are "chopped") and will report here if anything else comes up.

Thanks!