Sensitivity Descrepencies in Broadband Data (HH Channel)

[AaronJR474]

Good afternoon,

As part of our work on the continued expansion of the NZ GMDB (Hutchinson et al., 2024), we have conducted a study on the broadband data before its inclusion over previous quality concerns. In this regard, we have found that there are several instances where the sensitivity values of the HH channel appear to be incorrect, as evidenced by its bias relative to the strong-motion channels (see image). As part of this exercise, we identified about 11 stations (see table) with this issue, where once the period (in time, e.g., 2000 - 2010) which exhibited substantial bias is removed, the overall dataset of the HH channel now aligns. In most cases, it is quite clear that the initial sensitivity value appears to be incorrect.

Considering the points made, is there any way in which GeoNet can validate whether or not the sensitivities during these periods are accurate and/or consistent with the calibration logs? Interestingly, in some instances, there are cases where the BN channel also appears to exhibit instances where a "bad" initial sensitivity value exists.

Thank you in advance for your time and consideration of this issue.

PS: We have also utilised HN-HH and BN-HH ratios, which all exhibit this same trend.

<style> </style>

Station	Incorrect Sensitivity	Sensitivity Adjusted
BKZ	<2013	2013 – Present
COVZ	<2016	2016 – Present
DSZ	<2004	2004 – Present
HIZ	<2014	2014 – Present
KNZ	<2009	2009 – Present
OUZ	<2010	2010 – Present
TLZ	<2019	2019 – Present
TOZ	<2008	2008 – Present
URZ	<2017	2017 – Present
VRZ	<2014	2014 – Present
WHZ	<2006	2006 - Present
KUZ	???

[elidana]

thanks a lot for touching base and your detailed report @AaronJR474 ! this is much appreciated.

We will need to investigate this with the team here at GeoNet, and will get back as soon as we have a bit more info.

thanks a lot for taking the time to report this, and if you notice any other discrepancy in other stations please feel free to update this ticket accordingly!

Ngā mihi nui,

Elisabetta

[salichon]

Kia Ora @AaronJR474
Thanks a lot for this, as a start would you send a set of event IDs for which you got the most obvious discrepancies with the sensitivity ?
cheers

[AaronJR474]

Good morning, and I appreciate your quick reply. Regarding your request, @salichon, I can facilitate this by sending a couple of zip files containing HN, BN, and HH channel rejected records from our criteria, which currently identify some of these trends. However, I would also urge you to actually look at the "bad" sensitivity periods as seen in the former figure (summarised in the table), because this "drop" in seismicity in the HH channel is not an earth physics phenomenon. To facilitate this, in addition to the table sent previously, I am also attaching plots of all HH stations.

PS. The single .rar file attached contains all referenced files. Link to OneDrive since the file is too large:

https://ucliveac-my.sharepoint.com/:u:/g/personal/ara299_uclive_ac_nz/EazKt5TRkIZOoNGS3Jjdo2UBTKuDz3N8gVgDe4YbzFxAJw?e=3Lw5jl

[AaronJR474]

Although the title says HH, I extended my study to the SH and EH channels, and my findings are as follows:

1. EH appears to be fairly consistent, but there are four stations with apparent sensitivity issues (DUWZ, TUWZ, CMWZ and CPWZ) with plots seen below.
2. In addition, due to the nature of the sensor, the max period appeared to be in the order of 1 - 2s, which was derived from reviewing co-located records ratios between HN/EH.
3. SH appeared to generally have a severe bias across the board with obvious near-source and far-source clipping/scaling issues as seen in the figure, with the decreased trend with Rrup. This sensor appears to be one of the older ones, and I assume the reliability of the data derived from the same is questionable or should be used with caution, regardless.

If you have any thoughts and/or confirmation on these points, feedback is welcome.