Data quality of bottom current sensors

HelpDesk
1

I have three questions (continued in reply chain) regarding the long time series of seafloor 3-D velocity records. There are many time periods when the data appear noisy, but could maybe be fixed by addressing a gain-related issue. Specifically, for the VEL3D instruments at Slope Base and Hydrate Ridge:

  • Slope Base (RS01SLBS-MJ01A-12-VEL3DB101)
  • Hydrate Ridge (RS01SUM1-LJ01B-12-VEL3DB104)
  1. Is there a gain issue for the HYS14 instrument (Vel3DB104) between the Aug 11 2017 deployment and the June 26 2018 deployment? Could this be fixed to allow the data to be used?
    Screenshot 2024-03-25 at 2.39.35 PM
    Screenshot 2024-03-25 at 2.39.35 PM1916Ă—1326 133 KB
2

Sorry, multiple images require multiple posts.

  1. Occasionally, there are BIG excursions where bottom currents momentarily spike above realistic values– what could those be? We see them on both instruments, at different times. For example, HYSB1 (Slope Base) in ~ May 2017:
    Screenshot 2024-03-25 at 2.39.50 PM
    Screenshot 2024-03-25 at 2.39.50 PM1906Ă—1366 143 KB
3
  1. For many times when we appear to have “good” data, there still seems to be some clipping mixed in. Somewhat arbitrarily, I convert any values above 0.5 m/s to NaNs. But, this still leaves big spikes that I don’t know whether or not to trust. What causes this “clipping”? Is it a real signal?
    Screenshot 2024-03-25 at 2.40.05 PM
    Screenshot 2024-03-25 at 2.40.05 PM1920Ă—883 109 KB

Just wondering what to make of the noise in general, and/or if any of the bad time periods are salvageable.

Thanks!

1 Like
4

Hello! Apologies for the belated reply - we’ve been looking into the 3D velocity data time periods you pointed out.

There was definitely a known issue during that 2017-18 deployment. The cause remains unclear, but the whole deployment should have been annotated as suspect/failed during that time period while we investigated. It’s not clear that a simple gain adjustment will fix it, but we’re taking another look at the data in case there was a missing calibration factor or some other “simple” explanation.

With respect to the frequent spikes, those are concerning. I would not trust those data, and they can be filtered out by using the gross range QC flags (which should be active now). I know the VEL3D data can be noisy, but it’s not clear if it’s electrical noise from the power supply or an instrument issue. We’ll get back to you when we’ve done some more cross-comparisons.

Thanks for your question, and we’ll follow up soon!

-Mike Vardaro, OOI-RCA data team