warn on channel at depth if elevation is station elevation

[crotwell]

Channel elevation is supposed to be elevation of the sensor, but it is easy to mess this up and put in the ground elevation in instead. For the majority of cases, channel elevation + channel depth == station elevation but this probably shouldn't be required, especially for sensors installed laterally away from a stations location in hilly terrain.

Two ideas to catch common errors:

1) Warn if channel elevation + channel depth != station elevation

2) Warn if channel elevation == station elevation AND channel depth != 0

First is a more comprehensive check, but might generate to many false positives. The second would miss some errors, but would catch the most common error of misunderstanding the meaning of channel elevation. I, at least, have been guilty of this.

[crotwell]

See https://github.com/iris-edu/stationxml-validator/pull/122 for example where I did this incorrectly.

[timronan]

This rule seems like it would throw a number of false positives and should likely not be included in the validator, even as a warning. We need to be careful about discrediting the validator warnings with false positives. It seems that potentially missing an instance of this error is better than having false positives thrown and having the community discredit the validator.

Example false positives: An infrasound microphone at the Earth's surface that is not colocated with it's datalogger does not have to be at the same elevation as the datalogger.

Deviated boreholes where the channel's surface location is different than the station's surface location.

[crotwell]

You could take care of your false positives by only doing this check if the channel lat/lon was the same as the station lat/lon.

[timronan]

Making a latitude and Longitude check does seems like it would eliminate the false positive issues. So the rule would be:

IF Station:Latitude = Channel:Latitude and Station:Longitude = Channel:Longitude THEN Station:Elevation must equal Channel:depth + Channel:Elevation

If a sensor was above the Earth's surface, in a building for example, would the user have to use a negative depth value for this check to hold true?

[crotwell]

Yep, that looks right to me.

And above ground sensors would have negative depth.

[chad-earthscope]

The proposed rule is based on an interpretation that the Channel:Elevation is the "true" elevation of the Channel/sensor and not equal to Station:Elevation when Channel:Depth != 0. When I read the tea leaves of the SEED manual that does appear to be what is meant, and it makes sense from a data usability perspective.

Unfortunately, spot checking a number of borehole stations from many operators, it appears that many have assumed that Station:Elevation and Channel:Elevation are equal and the true elevation of the sensor must be calculated by adding the Channel:Depth value. Not one odd-ball network, but major permanent networks.

Regardless what the "right" answer is it would be difficult to have such a rule until a consensus is determined. This is a good bookmark for the issue.

[crotwell]

Should this be addressed in a stationxml revision? At least clarify in the docs?

[rcasey-earthscope]

I have attached an explanation to the 'Discussion' label attached to this issue. We currently do not have a clarification or solution worked out so will not yet modify anything in the docs or logic until this is formulated.

[crotwell]

Just a note, if this interpretation of channel elevation is not followed, then rule 223 no longer checks to see if the channel is within 1 km cube of station and so should also be modified.

[rcasey-earthscope]

Hi Philip-

Thanks for making that connection.  We'll have to consider that as well.  There is plenty of good corrective feedback you've provided us and not everything is clear cut when we look at it from different sides.  We'll consider your recommendations (as well as Tim's) and come up with a decision.  We are also making a point of being deliberate about revisions to the documentation and rules so that clients have the ability to adapt to changes at well defined points in time.

-Rob

On Oct 2, 2020, at 6:56 AM, Philip Crotwell notifications@github.com wrote:

Just a note, if this interpretation of channel elevation is not followed, then rule 223 no longer checks to see if the channel is within 1 km cube of station and so should also be modified.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/iris-edu/stationxml-validator/issues/121#issuecomment-702748488, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFL4VN3KOEFOG3ZTOMOUFQ3SIXLYDANCNFSM4QYZIEZA.