### Summary

This metric cross-correlates a 10-minute window surrounding the predicted P arrival from a M>=6.5 event at neighboring stations. The measurement is the signed value of the correlation function having the largest absolute value.

### Uses

Station pairs with correlation peak values consistently less than 0 are either phase-reversed with respect to each other, or the channel dip sign may be incorrect. Because this metric does not indicate which of the two stations is reversed, we advise against using it as a constraint for **rrds** requests.

### Data Analyzed

**Events** – magnitude >= 6.5

**Traces** – two N.S.L.C (Network.Station.Location.Channel) per measurement where traces are similar Channels from neighboring stations (see Algorithm for details)

**Window** – 10 minutes (UTC) centered on the predicted arrival of the P wave from the event

**Data Source** – IRIS miniSEED archive or IRIS PH5 archive, and USGS event service

**SEED Channel Types** –
[BCDFHLM]H? | High Gain

### Algorithm

- For each event having M>=6.5,
- Request the event location,
- Find all stations and channels
- within 15 to 90 degrees of the event,
- whose Channel names describe a broadband sensor.

- For each station having 10 minutes of data centering on the predicted P time,
- find the predicted P time,
- request the 10 minute data window.
- Find the nearest neighbor station and channel that
- lies within 15 degrees of the current station,
- has a station name that differs from the current station,
- the first 2 letters of the channel name match the current channel,
- the channel orientations are within 5 degrees for horizontal channels or the 3rd letter of the channel name is Z for verticals,
- has data available for the same 10-minute time window.

- Find the predicted P time of the nearest neighbor station.
- Decimate to reach a common sample rate if needed.
- Low-pass filter traces using a 0.01 Hz corner.
- If the channel is vertical and the metadata dip already describes the channel as phase reversed, multiply channel amplitudes by -1.
- Cross-correlate the channel pair time series:

Rxy(rΔt) = (1/N-r) * sum( x(n)*y(n+r) ) where N = # of samples and n = 1…N-r

- Report the maximum absolute value of the cross-correlation function:

Reported value is Rxy(r’Δt) where r’ is the lag at which |Rxy(rΔt)| is maximized.

### Metric Values Returned

**value** – signed cross-correlation function value at the lag having the maximum absolute value

**snclq2** – nearest neighbor, labeled as N.S.L.C.Q (Network.Station.Location.Channel.Quality)

**target** – the trace analyzed, labeled as N.S.L.C.Q (Network.Station.Location.Channel.Quality)

**start** – 5 minutes (UTC) prior to the predicted P wave from a M >= 6.5 event

**end** – 5 minutes (UTC) after the predicted P wave from a M >= 6.5 event

**lddate** – date/time the measurement was made and loaded into the MUSTANG database (UTC)

### Notes

Polarity values that are consistently negative for a station pair indicates that one of the two stations-channels may have reversed polarity, but it doesn’t indicate which one. It’s usually necessary to view first motions from events to resolve this ambiguity.