To properly represent the status of current and historic data for the GSN, we will gather and publish performance metrics for the GSN to include specific waveform-related quality assessments, in addition to the metrics already calculated and reported at the IRIS Data Management Center (DMC) which include data availability, latency, and timing quality, We plan to include the following metrics:

  • Noise: deviations from long-term noise characteristics of data streams. This will be based on the probability density functions (PDFs) calculated on the power spectral density data stored at the IRIS DMC and the USGS (links point to PDF images).  A description of the ambient noise PDF technique is provided by the USGS
  • Sensor response accuracy: Confidence in the metadata accurately describing the actual system response for a particular data stream.
  • Linearity: The confidence we have that the sensors are responding linearly. That is, the amplitude and phase response are independent of amplitude. Results can be found in Hutt and Ringler, 2011.
  • Accuracy of Orientation: These data are currently available at the WQC page and are based on polarization anomaly work cited in Ekstrom and Busby, 2008.

Based on these metrics, we intend to publish a historical record for every channel of data for the GSN that describes our confidence in the quality of these data. The purpose of this historical time series will be to allow data users to determine if the GSN data they are utilizing in their analyses are of the highest quality possible by a number of standards. If there is a problem with a particular dataset, we hope that cross-referencing to the quality metrics will minimize the occurrences that may otherwise be misconstrued as a seismological phenomenon. As many groups utilize massive amounts of data in bulk analyses, a quantified quality metric may allow the pre-weighting of data by these metrics to minimize the adverse effects of sub-quality data.