Are You Ready? Introducing Research Ready Data Sets (RRDS)
The IRIS DMC is pleased to announce the Research Ready Data Sets web service (RRDS). This service makes it easy to select seismic data based on waveform quality metrics, greatly reducing the time and effort a researcher might spend on preprocessing of raw data. By specifying filtering criteria in an RRDS request, researchers can custom-design their own definition of quality and thus zero in on the important stuff: data subsets worthy of the “research-ready” designation.
RRDS incorporates the MUSTANG system’s collection of quality metrics such as ambient noise, dead channel identification, RMS variance, and data continuity. In combination with the more traditional selection parameters (e.g., station, location, time-range), the researcher can build finely-tuned requests, customized for particular fields of study.
As an example, a seismologist seeks vertical channel BHZ data from the Global Seismograph Network (GSN – IRIS/USGS) on the day of the Nepal M7.8 earthquake of 2015-04-25 (at 06:11:25 UTC) that meet the following quality criteria:
|maximum daily STA/LTA is greater than 20||max_stalta_gt=20.0|
|channel had no PDF-mode value exceeding 5dB below the NLNM (within the 4 to 8s period band)||noise_mode_diff_gt=lnm-5,4s,8s|
|no gaps encountered within a 24-hour window||num_gaps_eq=0|
The URL for such a request to the RRDS service will be:
For the equivalent request in POST style, you can submit a text file with the following content:
max_stalta_gt=20.0 noise_mode_diff_gt=lnm-5,4s,8s num_gaps=0 IU * 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00
The service evaluates these metrics and returns filtered data in what is known as “request format”. Here is a portion of that response:
nodata=404 IU ADK 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 IU AFI 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 IU ANMO 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 IU ANTO 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 IU BBSR 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 IU BILL 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 IU CASY 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 ...
Such a filtered response can then be submitted to one of our data access services to get the research-ready data of interest. For instance, POST-ing this request response to the IRIS dataselect web service by way of a script makes it easy to set up a quality-controlled data access pipeline.
The RRDS service also provides a format parameter which, if set to “report”, will include a log of which data were rejected by which criteria plus the data quality values used in the determination.
RRDS CRITERIA: max_stalta > 20.0 num_gaps = 0 noise_mode_diff_gt lnm-5,4s,8s RRDS REJECTION LOG: IU HRV 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 max_stalta=19.431 IU JOHN 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 max_stalta=17.681 IU KNTN 00 BHZ 2015-04-25T00:00:00 2015-04-26T00:00:00 max_stalta=16.663 ...
It’s our hope the research community will find the RRDS tool an exciting new addition to the suite of IRIS web services. We are planning to integrate its capabilities into DMC-supported data request tools (such as FetchData) to explore the potential of automated “research-ready” systems.
by Rob Casey (IRIS Data Management Center) and Mick Van Fossen (IRIS DMC)