Databases in the Field: A Broader Perspective

Anne Meltzer, Lehigh University
Chair, PASSCAL Standing Committee

IRIS/PASSCAL has been very successful in developing a facility to promote portable array seismology. Its success is measured by the continued demand on the instrument pool, the quality and quantity of data archived, and the advance of our science through publication of experiment results in scientific journals. The equipment in the PASSCAL facility represents a significant community resource. The quality of the data collected by PASSCAL experiments using this resource is such that it is and will continue to be of interest to investigators for many years, reaching far beyond the intent of the original PIs. PASSCAL has two priorities, the support of experiments and PIs in the field and the timely delivery of PASSCAL data to the DMC for archiving and use by the broader IRIS community. The archiving of data, whether from the GSN or PASSCAL is a primary commitment of IRIS and is an integral component of our Cooperative Agreement with the National Science Foundation. No one experiment justifies the tremendous resources NSF has invested in our community. Collectively we strive to justify this investment. Equipment in the PASSCAL facility is available to any research or educational institution, free of charge. Use of the equipment includes the technical support required to maximize the success of each experiment in the field. The only stipulation placed on those borrowing instruments from the PASSCAL facility is that they agree to submit their data to the DMC for use by the community at large after a two year proprietary period.

One reason PASSCAL has been so successful is that it has enfranchised a large number of seismologists at a wide range of institutions to conduct field experiments. No longer are field programs restricted to seismologists at only the largest research institutions. Part of PASSCAL´s mission is to continue to develop and provide hardware solutions and software tools that make the job of collecting data in the field easier and more fool proof. A recent development in the software domain is a suite of tools encompassed by the PASSCAL database (PDB) software. These tools provide a uniform integrated interface to help QC experiments in the field: view log files and waveforms, monitor network uptime, track instruments, make timing corrections, associate events, produce event gathers and produce archive volumes for submission to the DMS.

During PDB development, PASSCAL requested that three different 1996 experiments (Abitibi, Iceland, and Nanga Parbat) beta test the system. The Abitibi experiment consisted of 30 broadband stations for a six month period. One person did most of the work to service stations, quality control data, and produce archive tapes. I was the PI on the Nanga Parbat experiment and we also successfully used the PDB in the field. We operated a 60 station array for four months in the Himalayas of NE Pakistan. We certainly had no access to Internet or reliable phones. We serviced our stations on a three week interval and over half our stations could only be accessed by a 3-5 day trek. We had a PASSCAL field computer system installed in a rented house in Gilgit. It was clear, even from these first beta tests, that the PDB can be used effectively in the field to help with experiments. In my experience, not only did the PDB software not jeopardize the field component of the experiment, it was an asset. We found the PDB helpful for monitoring the uptime of our array, viewing log files, and identifying problems with instruments and timing. By continuing to update our network configuration, inconsistencies in field notes were identified and resolved immediately rather than at some later date when memory begins to fade. Perhaps most importantly, we were able to generate archive volumes for the DMC and time corrected associated events tapes for our own analysis in a timely fashion. We were able to begin the analysis phase of our project almost immediately after leaving the field rather than having to deal with merging field notes, correcting, and formatting data post-experiment.

To use the database software successfully, it is important for PIs to coordinate with the PIC during the planning stage. A database specific to each experiment can be built and installed on the field computer before the experiment begins. The database software is a sophisticated software package, but a day of pre-experiment training goes a long way toward understanding how to use it. It has a GUI interface and feedback from PIs continues to improve the ease of use of the software. Actually running the database software and updating network configuration files as necessary does not add significant time to data download procedures if it is integrated into the data processing flow in the field lab. If PIs let a significant amount of data build up, it is easy to fall behind. Generating SEED volumes does take time, but it takes time whether you do it after the experiment is completed or during the experiment. In addition to helping QC the experiment, one of the biggest benefits is in the amount of time saved post-experiment. If run correctly, PIs leave the field with a set of tapes ready for archiving at the DMC and a second set of tapes ready for data analysis. Archiving data at the completion of the field experiment allows the PI to take advantage of the incredible resources available at the DMC to produce network day volume or event volumes. This is a significant improvement over past operations where PASSCAL PIs spent months or sometimes years reducing and formatting their data for analysis. A quick scan of PASSCAL data sets recently contributed to the DMC shows experiments post-PDB have been submitted within a few months of when the experiment left the field compared to PASSCAL data acquired pre-PDB, which often took up to 6 years to be submitted.

It is not mandatory to use the PDB. It is mandatory for all passive source PASSCAL experiments to submit data to the DMS in SEED format. PIs are free to use any software they choose to produce SEED volumes. The PDB software does provide an effective means for producing archive volumes for PIs who might not otherwise have this capability. Since the three beta test experiments, nine experiments have successfully used the PDB in support of field experiments.

IRIS PASSCAL is a university consortium, not a commercial operation. We try to run the PASSCAL Instrument Centers in an efficient manner while still servicing the community needs. We tend to keep personnel costs to a minimum, maximizing our ability to continue to purchase additional instruments. This can lead to a certain amount of frustration for both instrument center personnel and PIs. However, it is quite clear that much of PASSCAL´s success is due to a working partnership between a talented and dedicated technical support staff and creative PIs. As we progress, experiments become more ambitious and complex. We stretch the envelope with each new experiment to maximize the amount of data recorded. We feel that tools like PDB go a long way toward increasing the success ratio of each experiment. The need for fairly sophisticated software in support of field experiments is a reality and will become increasingly important as we move toward the next generation instrument and real time systems. The notion that we will continue to track our experiments by hand is unrealistic. The experiments that are successful today benefit from the accumulated experience of previous experiments, just as the experiments that will run next year benefit from the advances provided by today´s experiments. PASSCAL is a community run facility. Our ability to advance and improve is dependent on input from the community. We appreciate the opportunity to engage in open debate and look forward to feedback from the community.

Return to: IRIS Newsletter Information
Return to: Title Page and Table of Contents
Continue to: Next Article