Subject: eq_response Resent-Date: Mon, 08 Nov 1999 17:08:04 -0500 Resent-From: Geoffrey Fox Resent-To: p_gcf@npac.syr.edu Date: Fri, 05 Nov 1999 17:22:04 -0800 From: "Kenneth J. Hurst" To: Geoffrey Fox Scenario 1 - Response to an Earthquake in Southern California In this example the goal is to rapidly form a consensus among researchers concerning the characterization of the deformation field and the location, size, and direction of slip on a fault following an earthquake. This consensus can be used to guide decisions on both civil and scientific responses to the quake. Following an earthquake in Southern California, the location and magnitude are calculated based on seismic data within minutes by Caltech/USGS, and currently are broadcast to several users via email and pagers. The information on location and magnitude could then be automatically used to define an area wherein instruments might be expected to record a signal. (disp) Data from these stations would be given priority in retrieval and analysis. In this example we will assume that the data in question is GPS data from the Southern California Integrated Geodetic Network (SCIGN) array. Retrieval in this case is done by telephone modem. As soon as the list of possibly affected stations has been generated, the database at the USGS is checked. If any of the stations on the list have not had data downloaded since the quake, computers at the USGS begin dialing the selected stations and retrieving the data. Data from these stations would then be processed for rapid analysis to determine the measured displacements of the stations. (GIPSY) If the measured displacements are large enough, emergency and scientific personnel and are notified via email and pager. These displacements are then automatically fed into an inversion routine (simplex) which solves for the best fit single fault displacement. This single fault displacement is in turn fed back into a forward elastic half space model which yields a preliminary map of displacements over the whole area. (disloc) At this point this map is shared between various scientists and emergency personnel, using TANGO. TANGO allows the collaboration and interaction of multiple people viewing and manipulating the same data set over the Internet. The emergency personnel can use the preliminary map in combination with GIS information about utilities, lifelines, etc. to help assign resources to various areas. The scientists will use the preliminary map to help design a strategy for collecting additional measurements. They can also collaborate on refining the single fault model, possibly breaking the single preliminary fault into several segments, introducing more realistic material properties, or including more data, before rerunning the inversion. This environment permits the rapid determination and dissemination of preliminary information about the earthquake and the collaborative refining of that information following an event. The rapid dissemination of information can greatly aid both the civil and scientific response to the quake. Resources can be more efficiently allocated to the areas where they are needed, and scientific measurements can be focused to provide information critical to refining our understanding of the earthquake system. Once an acceptable model of the earthquake has been determined, various models can be used to estimate the updated earthquake hazard for adjacent areas. Since there are currently several competing models for this it will undoubtably involve multiple runs of multiple models and significant discussion among scientific colleagues. Each of these models as well as the various pieces of the automated processes described above have been developed by different people under different assumptions, and is developed, run, and maintained on computers under the control of the developer. Technologies such as CORBA and EJB allow the maintainers of these computers to permit or deny access to collaborators in a rational way. A recent example of how a system like this could have been useful was provided by the 1999 Izmit Turkey earthquake. Following that earthquake, many geoscientists got together in a series of conference telephone calls to try to piece together what had happened, and what was an appropriate response. Some participants initially only knew what had been reported in the media. Others knew of specific pieces of data concerning the earthquake or of actions being taken by various groups and individuals. It is safe to say that no one had a complete picture. Much of the conference call was devoted to informing everyone about all the pieces of data and all the various initiatives that people were pursuing or might pursue. Similar calls and emails occurred after the 1992 Landers and 1994 Northridge earthquakes. Having a system such as has been described above wherein participants could share maps, descriptions, programs, data sets, and graphs and wherein they could interactively and collaboratively manipulate the data and programs both synchronously and asynchronously would immeasurably aid the rapid and accurate diagnosis of what has happened and what should be done next. Ken Hurst Mail Stop 238-600 voice: 818-354-6637 Jet Propulsion Lab / Caltech FAX: 818-393-4965 Pasadena, CA 91109 hurst@cobra.jpl.nasa.gov