THE SCEC COMMUNITY MODELING ENVIRONMENT: A CYBERINFRASTRUCTURE FOR EARTHQUAKE SCIENCE

The Southern California Earthquake Center (SCEC) has formed a partnership comprising SCEC universities, the Information Sciences Institute, the San Diego Supercomputer Center, IRIS, and the USGS to develop a “Community Modeling Environment” for seismic hazard analysis (SHA). The SCEC/CME project is funded under the National Science Foundation’s Information Technology Research Program, and its goals are formulated in terms of four "computational pathways." Pathway 1 is an SHA computational framework that supports a variety of earthquake forecast models and intensity-measure relationships. Pathway 2 utilizes the predictive power of wavefield simulation in the construction of intensity-measure relationships. Pathway 3 incorporates fault-system and rupture-dynamics models into earthquake forecasts. Pathway 4 assimilates various types of data into the structural representation of Southern California required by the other pathways. This presentation will summarize our progress in developing the software modules for each of these pathways and in erecting the cyberinfrastructure needed to automate their configuration and execution. To support computationally intensive simulations that generate large data volumes, we have established a virtual organization utilizing Globus middleware that includes grid-based connectivity and services shared between SCEC, USC, SDSC, and Pittsburgh Supercomputer Center. Knowledge-based systems have been developed to serve two important functions: a Grid MatchMaker utilizes ontology-based descriptions of grid resources and computational programs to schedule jobs, and a Composition Analysis Tool reasons across semantic descriptions of computational models and data types to assist users in specifying valid computational pathways. Data are stored in a digital library managed by the Storage Resource Broker, which provides a robust system for maintaining the association between the data and metadata. Visualization software supports analysis and validation of the data sets. We will illustrate how the CME is functioning as a collaboratory for knowledge synthesis, hypothesis formulation and testing, data assimilation, and prediction.