This file includes speaker notes that are in

This file includes speaker notes that are in the “Notes” module of PPT (go to View--->Notes Page)

Developing a Computational Environment for Coupling MOR Data, Maps, and Models: The Virtual Research Vessel (VRV) Prototype Jan Cuny Doug Toomey U of Oregon Dawn Wright Oregon State Judy Cushing Evergreen State

Best studied fast-spreading ridge segment Wealth of data, results, models under-utilized . . . formats, standards, tools incomplete/incompatible

hydrothermal activity/convection (geologists & geochemists) physical structure of axial magma chambers (seismologists)

Vision for VRV: A Computational Infrastructure • MORE than just archiving…. • data sharing, tool composition, and model coupling – physical observations (traditional data) – text attributes, video and graphics – programs, models, tools, and scripts for computational processing • New data and metadata, format conversion • Web interface for distributed computing

Good Fit to NSF ITR • Computer science clearly needed – Improvements to current technologies • Interdisciplinary, multi-institutional team, history of collaboration • EPR yes, but other sites (e. g. , Galapagos) and types of environmental data as well • Human resource development (undergrads, VRV-ET, “Saturday Academy”) • research plan "compelling" but obviously too ambitious!

Three Components (Solutions) 1 - Data Sharing • GIS, RDBMS, computational experiment management system (Vi. NE) are all needed • Non-spatial data and text metadata • Computational experimentation • More than physical access to files – More than flat files and simple tables

Arc. IMS Zoom in Query, simple analyses, add your own data

So far. . • Dawn – Our vision & NSF ’s ITR – The data sharing problem – GIS data visualization • Judy – Tool Composition & Model Coupling – Educational outreach – Expected outcomes

2 - Tool Composition for “Computational Steering” Experimental Data Processing Ocean Data Mat. Lab Adjust constraints Parameters Geodynamic Application Seismic Velocity Model Parameters Seismic Velocity Model Viz Mat. Lab Visualize model space Add physics Published result

Tool Composition Building a Computational Experiment

Tool Composition with Vine Describing Data for an Experiment

3 - Model Coupling -“Super. Models” start image mantle structure flow models seismic anisotrophy models melt generation regions mantle streamlines

Model Coupling Creating a “Super Model” • Steer a single model (Vine), • Launch that steering (Vine) across platforms, • Transfer data seemlessly across platforms • Describe the models « declaratively » – input, parameters, process, output • Describe « Process Interactions »

Model Coupling Launch Computational Steering across Platforms

Data Models and Databases Physical Access to Ridge Data MATLAB Computational Steering & Model Coupling Web Browser Seismic Anistrophy Model JDBC Driver Le Select program wrapper Flow Model Le Select Communication Modules JDBC SQL Engine Job Mgr data wrapper EPR Endeavor Vents Le Select view wrapper Ridge Global Schema

Data Models and Databases (prelim) Common Semantics (EPR & Endeavor)? • Location • Time. Stamp • Event • Observation

VRV - ET (Educational Tool)

Expected Outcomes Integrating data with metadata, tools and models - A (possibly virtual) database - Tools to visualize data (GIS and Mat. Lab) - Tools for Steering & Coupling - Publish models - Compose tools - Support migration paths for model coupling • Apply all to VRV for EPR • Educational Outreach -- VRV ET – UOregon, Portland Sat. Academy, Evergreen, etc.

Methods for Model Coupling Express model couplings so they can be implemented as coupling between simulations. • Use simulation code analysis and theoretical tools such as Petri Nets to express these couplings. • Describe models so that the coupling can be automated and model descriptions can be reused. •