The Open Earth Framework (OEF) A Data Integration Environment for Earth Sciences G. Randy Keller - Univ. Oklahoma Matt Fouch - Arizona State Univ. Chris.

The Open Earth Framework (OEF)

A Data Integration Environment for Earth Sciences

G. Randy KellerG. Randy Keller - Univ. Oklahoma- Univ. Oklahoma

Matt Fouch Matt Fouch - Arizona State Univ.- Arizona State Univ.

Chris Crosby Chris Crosby – SDSC– SDSC

Chaitan Baru Chaitan Baru – SDSC– SDSC

Dave Nadeu Dave Nadeu – SDSC– SDSC

John Moreland John Moreland - SDSC- SDSC

Motivations

• Integration of multidisciplinary data sets is essential to understanding the complex processes operating at the Earth’s surface and within its interior.

• Our current ability to collect massive amounts of digital data, develop structural models from these data, and generate high-resolution numerical models of dynamics is very well developed.

• Conversely, our ability to quantitatively integrate these data and models into holistic interpretations of Earth systems is poorly developed.

3

EarthScope Program http://www.earthscope.org

• 3.2 km borehole into the San Andreas Fault• 1099 permanent GPS stations• 74 borehole strainmeters • 5 laser strainmeters• 100 Permanent seismic stations• 1000s km2 high resolution topography/InSAR

• 400 transportable seismic stations occupying 2000 sites

• 30 magnetotelluric systems• 100 campaign GPS stations• 2146 campaign seismic stations• Geochronology

Study the three dimensional structure and evolution of the North American Continent

EarthScope Data Portal

portal.earthscope.org

CYBERINFRASTRUCTURE FOR THE GEOSCIENCES www.geongrid.org

Data layers - Input to aid in the construction of

3-D and ultimately 4-D models DEM (USGS, SRTM)Geology (mostly 1:500,000)

Landsat 7 / ASTER / LIDAR

Magnetic anomalies Gravity anomalies

Petrology/Geochron (e.g. NAVDAT)Drilling data (State surveys, USGS)

……….To construct 3-D models, start To construct 3-D models, start with tomography; add gravity, with tomography; add gravity, geologic interfaces, seismic geologic interfaces, seismic interfaces, ….interfaces, ….

Provide input Provide input for modeling for modeling of processesof processes

A Scientific Effort VectorA Scientific Effort Vector

Background Background ResearchResearch

Data Collection and Data Collection and CompilationCompilationSoftware IssuesSoftware Issues

ScienceScience

Back-Back- groundground

ResearchResearch

Data Collection Data Collection and Compilation and Compilation

Software IssuesSoftware IssuesScienceScience

ScienceScience - Analysis, Modeling, Interpretation, Discovery

Data Integration as a Workflow• All data integration activities can be characterized generically as

workflows that typically involve running the data through a series of processing stages to:– Find the data– prepare the data, – remove outliers, – format and filter data, – grid the data – derive other data products, – visualize the results, – produce a proposed “model” for a given region

• The integration process is necessarily iterative, leading to progressively refined earth system models, but it is rife with possible stumbling points and inefficiencies.

Data Integration Choke Points• Frustrations and inefficiencies that come from wrangling and

integrating disparate data to build a coherent model. • Bottlenecks in the process of going from disparate sets of data to

integrated models are workflow “choke points” that stall processing when data does not flow easily from one stage to the next.

• Custom ad hoc software “hacks” to stitch together tools and push past these choke points.

• Working through and tolerating such choke points is sometimes viewed as a “rite of passage” and necessary training.

• Learning to manage and prepare data is useful, but ultimately time would be better spent on analyzing data and building comprehensive models.

Back-Back- groundground

ResearchResearch

Data Collection Data Collection and Compilation Software and Compilation Software

IssuesIssuesScienceScience

“PhotoShop Science”• Choke point at the very end of the workflow, just shy of a

publication-worthy diagram.

• Screen shots & outputs from different tools overlain them by hand in software such as PhotoShop to obtain the desired result.

• Difficult to reproduce a similar figure with improved data, and we have to regenerate the figure from the beginning, through all of the workflows again.

• “PhotoShop science” is a problem to be solved rather than lived with.

Seismic Seismic tomography result tomography result from the CD-ROM from the CD-ROM projectproject

Ken Dueker, University of Wyoming

Use open source software and openly available dataUse open source software and openly available data

Start with a tomographic model

Add interfacesbased on geophysics

Add geologic detail

Add topography

Finish and place in a

regional context

The OEFThe OEF

3-D is essential in

geology

Is this a plume, an Is this a plume, an intrusion, a salt intrusion, a salt dome, or a reef?dome, or a reef?

The ultimate goal in geophysics is..The ultimate goal in geophysics is..Construction of 3-D volumes with as many physical Construction of 3-D volumes with as many physical

properties as possible assigned to each volume elementproperties as possible assigned to each volume element

Discontinuities are also important, and we need Discontinuities are also important, and we need to be able to insert them and manipulate themto be able to insert them and manipulate them

We also want the We also want the results to be results to be

compatible with compatible with various modeling various modeling programs (e.g. programs (e.g. groundwater, groundwater, geodynamics)geodynamics)

Figure by M. Billen, UCDavis

A number of geophysical techniques can produce 3-D voxel models (e.g., tomography), and others produce interfaces. The big challenges are to include interfaces in voxel-based models and to be able edit and visualize the models as one proceeds.

Reflection Reflection seismology seismology

provides an image provides an image of the subsurface of the subsurface whose geologic whose geologic interpretation is interpretation is often obvious.often obvious.