Yikes, there are so many exciting things to put in such a database it is almost overwhelming. I think it would need to have searchable sections: on rock mechanics (rock type/location and the measured physical properties of said rocks), locations, ages, and formations.

Data on

joint ages/orientations

fold ages/orientations

fault ages/orientations

strain analyses

depth

dike ages/orientations

. . .

Yes! Thank you much, and I agree. So far the testing database structure accounts for all these features, their orientations, and the material – probably not much on mechanical strength measurements, simply because they aren't included in most orientation data collected... But, could definitely be referenced to data regarding the average/relative strength of certain lithologies. I'll think about how I might integrate such an idea. The strain analyses in a useful geographic reference-frame are also less abundant (available ones will be included), but such other features you listed could be used to evaluate strain ellipsoids for different scales of deformation with clever querying. And I completely agree that the number of things/questions to investigate is almost endless.

What I am still wondering is: what are some specific questions/tasks that *require* such a large relational database to be in place to answer? I am basically trying to further justify this work and have questions in place that demonstrate/test it's usefulness in a Big Data style. I have a short list going, but I am trying not to bias the potential discussion.

Thank you again, and I would enjoy any other thoughts.

I think the justification is 1) Much of this data is not immediately identifiable in the accessible record, this database reduces redundant studies, 2) NSF requires that all data be archived for public use now and 3) for large scale questions. For example: I have a student interested in how the break-up of the Farallon plate around 80 Ma affected western North America. In a database like this I would search for joint/fault/fold orientation data for 90 and 075 Ma for western North America. I would create maps showing the location and strain for each time period and compare them.

Oh, and four, for teaching. How wonderful it would be to have the students be able to conduct their own research projects using readily available data! Charleston, SC does not have accessible rocks for quite a distance so even if we do a field trip with rocks, we don't usually get good data. This would allow me to assign specific, real-life research projects! Yowza

Have you used GeoMapApp? This is a graphical means of looking at geophysical data that can be very helpful for teaching

:-)

Fantastic.

Indeed one motivations is the new NSF requirements. A question related to that has been if-and-how to incorporate old or dark data into the new database – some folks say we should only worry about moving forward. We are working on extracting those existing data so they can be queried alongside new datasets.

GeoMapApp is great.

Hi Zachary, here are other more general examples of Big Data  analysis applied to SG&T issues: i) sismology - large geophysical databaseshelp identifying the physical variables that control earthquake mechanics, ii) petrology - the equations that describe metamorphic reactions in a P-T-x space are usually extrapolated by regression of multiple datasets. The accuracy and precision of these regression lines could be improved using very large datasets, and other 'dark' reactions could be evidenced this way to.

If this achieved, it helps and answer all questions in our minds about applied geology and can serve in academic purposes.....