Wednesday, March 20, 2013


Just now I'm sitting in the Denver airport after attending the SAGEEP meeting run by EEGS (environmental and engineering geophysical society). This is a annual meeting of scientists working on shallow geophysical problems. It is a small group, maybe hundred 100 people at this meeting. The entire membership somewhere between 400 and 600. They are struggling with the opportunity for a potential merger with the SEG organization. This group has been active on its own for nearly 30 years and you can imagine the concerns of becoming part of a much larger organization.

But I did not come here for politics. My interest was in seeing some small vendors in the exhibition hall who either would not be found at the annual SEG meeting or be lost in the shuffle at such a large affair. In this I was successful.

I had a chance to sit down and have some beers with people from Sensors and Software, a company from whom I had just purchased a GPR system. The issue was how to survey quickly across good terrain like that we might find in a rock quarry. Turns out they have a small wheeled vehicle beautifully suited to the purpose, including a distance activated trigger on the wheel. The deal was made.

Another need we had at University of Arkansas is software to build a near surface model from the kind of seismic dated generated by a 48 channel system and hammer source. In the high production world of commercial 3-D seismic data, this job falls to tomography software which is largely automatic. The tomography tends to generate smooth near surface models devoid of hard contacts and layers. This is fine as an effective medium solution to do statics for deeper reflection seismic data. But my work on the outcrop and in the shallow subsurface recognizes there are discrete layers and refractors we want to map as accurately as possible. This harkens back to the early days seismology when people did serious refraction interpretation, a skill largely lost in production 3-D seismic shops these days.

The most rigorous form of these refraction analysis solutions is the Generalized Reciprocal Method (GRM) of Derrick Palmer published in the 1980s. It is brilliant, but difficult to understand and implement. Thankfully Charlie Stoyer (ex-CSM professor) has done this laborious job for us. I sat down with Charlie for 30 minutes, we talked through the software (including an example), and came up with a nice, realistic near surface model. The workflow is involved and requires significant human input for data quality judgment and parameters. Exactly the sort of thing I want my students to understand. The interface is clean and tight, and allows quick QA/QC from shot records to to final model. I noticed some small tick marks on the final model and asked Charlie about them. He said "Oh, those are uncertainty bars. I can remove them." I said "Don't you dare, I want my students to see them." The deal was made; another sale.

Finally, I talked to a company which makes a trailer mounted thumper seismic source. This is a necessary addition to our arsenal of hammer source and Vibroseis unit available through the UArk civil engineering department. But I was late for a lunch meeting with Mike Sullivan of FairfieldNodal, a meeting later with John Stockwell of CSM, and dinner that evening with my daughter Sam in Boulder. The thumper company will send me a quote so we can communicate and figure out a way to close this deal too.

Other societies such as a APG and SPE hold regional meetings throughout the US and the world (my fledgling research group has submitted seven abstracts to the midcontinent AAPG meeting in Wichita this fall). Regional meetings don't replace national meetings, they compliment them, allowing workers dealing with regional problems to meet on a smaller scale, a more intimate scale, and vendors to deal with that group specifically.

SEG may want to reconsider the regional meeting model, or lose that market to others who do.

John Stockwell in his domain (below)

Thursday, March 14, 2013

Mathematica Strat Column

Update 3/25/2013

In my new life as the only geophysicist in a geology/geography department, one of my first jobs is to figure out the local stratigraphy.  These are the rocks we plan to investigate by outcrop and near surface studies, ultimately building a 3D digital geomodel.  The formation of interest just now is the Mississippian Lime, a hot oil target in Oklahoma and Kansas.  Now there have been stratigraphers working around NW Arkansas for over a century, may of them still alive and kicking.  I am the new kid on the block and sure to step on some toes as I try to understand the stratigraphic section.

First I scanned all available digital theses from U Arkansas geology students, then checked the state Geological Survey and recent guidebooks.  Lots of general agreement, but differences in detail.  Then I found this amazing 1970 report from the Picher mine district in far NE Oklahoma. This summarized 130 years of mine geology experience and analysis.  Those miners were working the area like an ant farm and gaining an intimate knowledge of the Boone (Miss Lime) formation.  Surface geologists over the same time have subdivided the Boone into 'upper' and 'lower' with an thin oolitic limestone in there somewhere.  The mine team subdivided the Boone into 6 lithostratigraphic units, which no one ever brought up to the surface.

But my plan is to take a fresh look at the Boone/Miss with modern technology and tools.  The 6 subdivision are useful as a starting point to divide the sections into units that someone might see in well logs and core, sections that someone might care about in the petroleum exploration business.  Note my use of the word 'unit', that avoids the uber-defined terms 'member' and 'facies'.  Surface geologist are very strict with nomenclature.

Anyway, I made a column in excel first.  The nature of a spreadsheet was perfect for this, each member was a couple of cells high and the hierarchy of stratigraphy could build up across columns.  Figure 1 shows this stratigraphic column.

Figure 1. Excel stratigraphic column for NW Arkansas with members having equal height.

Next I went into Adobe Illustrator and laboriously (think 13 hours) converted this to a depth column.  It is shown in Figure 2. 

Figure 2. Same column with members shown in true thickness.

Beautiful, but what happens when some experienced stratigrapher says: "No way, the Batesville is way thinner than that!" Another 12 hours? Not for this cowboy. So I thought about this and concluded that a strat column is just a bunch of rectangles in columns, all linked up.  It occurred to me that Mathematica's manipulate function could be used to build an interactive stratigraphic column.  The finest subdivision of the column is the member, everything else is made up from aggregation of members into formations, series, and systems.  So I spend some time (think 5 hours) building the widget with each member having a thickness sliding bar.  The default values represent the rock column beneath Old Main, except for the Bloyd formation that is missing.  Bloyd thicknesses are given as average values around NW Arkansas (Figure 3)

Figure 3. Interactive stratigraphic column coded in Mathematica with thickness (feet) slider bars.

To show how the whole thing adjusts, if the Chattanooga is much thinner we get Figure 4.

Figure 4. Mathematica stratigraphic column automatically adjusted for Chattanooga Shale thickness of 10 ft rather than the default 60 ft shown in Figure 3

It may not be quite as pretty as the hand-drawn version, but changes take an instant.  And as the stratigraphers beat on me, the code can be modified to my current understanding of the stratigraphy.  

Of course, I am only the latest to try and correlate rock units between regions.  There are some sacred documents that have to be considered.  First is the North American Stratigraphic Code (AAPG, 2005), which actually is some pretty interesting reading in places.  But even to summarize it here is out of the question, the minutiae of stratigraphy is here.  Read it if you want to know what all the arguing is about around the next geology field trip campfire. 

Every state in the USA has a geologic map, carefully prepared over many decades from legions of field workers.  These describe surface rocks, but the state maps rarely fit together.  Infamous 'state line faults' are more common than one would think; places where a geologic unit changes name or assigned age across a state boundary, standing as mute testimony to intellectual battles fought long ago. But progress is being made on this venerable and cantankerous problem. An effort is well underway to stitch together surface geology maps, not just across the United States, but the entire world (OneGeology, 2013). It is wildly ambitious, their motto "making geological map data for the Earth accessible" says it all.  But, in fact, the web site is a bit difficult to navigate, somewhat buggy, and quite slow.  Strange as it may seem, the best way to get at this treasure trove of global geological information, and more, is an iPad app called EarthObserver developed by the Lamont-Doherty Earth Observatory at Columbia University (LDO, 2010).  This amazing App draws on the OneGeology database to give you global geology maps, gravity, magnetics, tectonic plates and boundaries, spreading rates, and more.  Anyone teaching introductory tectonics could build the entire course from EarthObserver.  One could complain it needs better referencing; state and country outlines, latitude and longitude, that sort of thing.  But this is quibbling, it is fantastic.

For example, suppose you are up one night wondering what is the fastest plate boundary velocity on earth, and where is it?  The EarthObserver image in Figure 5 shows a good candidate not far from Samoa.  Maybe I should go there sometime and check it out with the money I saved buying the EarthObserver App.  It costs 99 cents.

Figure 5. EarthObserver screen shot of some fast-moving plates (24 cm/yr) near Samoa.  How fast is 24 cm/yr? It is about the speed as human hair growth, or about 6 times faster than the moon is receding from the earth due to tidal friction.

References cited:

AAPG, 2005, North American stratigraphic code, AAPG Bulletin, v. 89, no. 11, pp. 1547–1591
north american stratigraphic code - USGS National Geologic Map ...

OneGeology, 2013, Making Geological Map Data for the Earth Accessible

LDO, 2010, EarthObserver App: The whole world in your hand