[This is a contribution to Accretionary Wedge #5]
One of the problems that a geologist is often faced with is the difficulty of reconstructing a complex three-dimensional geometry and history from limited information that is often one-dimensional (e.g., well data, cores) or two-dimensional (outcrops, 2D seismic sections). Humans in general, and geologists in particular tend to look for evidence where the light is better, and we are tempted to think that the beautiful core we have described, the one good outcrop face we have, the one textbook-quality seismic line on our wall is a good representation of the geology and stratigraphy of a much broader area, and that one can build a coherent story without knowing much about the third dimension.
That, of course, may well be true of ‘layercake’ stratigraphy: after all, a single thickness value can be used to fully characterize the geometry of a layer that has the same thickness over a large area. But, as Brian points out, ‘layercake stratigraphy’ should be considered an oxymoron: every sedimentary layer shows some thickness variations if traced for a long enough distance, even if some layers change their thickness more slowly than others. Stratigraphy is only layercake-like for human observers; subtle but persistent variations in thickness and relief can become striking geometries with some vertical exaggeration. Again, if this variation only occurred in one direction, a two-dimensional section along the same direction would summarize very well the whole story.
However, complex three-dimensionality is the rule rather than the exception in geology. Take for example a meandering river: its geometry is complex enough as it is, a single snapshot of a snaky morphology in time. But try imagining what happens as point bars and levees are deposited and cutbanks are cut; the channel changes its position over time and, over thousands and hundreds of thousands of years, it leaves behind an extremely complicated stack of deposits that would probably be difficult to fully understand even if you somehow could see and describe everything at the greatest detail in 3D. Obviously, a nice outcrop or a number of cores through such a deposit can provide a wealth of information, but we would be fooling ourselves if we thought that a single fining-upward sequence with some cross-bedding (that is, the classic point-bar facies model) was enough to understand a fluvial system.
But strong three-dimensionality is not restricted to fluvial deposits; look at any present-day depositional system in Google Earth and you will find that alluvial fans, deltas, barrier islands and tidal inlets, wind-blown dune fields are all intricate patterns, usually with lines running in more than one direction. Yet many of the classic facies and stratigraphic models are either one- or two-dimensional. Maybe, probably, these are necessary and useful simplifications and conceptual models, but they can only be useful if one is also aware how far they are from capturing the full 3D complexity of nature.
That being said, I have to add that 3D is not always better than 2D. Nowadays, some of the best three-dimensional geological datasets are 3D seismic surveys, and, with the increasing availability of such gold-mines of stratigraphic beauty (there are other uses as well, but let’s focus on one thing for now 🙂 ) it is easy to fall victim to the temptations of colorful three-dimensional displays. Despite claims like ‘3D interpretation and visualization are the future’, the truth is that a good set of old-fashioned maps and cross sections are more valuable in the long term than some glossy presentation slides with no exact spatial location.
Unless, of course, you can visualize and share your data relying on an easy-to-use and truly three-dimensional viewer. Like Google Earth. Even William Smith would be excited about that.
Detail from “Geological view and section through Dorsetshire and Somersetshire to Taunton, on the road through Yeovil toWimborn[e] Minster, &c.”, by William Smith, 1819. Source: Oxford Digital Library