Saturday, March 27, 2010

Depositional facies models - the need to move on?

I am reading John S. Bridge's River and Floodplains, which I highly recommend for anyone interested in fluvial processes and the resulting deposits.  His comments on depositional facies models deeply resonate with me:
Most facies-association models in the literature are inadequate for anything more than a very general qualitative environmental interpretation because: (i) the deposits are normally represented in only one or two vertical sections, normal and parallel to the flow direction; (ii) the information shown in these two-dimensional sections lacks critical detail of the spatial variation of strata thickness and orientation, grain size, internal structures, paleocurrents, and biological features; (iii) the nature of preservation of the strata is not indicated, because the models are not dynamic; and (iv) the models are purely qualitative and many do not even contain scales.  The reasons for this state of affairs are: (i) incomplete knowledge of depositional environments (ii) incomplete knowledge of the evolution of depositional environments and their deposits over time periods in excess of a few hundred years; and (iii) uncertainties in translating essentially two-dimensional (parallel to land surface) information representing a single snapshot in time into a fully three-dimensional dynamic model.
I really agree with Bridge's comments and strongly believe that the principal focus on descriptive and qualitative, amplified by the proliferation of conceptual models in sedimentary geology was and is detrimental to the advancement of the science.  Which reminds me of a Twitter exchange I had with @clasticdetritus a while ago.  It started from my venting over how overwhelmingly descriptive and qualitative sedimentary geology is, and how beneficial would be to have a much stronger link between designing (quantitative) experiments to test concepts.  The Twitter thread is shown below, in "proper stratigraphic order", with the newest comment at the top.

This was a fun exchange, unfortunately for me it was so late in the day, and the 140 character limitation was not helping.
There is a solution to this problem, and I will cite again from Bridge, because he expressed it so well:
The solution to these problems require much more serious, comprehensive studies of modern processes and deposits, supplemented by scale-model experimentation and quantitative theoretical modeling
What are your thoughts on this?

7 comments:

  1. I was addressing this on Twitter ... figured it'd be easier to do it right here! :)

    Firstly, I agree wholeheartedly.

    But, a couple of things I'll add.

    (1) I would argue that the community has recognized this and the past decade (2000-2009) has shown a huge leap regarding researchers utilizing both numerical and physical experiments. That is, sedimentary geologists and geomorphologists are working together much more and some younger researchers combine multiple disciplines very effectively.

    (2) As always, it depends on the questions one is asking with a model. Yes, many of these qualitative/conceptual models are grossly inadequate to address specific questions related to physics of sediment transport and morphodynamics of sedimentary bodies. But, does every geologic question require this level of rigor? Does every geologic question that involves the interpretation of ancient sedimentary environments require a fully 3-D dynamic model to sufficiently put the question being asked into context?

    In other words, qualitative/conceptual models have value!

    But, that's not to say we should be satisfied with them. Further work of the physics and dynamics of sedimentary systems will lead to *better* qualitative/conceptual models, but they won't, and shouldn't, go away.

    ReplyDelete
  2. Clasticdetritus, I agree with both your points, and I recognize that conceptual models have value. I still hold though that too many conceptual/qualitative models are not rigorous enough and do not stand when tested against even simple process-based principles. I also think that the over-emphasis on qualitative has important implications of how observations are recorded and data collected in the field, which in turn makes it so much harder to link the conceptual with numerical. Thanks for your great, engaging comments.

    ReplyDelete
  3. Is there any certain paper/study you can point me to that is a good example of the kind of model you want to see more of?

    ReplyDelete
  4. Since I brought up Bridge in my post, some examples from his published work would give you an idea: the companion papers Bridge&Mackey (1993) and Mackey&Bridge (1995); Bridge et al (2005) published in JSR on rocks in San Jorge Basin.

    ReplyDelete
  5. Can't seem to find the 2005 JSR paper ... my search keeps coming up w/ this 2000 one: http://jsedres.geoscienceworld.org/cgi/content/abstract/70/2/341

    Is that the one? If not, can you paste the link or full citation in here?

    ReplyDelete
  6. Some of the SAFL folks did some work testing the LAB (Leeder, Allen, Bridge) alluvial models over last decade ... e.g., http://jsedres.geoscienceworld.org/cgi/content/abstract/75/4/710

    ReplyDelete
  7. Sorry, my mistake, it is 2000 and your link identifies the paper I meant. SAFL work is v. good as well. To change the environments a little, Kneller's work is another good example. More examples to come (hopefully) soon :)

    ReplyDelete

Related Posts Plugin for WordPress, Blogger...