The current Accretionary Wedge made me wonder: what qualifies as an important geological experience? My days in the field, hauling equipment and running logs? Preparing my first publication or presentation? The many geological wonders I’ve been able to visit and explore? I’ve been lucky to have so many great experiences in the geosciences! One important experience that comes to mind, though, is geophysical modeling during my master’s degree.

My thesis involved investigating contaminated groundwater moving from a closed landfill to a nearby lake. Did this plume have a geophysical signature that could be used to map it?(This map could then inform decisions about where to focus future cleanup activities.) The investigation used a variety of geophysical methods on land, in boreholes, and on the nearby lake.

(2D resistivity profile.)

Sample 2D resistivity inverted model. Pretty, but is it any good?

The glory of portable computers and ever-increasing computing power and storage is that we can now collect a lot of geophysical data in the field very quickly. Powerful software programs easily import these data and provide default settings for processing, analysis, and display in order to generate fast “results.” I was able to quickly import the data and produce stunning visualizations of the data, like this two-dimensional (2D) electrical resistivity inverse model.

If It Seemed Too Easy…

After the initial glow of “wow, that was easy!” I realized I was far from done. I had so many questions:

  • How did the model correspond with other geophysical data and local stratigraphy?
  • Was the model meaningful?
  • Was the model equally valid and reliable in all areas?
  • How does the software get from the initial data I provide to the inverse model?

I had so many questions, but not enough answers! I dug into the software. I learned more about the modeling algorithm. I learned about the variety of settings you could use when running the model. I learned about sensitivity analyses and maximum/minimum versions of the model values.

I also dug into case studies of 2D resistivity surveys at other contaminated sites. I re-reviewed case studies I had already read and found new ones. I realized that most of these case studies didn’t discuss sensitivity analyses or modeling parameters. Did they run the models relying on software default settings? Or did they customize the model to the individual survey and site hydrogeology and not mention it? Since it wasn’t documented in the articles, there was no way to know.

I re-ran the model – and then I ran it again and again, as I learned how to more effectively use the data collected in the field to generate a more accurate model of the distribution of electrical resistivity in the study area.

(2D resistivity profiles.)

Image shows three profiles: The minimum and maximum electrical resistivity models that fit the data and the best-fit model.

It’s Only a “Black Box” if You Don’t Look Inside

Near-surface geophysics sometimes gets a bad reputation as a lot arm waving around “black box” tools. I realized that the same convenience offered by increasingly sophisticated software can be the user’s downfall: it is too easy to run programs that produce great visuals, without ever having to understand what the software is actually doing. If we don’t take the time to understand the tools and software we use, then yes, it is like working with a mysterious box and we should question the results. We must take the time to fully understand the tools we use, whether we’re talking about a compass in the field or software in the office. It’s a lesson that has served me well, whether I am working with my own data or evaluating data analysis and modeling conducted by others.

Side Note:

This is my first Accretionary Wedge contribution. I’m very excited to get involved in this regular geoblogging event. I want to thank some of the other contributors this month for providing some inspiration and ideas for my own post. Folks have posted about: