Scientists talking to scientists…a 19th century behavior in need of a 21st century update

The past thirty years or so has seen an explosion in the ways scientists try to communicate with the public.  Blogs, planetarium shows, movie consulting, making promotional videos, testifying to school boards, running teacher workshops, writing popular books, making museum displays all have been gaining popularity, not to mention the rather explosive growth in ‘educational’ cable channel science shows. We can save for another day a discussion on whether any of this works, but how about communication between scientists?

Arguably this has been going downhill.  In the 18th century, individuals wrote each other letters and would publish books. In the 19th century, journals got going (which really began as writing a letter to a group).  In the 20th century we largely added the review process to journals.  And in the 21st century we have…tweets? Actually, we have a strangely dysfunctional system.  You can contact nearly any scientist via email in minutes.  You can read most research papers from your computer (provided you have access to the journals in question). The one thing we do have are data centers which, presently, focus on primary data.  These are a wonderful advance, but these are not the means by which scientists convey their discoveries.  No, we still use journals (or, if you are feeling more cutting edge, you post a preprint on ResearchGate or some similar platform).

Here’s the funny thing.  As we’ve moved to electronic journals (GSA’s Geosphere and AGU’s G-cubed are just two examples of journals envisioned from the start as e-journals) the striking thing is that these papers look almost precisely like the old papers in the old journals.  With one notable exception: journal articles increasingly fail to provide intermediate results or sufficient detail on how they got their results. In the olden days papers would have tables with results of analysis: for instance, seismic refraction survey might show the seismograms, would have a set of picked arrival times, and would have a set of inferred apparent velocities, all of which might be written down in the table.  From this, the author would infer some seismic velocity structure. If you wanted to check the author’s work, you could look at the seismograms to see if you agreed with the picks, you could plot the picks yourself and see if you agreed with the set of apparent velocities, and you could calculate the velocity model that should be consistent with the apparent velocities.

The same kind of work today, with lots more seismograms, lots fancier tools for identifying arrivals, usually produces a paper with figures of the seismograms overlain by the arrivals predicted by the seismic model and a cross section of the seismic model. Usually there are some aspects of the study that will merit detailed mention (perhaps trying to see how sensitive the final model is to assumptions of gradients or something like that), but if you want to reproduce this work, um, good luck. You probably can get the waveforms from a database, but then you are off to reproduce the whole study on your own.  Even if you want to just use the final wave speed model, you probably are going to have to pick off the values from an image (and, yes, GG has done this).

This sucks. We’ve crippled the ability to build on others’ work.

One solution is to encourage all the intermediate stuff to be put into supplemental materials, and this is a help (especially if the code used to fit the picks is there).  But can we imagine something better?

Imagine reading the article and thinking “you know, they don’t see this high wavespeed body that should be here”. So you go into the model figure and grab a line bounding a body and move it.  You can then see on the figure of the seismograms where arrivals should be and whether or not this works.  Or flip it around and change an arrival you don’t believe and see how that changes the model.

Impossible? Not today.  You could probably build this into a Flash animation; you could certainly do this with a Java application. (The nice thing about the Flash animation is that it can be embedded in a pdf). Something similar could be made for gravity and magnetics interpretations, and if you really want to have fun, having the ability to change assumed parameters in numerical inversions would be most enlightening.  Obviously there would be limits (else the publication would be a full-featured modeling program), but why not do something like this?

There are baby steps we can take today.  GG has a paper just out that has three dynamic figures embedded in the pdf version of the paper.  These are not as fancy as our goal above, but what they do do is allow the reader to consider a number of alternatives the author has explored, with a couple of tools to make it easier for the reader to consider these.  For instance, a cross-section tool allows presentation of many more cross sections than would be feasible otherwise, allowing the reader to see more than just the few most essential sections for the arguments in the paper and to sketch on the sections to compare between different alternatives.  This is a trivial tool to make (in essence, it combines a multilayer setup with a layer for sketching–easily adapted from suggestions on the web–and a simple calculation for converting x-y positions on a section to latitude and longitude).

There are plenty of other low-hanging fruit out there.  Simple x-y plots where points can be identified by clicking on them, or the regressions can be changed by excluding some points. Zoomable maps with navigation and legend tools (hover over the pastel green unit between the lime green and forest green to learn that it is the Pierre Shale and not the Lewis Shale, say). 3-D volume exploration tools.

Why do this in journals?  Because the journals are more long-lived than a professional’s personal website; one of the responsibilities they now have is to archive material (this is the cost that replaces the old cost of printing lots of paper).

So it is time to quit publishing 19th century papers and move forward. If you agree, agitate with your favorite journal.  Over time, GG will make some how-to web pages for the simplest of these tools….

Advertisements

Tags: ,

Trackbacks / Pingbacks

  1. All the Science that’s Unfit for Print | The Grumpy Geophysicist - February 5, 2015

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: