Mapping failure,advanced edition

GG has complained a few times about poor maps in the professional literature, but a new entry has carried this to a new level. A new and intriguing paper by Alessandro Verdecchia and Sara Carena contains this figure:


This has the now-familiar “computer stupid” projection that GG complained about (and others have defended): 1 degree of longitude is shown as having the same length as one degree of latitude, a relationship only true at the equator. The scale shown is correct for north-south distances, but overestimates east-west distances by about 20% or so.  Strike 1.

What carries this to the next level is that important information here is the orientation of the faults. A Mercator projection will preserve angles (that is what it was always good for), but this projection will not.  A fault trending N45E, for instance, will be drawn here as trending about 52 degrees away from a north-south line. Strike 2.

The stereonets shown for focal mechanisms are not so distorted, so you cannot compare the orientation of nodal planes with the trends of faults as accurately as you should be able to. Strike 3.

Grr. As GG has said before, there is no excuse in the age of GMT and ArcInfo for stuff like this.  We can hope that the authors’ codes do not mirror their abuse of mapping software.

This again is the sort of mistake (and there is no defending this one!) that biases GG to question the ability of authors producing such maps. This is not what you want to do as an author!

[UPDATE: See Sara’s comment below; the statements above are too harsh as they presume ignorance or laziness rather than a choice made for other reasons.  GG continues to disagree with the choice, but now recognizes this was a choice.]

Hopefully we’ll get a chance to come back and explore the suggestion here that the data available allow us to refine earthquake forecasts in complexly deforming regions.

Tags: , ,

3 responses to “Mapping failure,advanced edition”

  1. Sara says :

    Concerning this particular figure, it doesn’t just have a “computer-stupid projection”: it is an unprojected map, and the same-size degrees of lat/lon are a dead giveaway. I can also tell you that even for latitude the scale is not that accurate (it is probably off by 5-10 km just because of the change in line thickness during redrawing), and the beachballs have been redrawn too (i.e. they are not accurate either, especially in where the arcs intersect the circle). It is a 10×15 cm location figure merely for showing the location of a list of objects, their approximate size, and the type of focal mechanisms. The actual fault geometry is stored as a file and can be downloaded.

    That said, I disagree with the very idea of specific map projections in papers. I much prefer unprojected maps, because I can then project them as needed if I want to compare them to something else or measure anything. Instead, what many people do is project a map (because that’s the good thing to do, or so they’ve been told), and then invariably fail to state projection and datum (because they do not actually understand their importance). So it becomes very difficult for someone else to reverse-engineer it. It takes just 60 seconds to project an unprojected map in whatever projection you see fit, but a projected map with missing information is unusable. Unless you are working at latitudes above 60 degrees, dealing with an entire continent, or writing a paper about circular vs. elliptical craters, projected maps are more of a hindrance than a help.

    By the way, projected maps in papers do not bother me nearly as much as official government maps (USGS maps first and foremost) that lack projection and datum information. It is by no means a rare occurrence: most maps I have attempted to use lacked the basic information necessary to reproject them and required in the best case a lengthy trawling through other map sets to track it down, and in the worst case guessing dozens of possible projection/datum combinations. If these maps had not been projected, they may have looked a bit distorted, but they would have been easier to use 30 years later, when nobody remembered any longer the projection habits of the day.


    • cjonescu says :

      Thank you for the comment and the perspective of why an unprojected map might be preferred (not to mention showing that I was wrong about this being indefensible). I would never have guessed a motivation of “user can reproject however he or she wants.” I assume that you are Sara Carena, second author on this paper [I am beginning to think that geoscientists are more touchy about map projections than almost anything else, and really hate seeing “computer-stupid” instead of “unprojected”–many of the substantive comments on this blog have been from upset authors defending maps–“computer stupid” was a term I started using when computer-generated figures first started showing up and their appearance did represent laziness; I guess I should abandon the term given the intentional use by some authors].
      Despite that, I do continue to disagree. First, reprojecting a published map is one of the least efficient and least accurate means of getting the relevant information, so publishing an unprojected map is kind of pointless. As this paper does have very nice supplementary files (thank you), I would go there if I wanted to plot this data differently. Having encountered many badly drafted maps where locations would be in error by large amounts were I to extract them from a map, I’d only do this as a last resort. But second, and more importantly in my view, figures should be there to support the scientific points being made. The example in question is more than what I would consider a simple location figure, and it invites comparison between the beachballs and the faults (which is why I brought it up rather than the other unprojected maps), so choosing a projection (probably Mercator here) that makes such a comparison easiest is, I would argue, the better path to follow. I think it interesting that this kind of shows how a reader reacts differently to a figure than the authors might have intended (this is a space where editors and reviewers can be helpful).
      And I really disagree that the distortions south of 60° latitude are easily overlooked. If you put a scale bar on a map, it should mean something, and here it clearly does not mean anything. For large scale maps I can live with a subset I call computer-stupid Mercator (another term sure to anger), which is just squishing an unprojected map so the scale is the same north-south and east-west. But at small scales you really want to avoid distortion (well, *I* really want to avoid distortion). So it sounds like we are doomed to disagree.
      But you raise an excellent point, one where I am guilty, is the absence of projection information on published maps. Long ago projection information was more common. I think it would be beneficial to encourage inclusion of projection parameters somehow. It is a bit awkward in captions, but should be present, perhaps as a footnote. Even worse for really large scale maps is the absence of the datum (and this can persist into data tables). Having tried to locate a seismograph that had the wrong datum attached to its location, I can tell you datums matter.
      I will confess to being puzzled about the claim that a lot of USGS maps lack projection information. This has not been my experience at all (well, mostly–maps in professional papers often lack info), and I have reprojected many such maps for other purposes, so I am curious what series of USGS pubs lack map projection info.
      And if you want to see a lot of really crappy maps, pick up just about any history book and you’ll find “maps” that challenge the idea of “map.”
      Thanks for the visit and the comment.


Trackbacks / Pingbacks

  1. Mapa Culpa | The Grumpy Geophysicist - December 18, 2017

Leave a Reply

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

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

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s

%d bloggers like this: