In the latest blogpost of his “GIS, a transformational approach” series, David O’Sullivan looks into transformations from fields1 to points and from fields to lines.
David centers the blog post on the analysis of a DEM2 and thus introduces several interesting concepts from geomorphometry3 – surface-specific points4, geomorphons (short for “geomorphologic phonotypes”), courses, thalwegs, and surface networks – demonstrates how to compute these features using R and highlights methodological challenges.
It’s a great concise introduction to some geomorphological concepts. The rest of the “Transformational” series (one article still to come) is also worth reading, particularly if you use R for geospatial tasks. And the article has me wonder about surface networks:
The surface network representation allows the terrain to be partitioned into hills (regions surrounded by thalwegs) and dales (regions surrounded by ridges). (…) surface networks deserve more attention than they’ve gotten from the community over the decades, but their reliable construction has proved elusive. Perhaps now that there is at least one tool out there that is up to the task they’ll make their way into the mainstream (…).
Footnotes
Think, for example, a digital elevation model or a rainfall intensity raster.↩︎
Digital elevation model↩︎
The science of measuring and quantifying terrain or the the shape of the Earth’s surface.↩︎
David’s article mentions this law on the number of surface-specific points (where n means “number of”):
npass − npit − npeak = 2
I remember this law being (implicitly) stated by J.C. Maxwell in his publication On Hills and Dales (I didn’t remember the year, so I looked it up: in 1870, based on earlier work by Arthur Cayley) – probably because of its simplicity but also specificity. I remember, because I cited Maxwell and Cayley in my PhD thesis.↩︎