[GRASSLIST:2117] Re: Landscapes and Graph Theory

[Rich Shepard]

On Wed, 18 Jul 2001, Roy Sanderson wrote:

> Has anyone seen the paper in by Dean Urban and Timothy Keitt in Ecology
> vol. 82 No. 5 (2001) pp 1205-1218 on "Landscape Connectivity: a
> graph-theoretic approach".  This looks like a good way to analyse
> connectivity between habitat patches, and identify key patches (graph
> nodes) and dispersal corridors (graph edges) for many species.  I'm
> currently using GRASS 4.2.1, but I can't see anything about graph theory in
> GRASS 5.0.  Anyone know whether this sort of approach can be implemented
> easily in GRASS, or is it a case of getting down to some low-level C coding?

Roy,

  I haven't read Ecology since I escaped from academia about 15 years ago.
:-)

  My response has more to do with ecology than with GRASS. But, you might
want to look at r.le (landscape ecology) to see if there's something useful
there.

  While I've not read this paper, your description brings to mind Levin's
loop analysis, McArthur's broken-stick/niche pre-emption hypothosis, the
application of information theory (from cryptanalysis) and thermodynamics to
species diversity and the application of economic theory to relationships
among organisms. All borrowed, you notice, from other sciences and
disciplines. Almost all are impossible to measure and/or interpret in the
real world. Good in theory but total failures in practice. I have a nagging
suspicion that graph theory falls in the same category.

  For what it's worth, I briefly looked at the potential use of graph theory
and chaos theory to ecology over a decade ago and decided that they did not
meet the test of field verification.

  I encourage you to read the talk I gave last month at the American Society
for Surface Mining and Reclamation: "If Biodiversity is the Answer, What is
the Question?". It's on our web site
<http://www.appl-ecosys.com/publications/> and the paper is assmr-2001.pdf.

  If you want to gain valuable insight into species distributions, relative
abundances, and ecosystem processes I strongly urge you to become proficient
in the application and interpretation of spatial statistics (also called
geostatistics). Rather than esoteric theory suitable for publication in
Ecology, it will give you real-life understanding of what's happening.

HTH,

Rich

Dr. Richard B. Shepard, President

                       Applied Ecosystem Services, Inc. (TM)
            2404 SW 22nd Street | Troutdale, OR 97060-1247 | U.S.A.
 + 1 503-667-4517 (voice) | + 1 503-667-8863 (fax) | rshepard at appl-ecosys.com
                 Making environmentally-responsible mining happen.