[R-sig-Geo] A practical guide to geostatistical mapping

Rich Shepard r@hep@rd @ending from @ppl-eco@y@@com
Fri Aug 31 15:14:48 CEST 2018

On Fri, 31 Aug 2018, Tomislav Hengl wrote:

> I am working on a rmarkdown update of the practical guide to GM (the
> working title is "A practical guide to spatial prediction with R"). As
> soon as I clean up these reports on my desk I will get on it (in about 2
> months should be online). My apologies to you and all other users for
> untidy website / my tardiness.


   This is excellent news!

   I sent you a message yesterday at the e-mail address on the web site only
to have it bounce as 'user known in this domain.'

> In the meantime you might also find this useful:
> https://envirometrix.github.io/PredictiveSoilMapping/

   Thank you.

   I don't know if r-sig-geo is an appropriate forum for my questions that
are not specific to applying R packages, but to geostatistics themselves. If
not, I'd appreciate a pointer to a more appropriate place.

   One thing I've noticed in my readings about geostatistics is that (quite
appropriately) most are research oriented and written by academic
geostatisticians and ecologists. My work as an environmental science
consultant, an applied aqutic ecologist who left academia for the private
sector several decades ago, means that all data available to me are
generated by regulatory requirements, not by the needs for a research
project. And, the overwhelming number involve aquatic chemistry (and biota
such as fish) which adds the constant movement of the medium into

   This makes it difficult for me to translate examples such as the Meuse
example in Chapter 5 to my projects. Currently I'm looking at mercury
concentrations in a river system and the sampling locations have an
intereesting pattern of clumps on the mainstem by major tributaries and are
otherwise quite dispersed. As a non-mathemtical statistician I've currently
no idea how to conduct exploratory analyses on such data. And, there's the
temporal aspct to be considered, too.

   I've much to learn because there's a real need for the application of
spatio-temporal statistics in regulatory environmental science instead of
the deterministic, differential equation models currently demanded by

Best regards,


More information about the R-sig-Geo mailing list