[R-sig-Geo] Gstat 'polylines' implemented in R?

Tim Peterson timjp at unimelb.edu.au
Tue Sep 15 02:19:54 CEST 2015



On 04/09/15 18:50, Edzer Pebesma wrote:
>
> On 09/04/2015 06:31 AM, Tim Peterson wrote:
>> Hi all,
>>
>> I'm keen to use the gstat feature 'polylines' (appendix B in
>> http://www.gstat.org/gstat.pdf) to restrict the point sample selection
>> to within a polygon. That is, when undertaking localised kriging at a
>> grid cell, samples are only used in the estimation if a line between the
>> sample and grid cell does not cross a polygon boundary. The polylines
>> feature appears to be implemented within the c-code for gstat
>> (polygon.c) but I cannot find any means to use it within R-gstat nor any
>> mention on lists. Any tips would be most welcome.
> That is correct, it has not been interfaced from the R code.
>
> The idea is nice, but if you work with it you see very very much
> unwanted features in the interpolated surface: sharp edges that follow
> the "shadow" of the polygon, as seen from an observation with an extreme
> value. Back then we thought about a work-around for this using distances
> measured "around" the obstacles, but that leads to non-positive definite
> covariance functions.
Thanks the background - and thanks again for your contribution to the list.

I suspect my application is considerably simpler than those for which 
polylines was tested on. I'm trying to take account of coastlines 
whereby I do not want observations that are on one side of a bay (as in 
coastline) to be used in the kriging estimation of the adjacent side of 
the bay. Code I've written in GSLib and application to mapping the 
watertable of greater Melbourne did not show artefacts. Anyway, it seems 
to me that get_n_edges() is the C function called within select.c for 
achieving such a restriction on the point selection. If I get the time, 
I'll investigate using a 'set' command for controlling polylines within R.

>> Thanks,
>>
>> Tim
>>
>> ----------------------
>> Dr. Tim Peterson
>>
>> The Department of Infrastructure Engineering
>> The University of Melbourne, 3010 Australia
>> T: +61 3 8344 9950 <tel:%2B61%203%208344%209950>, M: +61 0438 385 937
>> <tel:%2B61%200438%20385%20937>
>>
>> Dept. profile :
>> http://www.ie.unimelb.edu.au/people/staff.php?person_ID=141135
>> Research Gate : https://www.researchgate.net/profile/Tim_Peterson7
>> Google Scholar:
>> http://scholar.google.com.au/citations?user=kkYJLF4AAAAJ&hl=en&oi=ao
>>
>> 	[[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo

-- 
----------------------
Dr. Tim Peterson

The Department of Infrastructure Engineering
The University of Melbourne, 3010 Australia
T: +61 3 8344 9950 <tel:%2B61%203%208344%209950>, M: +61 0438 385 937 
<tel:%2B61%200438%20385%20937>

Dept. profile : 
http://www.ie.unimelb.edu.au/people/staff.php?person_ID=141135
Research Gate : https://www.researchgate.net/profile/Tim_Peterson7
Google Scholar: 
http://scholar.google.com.au/citations?user=kkYJLF4AAAAJ&hl=en&oi=ao

	[[alternative HTML version deleted]]



More information about the R-sig-Geo mailing list