On Thu, 5 Dec 2013, Francesco Carotenuto wrote:
Thank you for the answer. The object is the prediction location in an
interpolation process. The function of "overlay" is involved in an
automatic process and the longitude (and the kind of projection too) can
change according to the dataset that is, in some cases, worldwide
distributed.
This doesn't help much - do you mean that the grid is of prediction
locations? Are the polygons used for anything other than visualization? If
you change the visualization projection (shifting the meridian (+lon_0)
for worldwide distributions) repeatedly, you will have to clip to suit. If
visualization, does it make sense to modify the meridian (the reader may
see the shifts as confusing)?
You are trying to deal with multiple somethings that you are interpolating
(to a grid?) and displaying in possibly different windows possibly
depending on the extent of the prediction grid. Your workflow isn't
obvious.
You haven't explained why you need the projection so far (other than
aesthetics, maybe?). You may be able to interpolate on the sphere, so
unless the target grid cells must be of fixed area - and possibly not even
then, you don't need to make life complicated. There are problems in
creating global visualizations in any case, as the lines/polygons need
clipping at +/- 180 degrees from the meridian.
Roger
Dr. Francesco Carotenuto (Ph.D.)
Department of Earth Sciences
Federico II University
Largo S. Marcellino 10,
80138 NAPLES (ITALY)
http://francesco-carotenuto.blogspot.it/
http://3dpaleontology.blogspot.it/
Il Giovedì 5 Dicembre 2013 21:23, Roger Bivand <roger.biv...@nhh.no> ha scritto:
On Thu, 5 Dec 2013, Francesco Carotenuto wrote:
Dear list, as in
the title, I found that when projecting a world map by mollweide projection and
setting a?? longitude value, some strange
parallels are drawn. The real problem is that these lines ???interact??? with a
raster when performing the ???over??? function.
How can I manage
it?
The underlying question is what you want to do with the objects after
projection. The problem in your example is that some polygons/lines cross
-180+66 degrees, so get reflected back across. A possible solution is to
clip the polygons before projection at -180+66 degrees +/- a small number,
but before looking at this in more detail, I need to know what the
projected object is intended for. When +lon_0=0, there is no such effect,
as this data set is clipped at -180 already.
Roger
Here I provide a
simulated example to let you understand what I mean
Thanks.??
#Example of a
world map mollweide projection with a specific longitude value
library(sp)
library(maptools)
library(rgdal)
data(wrld_simpl)
w_pol<-wrld_simpl
w_pol<-spTransform(w_pol,
CRS("+proj=moll +lon_0=66"))
plot(w_pol)
??
# Intersecting a
grid with the projected world map
gr<-expand.grid(x
= seq(-180, 180, length.out = 100), y = seq(-90, 90, length.out = 100))
coordinates(gr)<-~x+y
proj4string(gr)<-CRS("+proj=longlat
+ellps=WGS84 +datum=WGS84 +no_defs")
gr<-spTransform(gr,
CRS("+proj=moll +lon_0=66"))
plot(gr[w_pol,])
??
??
Dr. Francesco Carotenuto (Ph.D.)
Department of Earth Sciences
Federico II University
Largo S. Marcellino 10,
80138 NAPLES (ITALY)
http://francesco-carotenuto.blogspot.it/
http://3dpaleontology.blogspot.it/??
[[alternative HTML version deleted]]
--
Roger Bivand
Department of Economics, Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: roger.biv...@nhh.no
--
Roger Bivand
Department of Economics, Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: roger.biv...@nhh.no
_______________________________________________
R-sig-Geo mailing list
R-sig-Geo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
