Projected SpatialLinesDataFrame has a segment spanning the globe

Ry Guy <rr2box42 <at> gmail.com> writes:

Hi All,

I'm trying to use the 'coastsCoarse' dataset from library 'rworldmap' as a 

base map for plotting a global

dataset. It works fine when I use unprojected data. However, if I project 

the data and base map,

coastsCoarse ends up with one horizontal line segment that spans the whole 

map. You can see what I mean like this:

Hi Ryan,
I think I've identified the offending segment (see below), but haven't yet 
been able to remove it. I'm sure someone else on the list better at 
manipulating sp objects will be able to help with getting around the error 
below :
'setting coordinates cannot be done on Spatial objects, where they have 
already been set'
Once we've fixed this I'll try fixing in rworldmap by removoing the 
corresponding point in the unprojected file.

Best wishes,
Andy

# ---------------------------------------
library(rgdal)
library(rworldmap)

data(coastsCoarse)
plot(coastsCoarse) # Looks good

mercator.proj4 = "+proj=merc +lon_0=0 +k=1 +x_0=0 +y_0=0 +ellps=WGS84 

+datum=WGS84 +units=m +no_defs"

coastsMerc = spTransform(coastsCoarse, CRS(mercator.proj4))
spplot(coastsMerc) # horizontal line across plot connects east and west 

Beringia...

#the problem is with line 94
plot(coastsMerc[94,])

#look at the last and the first point see which is the problem
coordinates(coastsMerc[94,])

#looks like the last point is the problem try removing it
numToRemove <- nrow(coastsMerc[94,]@lines[[1]]@Lines[[1]]@coords) 


coordinates(coastsMerc[94,]) <- 
coastsMerc[94,]@lines[[1]]@Lines[[1]]@coords[-numToRemove,]
#Error in `coordinates<-`(`*tmp*`, value = c(-1809707.97165956, 
-1823151.78413843,  : 
#setting coordinates cannot be done on Spatial objects, where they have 
already been set

It seems like the boundary of northern Russia didn't get cut off 

correctly. I imagine the fix is

straightforward, but I've been banging away at it to no avail. Does anyone 

have a simple solution that

either (a) locates and removes the offending pan-global segment or (b) 

doesn't produce it to begin with?

Thanks much,
Ryan

On 11/01/2013 02:15 PM, Andy South wrote:

Ry Guy 

<rr2box42 <at>
 gmail.com> writes:

Hi All,

I'm trying to use the 'coastsCoarse' dataset from library 'rworldmap' as
a 

base map for plotting a global

dataset. It works fine when I use unprojected data. However, if I
project 

the data and base map,

coastsCoarse ends up with one horizontal line segment that spans the
whole 

map. You can see what I mean like this:

Hi Ryan,
I think I've identified the offending segment (see below), but haven't
yet 
been able to remove it. I'm sure someone else on the list better at 
manipulating sp objects will be able to help with getting around the
error 
below :
'setting coordinates cannot be done on Spatial objects, where they have 
already been set'
Once we've fixed this I'll try fixing in rworldmap by removoing the 
corresponding point in the unprojected file.

Please let me know where you learned that you could use coordinates() to
assign coordinates to a Spatial* object.

Best wishes,
Andy

# ---------------------------------------
library(rgdal)
library(rworldmap)

data(coastsCoarse)
plot(coastsCoarse) # Looks good

mercator.proj4 = "+proj=merc +lon_0=0 +k=1 +x_0=0 +y_0=0 +ellps=WGS84 

+datum=WGS84 +units=m +no_defs"

coastsMerc = spTransform(coastsCoarse, CRS(mercator.proj4))
spplot(coastsMerc) # horizontal line across plot connects east and west 

Beringia...

#the problem is with line 94
plot(coastsMerc[94,])

#look at the last and the first point see which is the problem
coordinates(coastsMerc[94,])

#looks like the last point is the problem try removing it
numToRemove <- nrow(coastsMerc[94,]@lines[[1]]@Lines[[1]]@coords) 


coordinates(coastsMerc[94,]) <- 
coastsMerc[94,]@lines[[1]]@Lines[[1]]@coords[-numToRemove,]
#Error in `coordinates<-`(`*tmp*`, value = c(-1809707.97165956, 
-1823151.78413843,  : 
#setting coordinates cannot be done on Spatial objects, where they have 
already been set

a wild guess, but this might work:

coastsMerc at lines[[94]]@Lines[[1]]@coords <-
 coastsMerc[94,]@lines[[1]]@Lines[[1]]@coords[-numToRemove,,drop=FALSE]

It seems like the boundary of northern Russia didn't get cut off 

correctly. I imagine the fix is

straightforward, but I've been banging away at it to no avail. Does
anyone 

have a simple solution that

either (a) locates and removes the offending pan-global segment or (b) 

doesn't produce it to begin with?

Thanks much,
Ryan

R-sig-Geo@

-- 
Edzer Pebesma
Institute for Geoinformatics (ifgi), University of M?nster
Heisenbergstra?e 2, 48149 M?nster, Germany. Phone: +49 251
83 33081 http://ifgi.uni-muenster.de GPG key ID 0xAC227795

R-sig-Geo@

Thanks much to you both--Edzer's suggestion does work to remove the 

vertex. Based on the input I managed an alternative, which is to correct the
original vertex to not pass 180 degrees longitude (it's currently at
180.0000004, which is apparently the problem):

ind = which(coastsCoarse <at> lines[[94]] <at> Lines[[1]] <at> coords[,1] > 

coastsCoarse <at> lines[[94]] <at> Lines[[1]] <at> coords[ind,1] = 180

Cheers,
Ryan

Edzer Pebesma-2 wrote

Please let me know where you learned that you could use coordinates() to
assign coordinates to a Spatial* object.