Introduction
Soon after map projections were added to Oce, bug reports showed that coastline plots in some projections were subject to anomalous lines that run horizontally on the plot. A adhoc scheme was code to try to prevent this, but it does not always work. Problems are compounded for filled coastlines.
I had thought this was a basic problem of all projection coding, until I happened to try using the projr
package instead of mapproj
to carry out the projections. The result is that the annoying lines went away.
I am still thinking about whether to make a transition in Oce. The proj4
syntax is too tricky for simple inclusion in Oce, and tricky, and users (including me) are accustomed to the existing syntax, so it would be necessary to convert the internal linkage in a somewhat nontrivial way. Nevertheless, the point of the present post is simply to illustrate that the problem is solved, at least in one test case.
Test case
First, load libraries and extract the coastline.

library(oce)

## Loading required package: methods
## Loading required package: mapproj
## Loading required package: maps

library(proj4)
library(mapproj)
data(coastlineWorld)
lon < coastlineWorld[["longitude"]]
lat < coastlineWorld[["latitude"]]

Next, plot with existing (mapproj) projection.

par(mar = c(3, 3, 1, 1), mgp = c(2, 0.7, 0))
xy < mapproject(coastlineWorld[["longitude"]], coastlineWorld[["latitude"]],
proj = "mollweide")
plot(xy$x, xy$y, type = "l", asp = 1)

Finally, plot with proposed (proj4) projection.

par(mar = c(3, 3, 1, 1), mgp = c(2, 0.7, 0))
xy < project(cbind(lon, lat), "+proj=moll")
plot(xy[, 1], xy[, 2], type = "l", asp = 1)

Conclusions
At least in this example, the proj4
package produces better coastlines, somehow being clever enough to cut the polygons that cross the “edge” of the earth.
I will do some more tests to see if this is a fluke case, but if I think proj4
is promising, I will see how to invent a scheme for handling the mapproj
arguments called parameters
and orientation
with proj4
. THis seems to be a bit tricky, at first glance, but let’s keep the cart behind the horse for now.
Resources
 Source code: 20140410ocemapprojection.R
Rbloggers.com offers daily email updates about R news and tutorials on topics such as: Data science, Big Data, R jobs, visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...