Ugly horizontal line in map

[zecamuelbert]

Hi, Dan.
Not sure if this has to do with the bug reported for the EOW issue. But, a horizontal line is showing in the map I'm drawing. It appears when I draw the coastline after using mapImage.
The code:

par(mar=c(4, 4, 1.5, 1.5))
data(coastlineWorld)
data(topoWorld)

ram_sites <- tbl_df(read.csv("~/PROJETOS/OTN_CNPq/ReNOMO/Questionario/ReNOMO-RAM_geral.csv", stringsAsFactors = FALSE, check.names = FALSE))

lonlim <- c(-67, -24)
latlim <- c(3,-34)
#topo <- decimate(topoWorld, 2) # coarsen grid: 4x faster
#topolat <- subset(topoWorld, latlim[1] < latitude & latitude < latlim[2])
topo <- subset(topoWorld, lonlim[1] < longitude & longitude < lonlim[2])

#faz o mapa
mapPlot(coastlineWorld,  longitudelim = c(-50, -40), latitudelim = c(3,-34), projection="+proj=mill", col = 'lightgray', drawBox = TRUE, axes = TRUE)
breaks <- seq(-5000, 1000, 100)
mapImage(topo, col=oce.colorsGebco, breaks=breaks)
mapLines(coastlineWorld) # here is where the problem appears.

#plota pontos pretos grandes
mapPoints(ram_sites$Longitude, ram_sites$Latitude, col = "black", type = "p", pch=17, lwd=3)
mapText(ram_sites$Longitude, ram_sites$Latitude, ram_sites$Instituicao, pos=4)

Any advice?
Thanks, Zeca

[dankelley]

Hi Zeca, can you email me (dan dot kelley at dal dot ca) your code, and the data file(s) involved? It's easier when I can see things myself. My guess is that it is the "ugly lines" issue. Thanks! Dan. (PS. snowstorm here today. I'll send you a neat video of it, in reply to your email.)

[richardsc]

I can confirm that it is the UHL (ugly horizontal lines).

There are two ways that I know of to fix it:

1. call mapPolygon() instead of mapLines()
2. use the coastlineCut() function.

I made a minimum working example from your code^1, and while method 1 seems to work, for reasons I don't understand method 2 doesn't ...

[dankelley]

You're, fast, Clark! I suggest code like below.

library(oce)
par(mar = c(4, 4, 1.5, 1.5))
data(coastlineWorld)
data(topoWorld)
lonlim <- c(-67, -24)
latlim <- c(3, -34)
topo <- subset(topoWorld, lonlim[1] < longitude & longitude < lonlim[2])
mapPlot(coastlineWorld,
    longitudelim = c(-50, -40),
    latitudelim = c(3, -34), projection = "+proj=mill", col = "lightgray", drawBox = TRUE, axes = TRUE
)
breaks <- seq(-5000, 1000, 100)
mapImage(topo, col = oce.colorsGebco, breaks = breaks)
mapLines(coastlineWorld)
# Zeca, try doing something like this, instead of the line above.
cl <- subset(coastlineWorld, -70 < longitude & longitude < -20 & -40 < latitude & latitude < 10)
mapLines(cl, col = 2, lwd = 3) # make wide red line to show clearly

[dankelley]

Also, Zeca and Clark, here's a snapshot of a video I made this morning. It is the entrance to the Gorsebrook Junior High school (not sure I have the name right). Either school is cancelled today or somebody came late and made this cool snow art with their footprints. I love it!!

[dankelley]

Also, Zeca, I noticed that there was no lat axis, so I added that. Also, I'm using polygons, like Clark suggested, so the land will be more obvious. Finally, note that I added some things to the projection specification. There are a LOT of choices for projection (there's a vignette on that, and its in my book too, plus many blog postings by Clark and me and others).

library(oce)
par(mar = c(4, 4, 1.5, 1.5))
data(coastlineWorld)
data(topoWorld)
lonlim <- c(-67, -24)
latlim <- c(3, -34)
topo <- subset(topoWorld, lonlim[1] < longitude & longitude < lonlim[2])
mapPlot(coastlineWorld,
    longitudelim = c(-50, -40),
    latitudelim = c(3, -34), projection = "+proj=mill +lat0=-20 +lon0 = -45",
    col = "lightgray", drawBox = TRUE, axes = TRUE
)
breaks <- seq(-5000, 1000, 100)
mapImage(topo, col = oce.colorsGebco, breaks = breaks)
cl <- subset(coastlineWorld, -70 < longitude & longitude < -20 & -40 < latitude & latitude < 10)
mapPolygon(cl, col = gray(0.9))
mapAxis(2)
mapGrid()

[zecamuelbert]

Gee, you guys were fast!!
The suggestions worked fine. I got rid off the UHL. Thanks.
But, Dan. I still not getting the lat axis. I thought using 'axes - TRUE' would do the job. And, then I introduced the 'mapAxis(2)' you suggested. Not sure what is happening.
Also, is there a way to get a higher resolution for the bathymetry to get rid off the pixels.
Thanks,
Zeca

[dankelley]

1. I think axes=TRUE should make the axis appear. I'll look at that over the weekend, opening an issue probably. But you can alwasy add axes later so it is a bit of a low-priority thing.
2. For higher resolution -- oh, yes, you can get very high resolution. Try using download.topo() from the oce package, or dod.topo() from the dod package (installable with devtools::install_github("dankelley/dod")). Warning: looking at high resolution topography can be addictive. I could stare for hours at the amazing images you can get. The datasets are amazing. And, of course, if you want very very very high resolution, there can even sometimes be advanced maps that show down to centimeters...

[zecamuelbert]

I actually have 'axes=TRUE'. There was a typo in my msg.
But, no worries. This is something that I can add latter. I also need to explore a better projection.
I'll be careful with high resolution image.
Thanks,
Zeca

[dankelley]

I can never resist ... look at this cool image, and this is just with the default res of 4 minutes. (See the relict river values south of the Cabot Strait zone.)

PS. I know it's bad to use those colours but they let me see things.

library(oce)
data(coastlineWorldFine, package = "ocedata")
topoFile <- download.topo(-63, -50, 40, 50, resolution = 4)
topo <- read.topo(topoFile)
cm <- colormap(zlim = c(-5500, 0), col = oceColorsTurbo)
imagep(topo, colormap = cm)
polygon(coastlineWorldFine[["longitude"]],
    coastlineWorldFine[["latitude"]],
    col = "gray"
)

[richardsc]

For completeness, I'd also point out that the marmap package connects to the same NOAA etopo data, and any object that is created from a marmap download can be converted to a topo object with as.topo().

[zecamuelbert]

I like it. Already enhanced by map!!
Still needs some fixes. Will do it over the weekend.
Thanks Dan and Richard for this fun exchange!
Zeca

[dankelley]

Oh, you may now be addicted to looking at bathymetry !!!

If you think we have an answer, perhaps click one of the comments to indicate so, and close the discussion, so we can tell that things are settled now. If new things arise, please open up new discussions or issues. (PS I will open an issue tomorrow about the map axis but I'm doing something else today.)

[zecamuelbert]

Yes, it is addictive.... like the first gray scale satellite images when I was at Dal!!!

It is my first time using Github discussion. Not sure about the protocols. Will mark as settled!
Thanks!