I want to properly plot this netcdf file: http://www.filedropper.com/sshgridsv16092015060412nc (Downloaded originally from here: https://opendap.jpl.nasa.gov/opendap/allData/merged_alt/L4/cdr_grid/contents.html)
But run into issues:
I should be able to just plot the raster (the SLA variable):
library(RNetCDF)
library(raster)
library(maptools)
d <- raster("ssh_grids_v1609_2015060412.nc.nc4", varname = "SLA")
plot(d)
#plot SLA
But the result is very weird as you can see with the provided file.
Especially when plotting a world map on top:
data(wrld_simpl)
plot(wrld_simpl, add = T)
They don't match at all :/
So I thought maybe the problem is with the longitude (ranging from 4.839944e-09 to 360)
Then I read that raster::rotate(d)
should be perfect for that (to get longitude to -180 to 180), but it won't let me. I get this warning message:
Warning message:
In .local(x, ...) :
this does not look like an appropriate object for this function
and
plot(d)
still looks the same.
Any advice would be greatly appreciated!
Cheers
The Netcdf file is not only "rotated" in longitude but x and y are also at the wrong position. The way it has been entered in the netcdf is not usual apparently.
I downloaded the netcdf directly on the OpenDap server because your filedropper link seems to be corrupt.
Anyway, here is my proposition:
library(raster)
library(maptools)
d <- raster("ssh_grids_v1609_2015060412.nc.nc4", varname = "SLA")
# transpose x to y and double flip the map
m.r <- flip(flip(t(d), direction = "y"), direction = "x")
# then rotate from 0:360 to -180:180
rm.r <- rotate(m.r)
data(wrld_simpl)
plot(rm.r)
plot(wrld_simpl, add = T)
Related
I have a point shapefile ("search_effort.shp") that is highly clustered and an NDVI raster (resolution in m: 30.94948, 30.77829). I would like to subset my search_effort.shp by selecting 1 point per raster grid cell and create a new search_effort shapefile. I am using R version 4.0.3
I think I could have used Package ‘gridsample’ (in 'raster' v1.3-1), but it was removed from the CRAN repository and I would prefer not to use the archived version. Is there another way to do this in R?
I have also tried sample.grid but I do not know how to specify my raster as the grid, and have tried the following:
# NDVI raster to be used as the reference extent
NDVI_extent <-readGDAL('C:/Model_layers/NDVI.tif')
# Load the file names
layername <- "SearchEffort"
# Read in the shapefile
search_effort <- readOGR(dsn= ".", layer = layername)
plot(search_effort)
# Set the reference extent
r <- raster(NDVI_extent)
# Extract coordinates from the shapefile
search_effort#coords <- search_effort#coords[, 1:2]
#Subset points
sample.grid(search_effort, cell.size = c(30.94948, 30.77829), n = 1)
I get the following error:
"Error in validObject(.Object) : invalid class “GridTopology” object: cellsize has incorrect dimension."
I get the same error regardless of the cell.size I specify.
Example data
library(raster)
r <- raster(res=30)
values(r) <- 1:ncell(r)
x <- runif(1000,-180,180)
y <- runif(1000,-90,90)
xy <- cbind(x, y)
Solution
library(dismo)
s <- gridSample(xy, r, n=1)
Illustration
plot(as(r, "SpatialPolygons"))
points(s, col="red")
points(xy, cex=.1, col="blue")
I am converting many rasters to polygon. But in quite a few cases, I am seeing unexpected subgeometries, and I can't seem to get rid of them.
This is with R v3.3.3 and raster package v2.5-8.
Here is an example that should reproduce the problem I am having.
You can download the raster that I use here.
# first, read in raster and coarsen to something more manageable
library(raster)
library(rgeos)
env <- raster('adefi.tif')
env2 <-aggregate(env, 8)
# Reclassify such that cells are either 1 or NA
env2[!is.na(env2)] <- 1
# this is what the raster now looks like:
plot(env2)
# Now I convert to polygon, choosing to dissolve
p <- rasterToPolygons(env2, dissolve=T)
plot(p)
# I find that I can't get rid of these subgeometries
p <- gUnaryUnion(p) # identical result
gIsValid(p) # returns TRUE
I'm not sure where the problem is... Is it in how the raster package converts to cell polygons? Or is it how the rgeos package dissolves those cell polygons together?
Is there a work-around?
It looks like a projection issue. This works for me:
library(raster)
library(rgeos)
env <- raster(file.path(fp, "adefi.tif"))
env2 <- aggregate(env, 8)
env2[is.na(env2) == F] <- 1
# Project Raster
proj_env2 <- projectRaster(env2, crs = CRS("+init=epsg:3577"))
p <- rasterToPolygons(proj_env2, dissolve = T)
plot(p)
Not sure why the need for reprojection since epsg:3577 looks to be the same as the original projection, but I usually confirm projection using proj4string() or spTransform() to make sure everything will line up.
I have a raster in an equal area Behrmann projection and I would like to project it to the Mollweide projection and plot.
When I do this with the following code, however, the plotting doesn't seem right, as the map extends to the sides, and there are outlines of various landmasses where I wouldn't expect them.Also, the map extends beyond the plot window.
Can anyone please help me get this to plot nicely?
Thanks!
The data file used can be downloaded from this link.
Here is the code I have so far:
require(rgdal)
require(maptools)
require(raster)
data(wrld_simpl)
mollCRS <- CRS('+proj=moll')
behrmannCRS <- CRS('+proj=cea +lat_ts=30')
sst <- raster("~/Dropbox/Public/sst.tif", crs=behrmannCRS)
sst_moll <- projectRaster(sst, crs=mollCRS)
wrld <- spTransform(wrld_simpl, mollCRS)
plot(sst_moll)
plot(wrld, add=TRUE)
Alright, since the example at this page seems to work, I tried to mimic it as much as possible. I think problems arise because the far left and far right side of the raster image overlap. Cropping and an intermediate reprojection to Lat-Lon as in the example seem to solve your problem.
Perhaps this workaround can be a basis for a more elegant solution that directly addresses the problem, as it is not benificial to reproject a raster twice.
# packages
library(rgdal)
library(maptools)
library(raster)
# define projections
mollCRS <- CRS('+proj=moll')
behrmannCRS <- CRS('+proj=cea +lat_ts=30')
# read data
data(wrld_simpl)
sst <- raster("~/Downloads/sst.tif", crs=behrmannCRS)
# crop sst to extent of world to avoid overlap on the seam
world_ext = projectExtent(wrld_simpl, crs = behrmannCRS)
sst_crop = crop(x = sst, y=world_ext, snap='in')
# convert sst to longlat (similar to test file)
# somehow this gets rid of the unwanted pixels outside the ellipse
sst_longlat = projectRaster(sst_crop, crs = ('+proj=longlat'))
# then convert to mollweide
sst_moll <- projectRaster(sst_longlat, crs=mollCRS, over=T)
wrld <- spTransform(wrld_simpl, mollCRS)
# plot results
plot(sst_moll)
plot(wrld, add=TRUE)
We tried to reproduce the beautiful example of bl.ocks.org/diegovalle/5166482, using d3.js and leaflet, but with our own data, which is on a regular lon-lat grid.
In R, we first retrieve the data from a mysql table, and write them to a shapefile:
lonmin <- -10; lonmax <- 10
latmin <- 30; latmax <- 50
dlon <- dlat <- 0.5
lon <- seq(from=lonmin, to=lonmax, by=dlon)
lat <- seq(from=latmin, to=latmax, by=dlat)
nlon <- length(lon); nlat <- length(lat)
# cl <- a mysql request
solRad <- matrix(cl$solRad, ncol=nlon, nrow=nlat)
# Plot the data
levels=seq(from=-40, to=1000, by=40)
filled.contour(solRad, x=lon, y=lat, levels=levels, col=col)
# Write a shapefile
require(maptools); require(rgdal)
writeOGR(ContourLines2SLDF(contourLines(lon, lat, solRad, levels=levels)),
"solRad.shp", "contours", "ESRI Shapefile")
You can look at the filled.contour output ![here] http://www.jonxion.ch/5166482/solRad.png. We then transform the shape file to a topojson file, which you can find by replacing the .png in the above link by .json.
Finally, we render it with D3.js and Leaflet, leading to this faulty result [here] http://www.jonxion.ch/5166482/
We browsed many tutorials and other examples without finding the cue to our problem. What are we doing wrong?
Might it be a d3.js limitation? We recognise that our data is more complex, but Diegovalle's data contains unclosed contours too (see its upper left corner). Would writing ContourPolygones instead of ContourLines solve our problem? Does such routines exist? Or, is there an alternative technique to d3.js? Thank's in advance for your help!
I am wondering if this is possible to do this R .
I have one data as SpatialLinesDataFrame and another as spatialPolygonDataFrame. Is it possible to overlay these two data ?
When I try to overlay these I get the following error:
jd <- overlay(res,hello)
Error in function (classes, fdef, mtable) : unable to find an inherited method for function
‘overlay’ for signature ‘"SpatialLinesDataFrame", "SpatialPolygonsDataFrame"’
In the above code res is the SpatialLinesDataFrame and hello is SpatialPolygonDataFrame.
I have an shapefile and then I have data points with x,yand z
coordinates. I want to show the contour lines on the shapefile.
The procedure I used is using akima package to do the interpolation. The
code I used to interpolate is
fld <- interp(x,y,z)
Then I changed this to spatial object by using following code:
res <-ContourLines2SLDF(contourLines(fld))
The above command would store the contourlines as spatial data.
Then I read the shapefile and I plot both shapefile and res as follows:
p1 <-
spplot(hello,sp.layout=list(list("sp.lines",res)),col="blue",lwd=0,fill="grey",colorkey=F)
p1
"hello" is my shapefile and "res" is the object I created as shown above.
The problem is contour stored in "res" extends beyond the shapefile. So I
want to clip that contour with the shapefile and only display the contour
within the shapefile area.
So I am looking for a way to clip the contour layer with the polygon layer.
I have attached the image I got with my code.
In the image you can see the lines out of the shapefile. I also want to know
how can I display the contour levels on the map.
Thank you so much.
Jdbaba
I also want to know what does overlay does exactly. Does it intersect the area of both the data ?
Thank you.
It sounds like you're trying to clip your lines to the polygon extent. Use gIntersection from the rgeos package. Here's a reproducible example:
library(rgeos)
xx <- SpatialPoints(coords=matrix(data=c(0,0), nrow=1))
xx <- gBuffer(spgeom=xx, width=1)
yy <- SpatialLines(list(Lines(Line(matrix(c(-1,1,-1,1), nrow=2)), ID=1)))
zz <- gIntersection(yy, xx)
You can overlay the plot like so:
plot(xx)
plot(zz, add = TRUE, col = "blue")
Noah's answer has worked quite well for me. However, the output of his answer is a SpatialLines object, which cannot be saved as a shape file.
My two cents here is about how you can convert your SpatialLines object into a SpatialLinesDataFrame and save it as a shape file.
res.df <- fortify(res) # create data frame of res, your original SpatialLinesDataFrame
data <- data.frame(id = unique(res.df$id)) # get ids of road segments
rownames(data) <- data$id
# transform SpatialLines object into SpatialLinesDataFrame
zzSpatialLineDF <- SpatialLinesDataFrame(zz, data) # convert zz object keeping road ids
# 5 Save Shape File to your working directory
writeOGR(zzSpatialLineDF, dsn = '.', layer ='zzSpatialLineDF', driver = 'ESRI Shapefile')