How to assign and match land cover type from Corine to a dataframe with a set of lon lat coordinates-CodePudding

I'm trying to find out the land use type for a set of coordinates that define the location of plant species across Europe. However I'm stuck in the process of assigning the land use to the respective coordinates. Any advice would be more than welcome!

First, I download the land use raster file from here: https://land.copernicus.eu/pan-european/corine-land-cover

#Read raster file (year 2006 but could be any)
clc <- raster("U2006_CLC2000_V2020_20u1.tif")

Then, I read the Corine land use classes and rename the levels of the raster file with this classes

#Read Corine classes
clc_classes <- foreign::read.dbf("CLC_1990/DATA/U2006_CLC2000_V2020_20u1.tif.vat.dbf",
                                 as.is = TRUE) %>%dplyr::select(value = Value,landcov = LABEL3)

This is a small subset of coordinates from my full list of coordinates (over 200.000 in total):

lon <- c("51.105", "51.195", "51.188", "51.239")
lat <- c("4.392", "4.395", "4.896", "4.468")
sp <- c("sp1","sp2", "sp3","sp4")
#Create minimal dataframe 
d <- data.frame(lon,lat,sp)

But now I really do not know how to proceed and create the final dataframe with land use type given the matching with the raster file

My intention is to add a 4th column as follows after the matching of my coordinates with the land use type of the raster file.

#Example of how this fourth column would be like:
d$land_use <- c("Olive groves", "Olive groves", "Vineyards", "Pastures")

CodePudding user response：

The data (another file from the same website).

library(terra)
r <- rast("U2018_CLC2018_V2020_20u1.tif")

As you can see, r knows about the class labels.

r
#class       : SpatRaster 
#dimensions  : 46000, 65000, 1  (nrow, ncol, nlyr)
#resolution  : 100, 100  (x, y)
#extent      : 9e 05, 7400000, 9e 05, 5500000  (xmin, xmax, ymin, ymax)
#coord. ref. : ETRS_1989_LAEA (EPSG:3035) 
#source      : U2018_CLC2018_V2020_20u1.tif 
#color table : 1 
#categories  : LABEL3, Red, Green, Blue, CODE_18 
#name        :                  LABEL3 
#min value   : Continuous urban fabric 
#max value   :                  NODATA 

head(cats(r)[[1]])
#  Value                                     LABEL3       Red     Green      Blue CODE_18
#1     1                    Continuous urban fabric 0.9019608 0.0000000 0.3019608     111
#2     2                 Discontinuous urban fabric 1.0000000 0.0000000 0.0000000     112
#3     3             Industrial or commercial units 0.8000000 0.3019608 0.9490196     121
#4     4 Road and rail networks and associated land 0.8000000 0.0000000 0.0000000     122
#5     5                                 Port areas 0.9019608 0.8000000 0.8000000     123
#6     6                                   Airports 0.9019608 0.8000000 0.9019608     124

Here are some example points for use with extract

pts <- matrix(c(3819069, 3777007, 3775822, 2267450, 2302403, 2331431), ncol=2)
extract(r, pts)    
#                        LABEL3
#1                Sea and ocean
#2 Complex cultivation patterns
#3           Natural grasslands

Or with your lon/lat points (you had the names reversed!), first transform these to the coordinate reference system of the land use raster:

lat <- c(51.105, 51.195, 51.188, 51.239)
lon <- c(4.392, 4.395, 4.896, 4.468)
xy <- cbind(lon, lat) 
v <- vect(xy, crs=" proj=longlat")
vv <- project(v, crs(r)) 

extract(r, vv)
#  ID                                     LABEL3
#1  1               Complex cultivation patterns
#2  2 Road and rail networks and associated land
#3  3                                   Pastures
#4  4             Industrial or commercial units

If you want the land use code instead

activeCat(r) <- "CODE_18"
extract(r, vv)
#  ID CODE_18
#1  1     242
#2  2     122
#3  3     231
#4  4     121