Make raster stack with different extent

Question

I am in trouble making raster stack which have slightly different extent. The answer (1st one) given here is useful but did not help in my case. For example, I want to make a raster stack using bio2 raster for Australia and this Australian raster. The second raster comes for Australia only and the first one is global. So I cropped the global bio2 raster to the same extent of Australian raster using crop() function, but the resultant raster extent (i.e., bio2.au) is slightly different (therefore, I cannot make raster using the cropped raster and the Australian raster, awc). Sample code is below:

library(raster)
awc <- raster("path to Australian raster")
bio2.g <- raster("path to Bio2 global raster")
# crop bio2.g to the same extent of awc
bio2.au <- crop(bio2.g, extent(awc))

# make a raster stack
st <- stack(awc, bio2.au)
Error in compareRaster(x) : different extent

I have also tried using quick=TRUE within the stack() function. But in this case the cell values in awc is lost. Note: the size of awc raster is 4gb.

# first make a list of rasters saved in the computer
li <- list.files("path to file", pattern = ".tif$", full.names = TRUE)
st <- stack(li, quick=TRUE)
st[[1]] # no cell values for awc

Your suggestions will be highly appreciated. My ultimate goal is to crop several bioclim rasters to the same extent of Australian raster awc and stack them together so that raster cell values are not lost.

Edit (after comment of @Cobin):

Below is the attribute of each raster

# global raster (bigger raster)
> r
class       : RasterLayer 
dimensions  : 21600, 43200, 933120000  (nrow, ncol, ncell)
resolution  : 0.008333333, 0.008333333  (x, y)
extent      : -180, 180, -90, 90  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0 
data source : D:\Worldclim2_Bioclim\wc2.0_bio_30s_02.tif 
names       : wc2.0_bio_30s_02 
values      : 0, 37.06667  (min, max)


# Australian raster (smaller raster)
> r1
class       : RasterLayer 
dimensions  : 43201, 49359, 2132358159  (nrow, ncol, ncell)
resolution  : 0.0008333333, 0.0008333333  (x, y)
extent      : 112.8921, 154.0246, -44.00042, -7.999583  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0 
data source : D:\SoilAWC5cm.EV1.tif 
names       : SoilAWC5cm.EV1 
values      : 2.997789, 27.86114  (min, max)

# new raster, after crop() function is applied
> r2 <- crop(r,extent(r1))
> r2
class       : RasterLayer 
dimensions  : 4320, 4936, 21323520  (nrow, ncol, ncell)
resolution  : 0.008333333, 0.008333333  (x, y)
extent      : 112.8917, 154.025, -44, -8  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0 
data source : C:\Users\Anwar\AppData\Local\Temp\Rtmpmg9fyF\raster\r_tmp_2018-11-23_164300_11308_65747.grd 
names       : wc2.0_bio_30s_02 
values      : 1.933333, 18.15833  (min, max)

# rebuild r2 to match r1
> r22 <- raster(vals=values(r2),ext=extent(r1), nrows=dim(r1)[1],ncols=dim(r1)[2])

Error in setValues(r, vals) : 
  length(values) is not equal to ncell(x), or to 1

Dandiprat

global raster (bigger raster) resolution is different from Australian raster (smaller raster). If you want to stack r1 and r2, you should resample the raster insuring they have same resolution, extent, crs. Such as function resample(r1,r2, method='near') – Presocratic

Thank you @Cobin. This trick worked successfully. – Dandiprat

Answer 1

I suppose that the extent of two raster are differet though the raster masked by crop function.You should check the both of awc and bio.au extent base on same reolution, rows and columns. Because I couldn't download data from hyperlink, I give an example of my own data.

r <- raster('/big_raster')
r1 <- raster('/small_raster')
r2 <- crop(r,extent(r1))

r1
class       : RasterLayer 
dimensions  : 74, 157, 11618  (nrow, ncol, ncell)
resolution  : 0.0833333, 0.0833333  (x, y)
extent      : 89.2185, 102.3018, 30.96238, 37.12905  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0 
data source : D:\D\temp\Rtest\modis8km.tif 
names       : modis8km 
values      : -32768, 32767  (min, max)

r2
class       : RasterLayer 
dimensions  : 74, 157, 11618  (nrow, ncol, ncell)
resolution  : 0.08333333, 0.08333333  (x, y)
extent      : 89.25, 102.3333, 31, 37.16667  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0 
data source : in memory
names       : g201401a 
values      : -32768, 7789  (min, max)

Though r1 and r1 with same resolution and dimension, the extent have tiny offset. It cause stack error.

 stack(r1,r2)
 Error in compareRaster(x) : different extent

So, you should rebuid the r2 to match r1:

r22 <- raster(vals=values(r2),ext=extent(r1),crs=crs(r1),
                  nrows=dim(r1)[1],ncols=dim(r1)[2])

Now stack(r22,r1) will be successful.

Presocratic

Thank you. My r1 extent is: 112.8921, 154.0246, -44.00042, -7.999583, and my r2 extent after cropping is: 112.8917, 154.025, -44, -8. But I have got the following error: r22 <- raster(vals=values(r2),ext=extent(r1), nrows=dim(r1)[1], ncols=dim(r1)[2]). Error in setValues(r, vals) : length(values) is not equal to ncell(x), or to 1 @Presocratic – Dandiprat

The main problem is that the small rasster is not snapping big raster.The rasters have different rows and collumns (width and height) after crop.You should show bio2.au and awc atrribute in question. Maybe cellFrom function to crop raster is preciser. @Crayfish – Presocratic

Thank you. I have now added raster attributes in my main question. @Presocratic – Dandiprat