geom_stars fails to plot using curvilineargrid

Question

I am trying to plot a heat map of gridded data on a curvilinear grid. From what I have read the stars package provides this functionality using geom_stars on a stars object created using st_as_stars with a curvilinear option. There is a nice blog post on it here. However, when I try to use this method it fails. Simple reprex below:

 library(stars)   
 library(ggplot2)

 lng <- rbind(c(-2.099361, -2.098536, -2.097710, -2.096885),
 c(-2.099350, -2.098525, -2.097699, -2.096874), c(-2.099339, -2.098514,
 -2.097688, -2.096863), c(-2.099328, -2.098503, -2.097677, -2.096852))
 
 lat <- rbind(c(57.10963, 57.10963, 57.10962, 57.10962), c(57.11008,
 57.11008, 57.11007, 57.11007),c(57.11053, 57.11053, 57.11052, 57.11051), c(57.11098, 57.11098, 57.11097, 57.11096))
 
 arr <- rbind(1:4, 5:8, 9:12, 13:16)
 
 arr_star <- st_as_stars(arr, curvilinear = list(west_east = lng,
 south_north = lat))
 
 ggplot() + geom_stars(data = arr_star, alpha = 0.75)

The error message is:

Error:
! Tibble columns must have compatible sizes.
• Size 2: Column `curvilinear`.
• Size 16: Columns `X1`, `X2`, and `X`.
ℹ Only values of size one are recycled.
Run `rlang::last_error()` to see where the error occurred.

Looking at the code for geom_stars, it seems to be trying to use dplyr::as_tibble on arr_star and it makes sense that this would fail. Any idea where I'm going wrong?

Answer 1

Also don't think that this is desired behavior. Here's a working approach:

1. Transform your data matrix to a stars object using st_as_stars().

   arr <- st_as_stars(
     rbind(1:4, 5:8, 9:12, 13:16)
     )
   

2. Add coordinates to arr as you've done in your example code:

   arr <- st_as_stars(
     arr, curvilinear = list(X1 = lng, X2 = lat)
   )
   

   Now, arr has a curvilinear grid with WGS 84 reference system:

   print(arr)
   

   stars object with 2 dimensions and 1 attribute
   attribute(s):
       Min. 1st Qu. Median Mean 3rd Qu. Max.
   A1     1    4.75    8.5  8.5   12.25   16
   dimension(s):
      from to refsys point                      values x/y
   X1    1  4 WGS 84 FALSE [4x4] -2.09936,...,-2.09685 [x]
   X2    1  4 WGS 84 FALSE    [4x4] 57.1096,...,57.111 [y]
   curvilinear grid
   

3. Plot using ggplot.

   ggplot() + 
     geom_stars(data = arr, alpha = 0.75)
   

Result