Issues related to `crs` and `default_crs` processing after ggplot2 4.0 Release

[Breeze-Hu]

Hi!

I'm in the process of using tidyterra for raster file plotting, and I've noticed a problem in the development version of ggplot2 that I didn't have before (the code that went there first is not working). The details are as follows:

example data

library(terra)
#> terra 1.8.54
library(ggplot2)
library(tidyterra)
#> 
#> Attaching package: 'tidyterra'
#> The following object is masked from 'package:stats':
#> 
#>     filter

data <- rast('./temdata.nc') 

defcrs <- paste(
  "+proj=lcc", 
  "+lat_1=25", 
  "+lat_2=47", 
  "+lon_0=105", 
  "+datum=WGS84", 
  "+units=m",
  sep = " "
)

ggplot()+
  geom_spatraster(data = data) +
  coord_sf(crs = defcrs) # This works.

# When I needed to adjust the range so that I used the following method. This reports an error:
ggplot()+
  geom_spatraster(data = data) +
  coord_sf(
    crs = defcrs,
    default_crs = sf::st_crs(4326),
    xlim = c(90, 122),
    ylim = c(24, 37),
    label_axes = list(bottom = "E", left = "N", right = "N")
  ) 
#> `geom_raster()` only works with linear coordinate systems, not `coord_sf()`.
#> ℹ Falling back to drawing as `geom_rect()`.
#> Error in `geom_spatraster()`:
#> ! Problem while converting geom to grob.
#> ℹ Error occurred in the 1st layer.
#> Caused by error:
#> ! vector memory limit of 18.0 Gb reached, see mem.maxVSize()

^{Created on 2025-06-26 with reprex v2.1.1}

ggplot2 tidyverse/ggplot2#6382 made some changes to the geom_raster function that I suspect are related.

> sessionInfo()
R version 4.5.1 (2025-06-13)
Platform: aarch64-apple-darwin20
Running under: macOS Sequoia 15.5

Matrix products: default
BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib 
LAPACK: /Library/Frameworks/R.framework/Versions/4.5-arm64/Resources/lib/libRlapack.dylib;  LAPACK version 3.12.1

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

time zone: Asia/Shanghai
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] tidyterra_0.7.2    ggplot2_3.5.2.9000 terra_1.8-54      

loaded via a namespace (and not attached):
 [1] tidyr_1.3.1        RColorBrewer_1.1-3 R6_2.6.1           codetools_0.2-20  
 [5] tidyselect_1.2.1   farver_2.1.2       magrittr_2.0.3     gtable_0.3.6      
 [9] glue_1.8.0         tibble_3.3.0       dichromat_2.0-0.1  pkgconfig_2.0.3   
[13] generics_0.1.4     dplyr_1.1.4        lifecycle_1.0.4    cli_3.6.5         
[17] scales_1.4.0       vctrs_0.6.5        grid_4.5.1         withr_3.0.2       
[21] compiler_4.5.1     purrr_1.0.4        rstudioapi_0.17.1  tools_4.5.1       
[25] pillar_1.10.2      Rcpp_1.0.14        rlang_1.1.6       

---

After checking, this version of ggplot at tidyverse/ggplot2@3914e13 is able to draw images properly, although there are warning messages (lcc: Invalid latitude; lcc: Invalid latitude...)

Best wishes,
Hu

[Breeze-Hu]

In exploring alternatives for solving the problem, I noticed that tidyterra differed from several other solutions (based on ggsaptial and ggplot geom_rect) in the final plotted results, mainly in the raster at the edge locations.

as follows:

library(terra)
library(ggplot2)
library(tidyterra)
library(ggspatial)
library(patchwork)

data <- rast('~/Desktop/R_learn/temdata.nc') 

defcrs <- paste(
  "+proj=lcc", 
  "+lat_1=25", 
  "+lat_2=47", 
  "+lon_0=105", 
  "+datum=WGS84", 
  "+units=m",
  sep = " "
)


p1 <- ggplot()+
  geom_spatraster(data = data, show.legend = FALSE) +
  coord_sf(crs = defcrs, 
           default_crs = sf::st_crs(4326), 
           xlim = c(90, 122), 
           ylim = c(24, 37),
           label_axes = list(bottom = "E", left = "N", right = 'N', top = 'E')
  )+
  labs(title = 'tidyterra plot')+
  theme_bw()

p2 <- ggplot()+
  geom_spatial_tile(crs = 4326, data = data |> as.data.frame(xy = TRUE), 
                    aes(x = x, y = y, fill = temdata,colour = after_scale(fill), 
                        linewidth = I(0.3)), show.legend = FALSE) +
  coord_sf(
    crs = defcrs, 
    default_crs = sf::st_crs(4326), 
    xlim = c(90, 122),
    ylim = c(24, 37),
    label_axes = list(bottom = "E", left = "N", right = 'N', top = 'E')
  ) +
  labs(title = 'ggsaptial plot')+
  theme_bw()


p3 <- ggplot()+
  geom_rect(data = data |> as.data.frame(xy = TRUE),
            aes(x = x, y =y, fill = temdata,colour = after_scale(fill), linewidth = I(0.3)),
            width = 0.25, height = 0.25, show.legend = FALSE) + 
  coord_sf(
    crs = defcrs, 
    default_crs = sf::st_crs(4326), 
    xlim = c(90, 122), 
    ylim = c(24, 37),
    label_axes = list(bottom = "E", left = "N", right = 'N', top = 'E')
  ) +
  labs(title = 'ggplot_base plot')+
  theme_bw()

p1+ p2+p3

---

As a comparison, terra plots this:

data |> project(defcrs) |> plot()

[Breeze-Hu]

Is it due to spatSample in geom_spatraster ? (I'm not really good at coding, so this conclusion may be incorrect, please bear with me)

# Try to keep the same number of cells on the template
  template <- terra::spatSample(template, terra::ncell(init_rast),
    as.raster = TRUE,
    method = "regular"
  )