Potential spatialize_lsm (direction = 4) bug about landscapemetrics

[mhesselbarth]

Hey,

Thanks for finding this. I think the argument does not get passed on properly. I will try to fix this

[mhesselbarth]

Okay, should be fixed in c5336d1

library(terra)
library(landscapemetrics)

#Create a 10 by 10 raster in which all cells are 1
r = rast(ncol=10, nrow=10)
values(r) = 1

#Randomly assign some of the values to NA
set.seed(123)  # For reproducibility
na_matrix = matrix(runif(100) < 0.8, nrow=10, ncol=10)

# Apply NA values to the raster based on the NA matrix
values(r)[na_matrix] = NA

#Assign crs
crs(r) = "EPSG:32633"

# show landscape patch ids
show_patches(landscape = r, class = 1, directions = 4, labels = TRUE)
#> $layer_1

# calculate the area
area_df <- lsm_p_area(landscape = r, directions = 4)

# get raster layer with values
area_ras <- spatialize_lsm(r, what = "lsm_p_area", directions = 4)

# those should be identical and four unique values
unique(area_df$value)
#> [1] 0.2592 0.1296 0.0648 0.1944
unique(area_ras$layer_1$lsm_p_area[])
#>       value
#> [1,]     NA
#> [2,] 0.1296
#> [3,] 0.0648
#> [4,] 0.2592
#> [5,] 0.1944

# this seems to be correct now
plot(area_ras$layer_1$lsm_p_area, col = viridis::viridis(n = 4, option = "E"))

# compare to function by the package
show_lsm(landscape = r, what = "lsm_p_area", class = 1, directions = 4, labels = TRUE)
#> $layer_1

^{Created on 2024-02-29 with reprex v2.1.0}

Could you please try to install the Github version and see if results are correct now.

[MStuhlmacher]

Great, thank you! It's on my list to try the updated version and report back

[MStuhlmacher]

I installed the Github version and re-ran my code from above and it now works as expected. Thanks for solving that so quickly!

[mhesselbarth]

Cool 👍 Will close this, please re-open if needed