Update XX_sample_50_km_buffer_and_make_grids.R

Code/XX_sample_50_km_buffer_and_make_grids.R

@@ -3,15 +3,74 @@
 
 # packages
 library(sf)
+library(terra)
 library(tidyverse)
-library(ggplot2)
+# library(ggplot2)
+
+
+# arguments for the script
+taxa = "Aves" # taxa of interest (birds)
+yr = 2010 # year of interest
+grid_size = 5000 # unit: meters
+gcs = 4326 # geographic coordinate system
+pcs = 32633 # projected coordinate system (same as the one used for 50km buffers: XX_create geojson for 5 50km buffers for potential study sites.R)
 
 
 # read in GBIF data
 # only for one buffer for now
 gbif_dat <- read_delim("Data/GBIF data for test cases/buffer_1_data/occurrence.txt")
 
 # read in buffer geojson
-buffer_spatial <- st_read("Data/GBIF data for test cases/buffer_1.geojson")
+buffer_spatial <- st_read("Data/GBIF data for test cases/buffer_1.geojson") %>% 
+  st_transform(crs = pcs)
+
+# make a grid of the buffer with the specified grid size
+circ_grid <- buffer_spatial %>%
+  st_make_grid(cellsize = grid_size, what = "polygons") %>% # Create a grid of the buffer
+  st_sf() %>% # Convert to a spatial object
+  st_filter(buffer_spatial, .predicate = st_within) # Filter the grid to only the cells within the buffer
+
+
+# convert the GBIF data to a spatial object
+gbif_dat_sf <- gbif_dat %>% 
+  filter(class %in% taxa) %>% # subset to only the taxa of interest
+  filter(year %in% yr) %>% # subset to only the year of interest
+  st_as_sf(coords = c("decimalLongitude", "decimalLatitude"), crs = gcs) %>% # Convert to a spatial object
+  st_transform(crs = pcs) %>% # Transform to the same coordinate system as the buffer
+  st_filter(circ_grid, .predicate = st_within) %>% # Filter the data to only the points within the grid
+  select(species, geometry) # Select only the species and geometry columns
+
+# # plot the grid and GBIF data
+# plot(st_geometry(buffer_spatial))
+# plot(st_geometry(circ_grid), add = TRUE)
+# plot(st_geometry(gbif_dat_sf), add = TRUE)
+
+
+# create a raster of the grid
+circ_vect <- vect(circ_grid) # convert to a SpatVector
+r <- rast(ext(circ_vect), resolution = grid_size) # create an empty SpatRaster with the desired resolution
+ras <- rasterize(circ_vect, r) # rasterize the grid
+ras[!is.nan(values(ras))] <- 0 # set all cells to 0
+# plot(ras)
+
+# Create a raster of species counts
+uni_spp <- unique(gbif_dat_sf$species) # get the unique species
+ras.emp <- ras # empty raster to add the species rasters to
+for(i in 1:length(uni_spp)){ # loop through each species
+  rasx <- ras.emp # copy the empty raster
+  sp <- uni_spp[i] # get the species
+
+  sp_dat <- gbif_dat_sf %>% 
+    filter(species %in% sp) %>% # subset to only the species of interest
+    vect() # convert to a SpatVector
+
+  e <- terra::extract(ras, unique(sp_dat), cells = TRUE) %>% # extract the cells
+    count(cell) # count the number of cells
+
+  rasx[e$cell] <- 1 # detection = 1; nondetection = 0
+  ras <- ras + rasx # add the species raster to the main raster
+}
+# plot(ras) # Check the raster
 
-plot(buffer_spatial)
+# Save the raster
+writeRaster(ras, "Data/GBIF data for test cases/buffer_1_data/spp_count.tif", format = "geotiff")