Extract raster data for a hex grid using exactextractr

This is the PRIMARY function for extracting raster data to hexagonal grids. It automatically handles CRS transformations and creates consistent hexagonal grids.

Usage

extract_raster_data(
  raster_files,
  study_area = NULL,
  cell_size = NULL,
  hex_grid = NULL,
  sample_fraction = 1,
  random_seed = 42
)

Arguments

raster_files

named vector of raster file paths OR named list of terra::rast objects

study_area

sf polygon (optional, for cropping and grid CRS reference)

cell_size

hex cell size (in units of the target CRS - see details)

hex_grid

optional sf hexagonal grid to use instead of creating a new one

sample_fraction

fraction of raster cells to use (default 1)

random_seed

for reproducibility

Value

list: data frame with cell_id, x, y, and values for each raster

Details

**CRS Handling:** - If `study_area` is provided, the function creates a grid in the study area's CRS - If no `study_area` is provided, the function creates a grid in the raster's CRS - The function automatically crops rasters to the study area (if provided) - All coordinate transformations are handled transparently

**Grid Creation:** - Creates hexagonal grids using `sf::st_make_grid` for robustness - Automatically adjusts cell sizes for geographic coordinates (degrees) - Can use a pre-existing grid via the `hex_grid` parameter

**Input Flexibility:** - Accepts both file paths (character strings) and `terra::rast` objects - When passing `terra::rast` objects, you can inspect CRS, dimensions, and data before extraction - Useful for working with in-memory rasters or pre-processed data

**CRS Handling Details:**

The function intelligently handles coordinate reference systems:

1. **Study Area Provided**: Grid is created in the study area's CRS - Example: If study_area is in UTM (EPSG:32630), grid uses UTM coordinates - Cell sizes should be in meters (e.g., 20000 for 20km hexagons)

2. **No Study Area**: Grid is created in the raster's CRS - Example: If raster is in WGS84 (EPSG:4326), grid uses geographic coordinates - Cell sizes should be in degrees (e.g., 0.1 for ~11km hexagons)

3. **Automatic Adjustments**: - Large cell sizes (>1) in geographic coordinates trigger warnings - Cell sizes are automatically adjusted for geographic coordinates if needed

**Input Types:**

The function accepts two types of input for `raster_files`:

1. **File Paths**: Named character vector of file paths “`r raster_files = c(ndvi = "path/to/ndvi.tif", elevation = "path/to/elevation.tif") “`

2. **Terra Raster Objects**: Named list of terra::rast objects “`r # Load rasters first to inspect them ndvi_rast <- rast("path/to/ndvi.tif") elev_rast <- rast("path/to/elevation.tif")

# Check CRS, dimensions, etc. print(crs(ndvi_rast)) print(dim(ndvi_rast))

# Pass to function raster_files = list(ndvi = ndvi_rast, elevation = elev_rast) “`

**Workflow Recommendation:**

For consistent results, always provide a `study_area` in your desired CRS: “`r # Create study area in UTM for accurate measurements study_area_utm <- st_transform(study_area_wgs, utm_crs)

# Extract with 20km hexagons (in UTM coordinates) extracted <- extract_raster_data( raster_files = c(var1 = "path/to/raster.tif"), study_area = study_area_utm, cell_size = 20000 # 20km in meters )

# Then apply smoothing smoothed <- smooth_variables(...) “`