The terra package

• The terra package
• Classes
  • SpatRaster
  • SpatVector
  • SpatExtent
• Creating SpatRaster objects
• Raster algebra
• High-level methods
  • Modifying a SpatRaster object
  • lapp
  • app
  • classify
  • Focal
  • Distance
  • Spatial configuration
  • Predictions
  • Vector to raster conversion
  • Summarize
• Plotting
• Writing files
  • File format
• Cell-level functions
  • Introduction
  • Accessing cell values
• Spatial prediction
  • Predict
    • GLM
    • Principal components
    • Random Forest
  • cforest
  • Interpolate
    • Thin plate spline interpolation with x and y only
    • inverse distance weighted (IDW)
    • Kriging
• Parallelization
  • Introduction
  • The constraint: terra objects hold a C++ pointer
  • Strategy 1 — wrap() and unwrap()
  • Strategy 2 — pass filenames, open them on the worker
  • Strategy 3 — partition by tiles
    • Picking tiles automatically
    • Edge effects: buffer = for focal-style operations
    • Memory safety
    • Writing the final result
    • Doing it by hand
  • Strategy 4 — exchange raw values
  • Function-level parallelism: the cores argument
    • built-in functions
    • Important caveats:
  • Within-process parallelism: TBB
  • Single computer vs. HPC cluster
    • One computer, multiple cores
    • HPC cluster of nodes (each with multiple cores)
• Miscellaneous
  • Session options