editor_options	markdown wrap 72

Cell2FireR: R bindings and platform for Fire Spread Simulations using Cell2Fire

Cell2FireR not only provides a bindings to Cell2Fire, but also provides a portal with tools to load and validate weather/land-cover data, manually insert ignition points and run fire simulation using Cell2Fire directly on the portal.

It also provides tools to extract aligned data formatted correctly for running FARSITE and ForeFire software locally.

Installation

Grab the latest version from GitHub then run the interactive app, or keep reading.

remotes::install_github("fpirotti/Cell2FireR")
library(Cell2FireR)
Cell2FireR::runWildfireApp()

Description

Installing this package provides two main elements:

• R bindings to the Cell2Fire software, a Cell Based Forest Fire Growth Model.
• an online graphical user interface for learning how to use the fire behaviour software.

The R bindings allow users to modify input arguments and run the simulation inserting arguments and pipelining call to the Cell2Fire executable and interpret the outputs converting them to maps and plots.

Cell2Fire R Wrapper (run_cell2fire.R)

This R script provides a high-level wrapper function, run_cell2fire(), designed to automate the preparation and execution of the Cell2Fire wildfire spread simulator. It handles directory management, coordinate transformation for ignition points, raster data validation, and command-line argument construction.

---

Overview

The run_cell2fire function acts as a bridge between R and the Cell2Fire C++ executable. It automates several tedious steps: * Workspace Management: Creates timestamped instance and result directories for every run. * Data Preparation: Links or copies required landscape rasters (Fuel, Elevation, Slope, etc.) into a standardized format for the simulator. * Auto-Conversion: Detects if fuel rasters are not in the required 32/64-bit format and converts them to INT4U automatically. * Spatial Alignment: Automatically transforms ignition point coordinates to match the Coordinate Reference System (CRS) of the fuel raster.

---

🛠 Parameters

Core Inputs

Parameter	Description
fuel	Required. Path to the fuel model raster.
fuel_model	Required. String specifying the model logic. It can be provide either by string or single character according to the lookup table below. "0. Scott & Burgan" = "S", "1. Kitral" = "K", "2. Canada FBP" = "C", "3. Portugal" = "P" also an heuristic check for key words is done to look for the fuel model to use.
out_folder	Base directory where simulation instances will be stored.
input_folder	Directory containing rasters in asc or tif format with the following possible rasters

If the 'fuels' raster file is in tif format, then it assumes all other rasters are also in that format. Otherwise, it assumes ascii format.

• elevation: terrain elevation [m]
• saz: slope azimuth
• slope: terrain slope
• cur: curing level
• cbd: canopy bulk density [kg/m3]
• cbh: canopy base height [m]
• ccf: canopy cover fraction
• probabilityMap: ignition probability map [%]
• fmc: foliage moisture content
• hm: tree height [m]

Landscape & Topography

• elevation, slope, saz (Azimuth), cur (Curvature): Paths to topographic rasters.
• firebreaks: Path to a raster defined where fire cannot spread.

Canopy Data (Crown Fire)

• crown: Boolean. If TRUE, the simulator includes crown fire logic.
• cbh, cbd, ccf, hm: Paths to canopy base height, bulk density, cover fraction, and height model rasters.

Weather & Ignition

• weather_mode: Supports "Single weather file" or "Multiple weathers".
• ignition_mode: Supports "Probability map" or "Single points on a Layer" (via CSV or Shapefile).

PLEASE TAKE NOTE

If you specify an --input-instance-folder, you do not need to provide a separate path for the topography, fuel and canopy rasters, but the files inside that folder must be named exactly as the following and be either in GeoTiff (.tif) or ESRI ASCII format (.asc):

• fuels.tif(.asc)
• elevation.tif(.asc)
• slope.tif(.asc)
• saz.tif(.asc)
• cur.tif(.asc)
• cbh.tif(.asc)
• cbd.tif(.asc)
• ccf.tif(.asc)
• hm.tif(.asc)
• fcc.tif(.asc) - forest canopy cover

---

Usage

Basic Example

source("run_cell2fire.R")

sim_results <- run_cell2fire(
    fuel = "data/fuels.tif",
    fuel_model = "0. Scott & Burgan",
    input_folder = "data",
    out_folder = "sim_runs", 
    elevation = "data/dem.tif",
    nsims = 10,
    sim_threads = 4
)

Return Object

The function returns a named list containing: * process: The active processx object (handling the background C++ execution). * command: The full path to the binary used. * args: The character vector of CLI arguments passed to Cell2Fire. * outputFolder: Path to the specific results directory for this run.

---

Simulation Logic Details

1. Fuel Data Validation: The script checks the data type of the fuel raster. If it is less than 32-bit, it warns the user and creates a compatible _INT4U version.
2. Lookup Tables: Based on the fuel_model selected, the script automatically picks the corresponding lookup table (e.g., spain_lookup_table.csv, fbp_lookup_table.csv) from the template directory.
3. Weather Handling: If no weather file is provided in "Single" mode, it copies a generic default Weather.csv to ensure the simulation doesn't fail.
4. Dry Run Mode: Setting dry = TRUE allows you to validate your inputs and see the exact command line string that would be executed without actually starting the simulation.

Usage

In order to run the simulator and process the results, the following command can be used:

• via Rscript call

Rscript -e "Cell2FireR::run_cell2fire()" --input-instance-folder ../data/Sub40x40/ --output-folder ../Sub40x40 --ignitions 1 --sim-years 1 --nsims 10 --grids --finalGrid --weather rows --nweathers 1 --Fire-Period-Length 1.0 --output-messages --ROS-CV 0.8 --seed 123 --stats --allPlots --IgnitionRad 1

• directly

library(Cell2FireR)

# Example wrapper or user call
template_location <- system.file("templates", package = "Cell2FireR")

# Choose binary based on OS automatically... or assign path of Cell2Fire to 
# bin_location variable
# 
# if (.Platform$OS.type == "windows") {
#   bin_location <- system.file("bin", "C2F", "Cell2Fire.exe", package = "Cell2FireR")
# } else {
#   bin_location <- system.file("bin", "C2F", "Cell2Fire", package = "Cell2FireR")
# }

bin_location <- "path to your cell2fire executable downloaded from https://github.com/fire2a/C2F-W/releases"
# Run the simulator
process_obj <- run_cell2fire(
  fuel = "path/to/fuel.tif",
  fuel_model = "0. Scott & Burgan",
  input_folder = "path/to/inputs",
  out_folder = "path/to/outputs",
  elevation = "path/to/elevation.tif",
  c2f_bin_path = bin_location,
  template_dir = template_location
)

# You can now monitor `process_obj` using processx methods like:
process_obj$process$wait()
process_obj$process$read_output_lines()

For the full list of arguments and their explanation use:

run_cell2fire()

Web app

An embedded web app is available in the folder inst/app. It allows interactive insertion of ignition points, weather data etc.. and other input arguments. Users can install in their own PC or in HPC servers with multiple processors thus taking advantage from parallel processing capabilities of Cell2Fire.

shiny::runApp('inst/app')

Uploading user data can be done with a files inside a Zipped archive. The name of the ZIP archive will be the one used for storing and viewing the specific dataset. Fuel and canopy files should be stored directly in the archive, without subfolders.

See a typical dataset to upload using the FIRE-RES Pan-European Fuel Map Server platform for 100 m resolution maps or try one from the direct link HERE

Acknowledgements

This work was supported by the Wildfire CE project "Fighting wildfires better together across borders" Interreg Central Europe.

==> Many parts were taken from the excellent development and implementation of the QGIS plugin from the Fire Management & Advanced Analytics group.

References

"Cell2Fire: A Cell-Based Forest Fire Growth Model to Support Strategic Landscape Management Planning"

by Cristóbal Pais, Jaime Carrasco, David L. Martell, Andres Weintraub, and David L. Woodruff (published in Frontiers in Forests and Global Change, 2021).