Least-Cost Pedestrian Evacuation Modeling.

evacpath

evacpath is an R package for road-constrained pedestrian evacuation modeling using least-cost path analysis. It takes generic hazard, road/pathway, and elevation inputs, then returns distance-to-safety and evacuation-time outputs for any projected study area.

The package is designed around a simple idea:

hazard zone + roads/pathways + DEM + projected CRS -> distance-to-safety and evacuation-time outputs

The workflow builds on open-source least-cost path methods for evacuation planning (Cordero et al. 2025) and uses leastcostpath for least-cost path and movement-potential modeling (Lewis 2023; Lewis 2021).

Current status

This is an initial package scaffold, not a polished CRAN release. It is ready for local development with devtools::load_all() and devtools::document().

Core workflow

library(evacpath)

result <- run_evacpath(
  hazard_zone = "path/to/hazard_zone.tif",
  roads = "path/to/roads.gpkg",
  dem = "path/to/dem.tif",
  target_crs = "EPSG:XXXX",
  region_name = "Study area",
  road_buffer_m = 2,
  escape_buffer_m = 5,
  final_road_buffer_m = 3,
  seed = 23401,
  walking_speed_mps = 1.22,
  clip_mode = "hazard",
  progress = TRUE
)

result$evac_polygons
result$distance_points

write_evac_outputs(
  result,
  output_dir = "outputs",
  prefix = "study_area"
)

Main package functions

Customizable assumptions

Most modeling assumptions are explicit parameters:

run_evacpath(
  ...,
  target_crs = "EPSG:XXXX",
  grid_resolution = NULL,
  grid_resolution_factor = 5,
  road_buffer_m = 2,
  escape_buffer_m = 5,
  final_road_buffer_m = 3,
  dem_resolution = NULL,
  walking_speed_mps = 1.22,
  clip_mode = "hazard"
)

Tsunami preprocessing

Some coastal workflows need two tsunami zones. The land-only inundation zone is used for origins and mapping. A second zone combines the land inundation footprint with water so that the coastline is not treated as an artificial escape boundary.

library(terra)
library(evacpath)

tsunami <- rast("path/to/inundation.nc")
dem <- rast("path/to/topography.nc")
roads <- vect("path/to/roads.shp")

zones <- prepare_tsunami_zones(
  inundation = tsunami,
  dem = dem,
  target_crs = "EPSG:XXXX",
  inundation_threshold = 0,
  dem_sign_multiplier = 1
)

roads <- clean_roads(
  roads,
  exclude = list(field = "road_type", values = "pier"),
  target_crs = "EPSG:XXXX"
)

result <- run_evacpath(
  hazard_zone = zones$hazard_zone,  # land-only TEZ for origins/output
  escape_zone = zones$escape_zone,  # TEZ + water for true inland escape boundary
  roads = roads,
  dem = zones$dem,
  target_crs = "EPSG:XXXX",
  region_name = "Study area"
)

Do not use the land-only hazard_zone to find escape points in tsunami workflows unless you intentionally want the coastline to act as a boundary. In most tsunami cases, use escape_zone = zones$escape_zone.

Local development

From the parent directory of the package:

install.packages(c("devtools", "terra", "remotes"))
remotes::install_github("josephlewis/leastcostpath") # if leastcostpath is not on CRAN for your setup

devtools::load_all("evacpath")
devtools::document("evacpath")
devtools::check("evacpath")

Notes

• Use a projected CRS in meters before distance modeling.
• Keep case-specific preprocessing outside the core package when possible.
• Do not assume every DEM needs to be multiplied by -1; make that decision per dataset.
• Do not assume every road layer has the same fields; use clean_roads() with region-specific settings.
• Use max_origins for exploratory runs and remove or increase it for final production runs.

Time-grid output

By default, run_evacpath() now clips result$evac_polygons / result$time_grid to the full hazard zone, not just the buffered road network. Movement is still calculated using the road-constrained conductance surface, but the mapped time grid is a Voronoi-style surface over the hazard/inundation zone. Use clip_mode = "road_hazard" only when you want the older road-buffer-limited output.

Example data and figures

Small example inputs are included in inst/extdata/:

inst/extdata/dem.tif
inst/extdata/rds.gpkg
inst/extdata/tsunami_inundation_depth.tif

These files are from the Jakarta example dataset and are used only to demonstrate how to load package data:

dem <- terra::rast(system.file("extdata/dem.tif", package = "evacpath"))
roads <- terra::vect(system.file("extdata/rds.gpkg", package = "evacpath"))
inundation <- terra::rast(system.file("extdata/tsunami_inundation_depth.tif", package = "evacpath"))

The detailed diagnostic example is available at vignettes/diagnostic-example.Rmd and inst/examples/diagnostic-example.Rmd. It walks through each major function separately, including tsunami-specific zone preparation, inset cropping of roads used for escape-point detection, road-aware escape-boundary generation, least-cost-path testing, and final time-grid mapping.

The README figures are generated with terra plotting code in inst/scripts/make-readme-figures.R. The figures use a focused projected window (x = 669000 to 675000, y = 9223000 to 9225000) so the examples stay readable.

For tsunami workflows, the preferred escape-point pattern is:

zones <- prepare_tsunami_zones(inundation, dem, target_crs = target_crs)

roads_for_escape <- crop_roads_to_inner_extent(
  roads = roads,
  zone = zones$escape_zone,
  inset_x_m = 250,
  inset_y_m = 250
)

escape_boundary_zone <- make_road_aware_escape_zone(
  escape_zone = zones$escape_zone,
  roads = roads_for_escape,
  road_buffer_m = 2,
  crop_buffer_m = 3
)

escape_points <- find_escape_points(
  hazard_zone = escape_boundary_zone,
  roads = roads_for_escape
)

Use ?evacpath after running devtools::document() or installing the package to view the package-level help page.

References

Cordero, E., Ruiz Vélez, R., Huérfano Moreno, V., Sherman, C., 2025. Enhancing tsunami evacuation strategies in Puerto Rico using open-source least-cost path analysis. J. Disaster Sci. Manag. 1, 18. https://doi.org/10.1007/s44367-025-00018-y

Lewis, J., 2023. leastcostpath: Modelling Pathways and Movement Potential Within a Landscape.

Lewis, J., 2021. Probabilistic Modelling for Incorporating Uncertainty in Least Cost Path Results: a Postdictive Roman Road Case Study. Journal of Archaeological Method and Theory 28, 911-924. https://doi.org/10.1007/s10816-021-09522-w.