rwPerformance: Random walk runout path performance

In jngtz/runout.opt: GPP regional runout model optimization and validation in R

Random walk runout path performance

Description

Computes the area under the receiver operating characteristic curve (AUROC) for runout paths simuluated using the random walk model component of the GPP tool in SAGA-GIS. The AUROC compares a runout polygon to the simulated path.

Usage

rwPerformance(
  dem,
  slide_plys,
  slide_src,
  slide_id = 2,
  slp = 33,
  ex = 3,
  per = 2,
  gpp_iter = 1000,
  buffer_ext = 500,
  buffer_source = NULL,
  plot_eval = FALSE,
  saga_lib = NULL
)

Arguments

dem	A DEM as a RasterLayer object
slide_plys	Runout tracks as a SpatialPolygonsDataFrame
slide_src	Source points as a SpatialPointsDataFrame or source areas as a SpatialPolygonsDataFrame
slide_id	Selects a single runout polygon from slide_plys by row to run GPP model
slp	Random walk slope threshold - below lasteral spreading is modelled
ex	Random walk exponent controlling lateral spread
per	Random walk persistence factor to weight flow direction consistency
gpp_iter	Number of model iterations
buffer_ext	(Optional) Defines buffer distance (in meters) around runout polygon to crop source DEM. This helps to reduce computational time
buffer_source	(Optional) Can define a buffer distance (in meters) to extend source point to a source area
plot_eval	logical. If TRUE will plot random walk path and runout polygon
saga_lib	The initiated SAGA-GIS geoprocessor object

Details

Runout source can be either point or area.

Value

The area under the receiver operating characteristic

Examples

## Not run: 
# Initialize a saga object
saga <- Rsagacmd::saga_gis()

# Load elevation model (DEM)
dem <- raster::raster(system.file("extdata/elev_12_5m.tif", package="runout.opt"))

# Load runout polygons and source points
runout_plys <- rgdal::readOGR(system.file("extdata/dflow_runout_ply.shp", package="runout.opt"))
source_pnts <- rgdal::readOGR(system.file("extdata/dflow_source_pnt.shp", package="runout.opt"))

# Run GPP random walk model for a rounout polygon
rw <- rwPerformance(dem, slide_plys = runout_plys[1,], slide_src = source_pnts,
    slp = 30, ex = 3, per = 2,
    gpp_iter = 1000, buffer_ext = 500, buffer_source = 50,
    plot_eval = TRUE, saga_lib = saga)

rw # returns AUROC


## End(Not run)

jngtz/runout.opt documentation built on July 17, 2025, 3:06 a.m.