rwGridsearch: Grid search optimization for random walk runout path

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

Grid search optimization for random walk runout path

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

rwGridsearch(
  dem,
  slide_plys,
  slide_src,
  slide_id = NULL,
  slp_v,
  ex_v,
  per_v,
  gpp_iter = 1000,
  buffer_ext = 500,
  buffer_source = NULL,
  save_res = FALSE,
  plot_eval = FALSE,
  saga_lib = NULL
)

Arguments

dem	A DEM as a RasterLayer object
slide_plys	Runout tracks as a SpatialPolygonsDataFrame
slide_src	SSource points as a SpatialPointsDataFrame or source areas as a SpatialPolygonsDataFrame
slide_id	Selects a single runout polygon from slide_plys by row
slp_v	A vector of random walk slope thresholds - below lateral spreading is modeled
ex_v	A vector or random walk exponents controlling lateral spread
per_v	A vector or random walk persistence factors 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
save_res	(logical), if TRUE, will save results in current working directory
plot_eval	If TRUE, will plot random walk path and runout polygon
saga_lib	The initiated SAGA-GIS geoprocessor object

Details

Runout is either simulated from a single source point or a buffered area round the source point.

Value

the area under the receiver operating characteristic

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