In joeroe/fieldwalkr: Spatial Sampling and Survey Simulation

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

First, let's load fieldwalkr and generate some some dummy data to simulate a survey.

library("fieldwalkr")
library("sf")
library("ggplot2")

frame <- rpolygon()                   # Sample frame
sites <- st_sample(frame, 100)        # Target points
units <- quadrats(frame, size = 200)  # Survey units

# Plot
theme_nocoords <- function() {
  return(theme(axis.text = element_blank()))
}

ggplot() + 
  geom_sf(data = frame, fill = "white") +
  geom_sf(data = sites) +
  ggtitle("sites") +
  theme_nocoords()

ggplot() + 
  geom_sf(data = frame, fill = "white") + 
  geom_sf(data = units, fill = NA) +
  ggtitle("units") +
  theme_nocoords()

Detection functions

Modelling detection probability in fieldwalkr employs the concept of a detection functions. This concept is drawn from search theory, and specifically Banning and colleague's application of search theory to archaeological survey [@Banning2002-uv; @Banning2006-js; @Banning2011-wo].