addcov2mask: Add covariates to mask from raster data

In david-borchers/secrgam: Spatially Explicit Capture-Recapture with Generalised Additive Density Models

Description

Adds covariates to a habitat mask from a GIS landcover data in raster format.

Usage

addcov2mask(mask, raster, names = NULL, distance.to = NULL, drop = NULL,
  plot = TRUE, cols = NULL, domulti = FALSE)

Arguments

mask	mask object (as used by secrgam.fit and secr.fit)
raster	list with components x,y, and z in the same style as used by contour, persp, image etc. x and y are the X and Y grid values and z is a matrix with the corresponding values of the surface (this gets passed to interp.surface)
names	names to associate with the values in raster (for multiple layer types)
distance.to	names of any layers for which ‘distance to’ covariates are to be calcuated
drop	names of layers to be deleted from the mask
plot	if TRUE then an image plot will be drawn using prep4image
cols	colours to associate with the layers (only used if plot = TRUE)
domulti	logical for categorical integer variable (like land use class) into columns of binary data for each category at each mask point if TRUE. If FALSE then treats creates a single column with the integer integer values.

Details

NOTE: This is a function under development, with rather limited and specific functionality. The secr function addCovariates is probably more useful at this stage. If typeof(raster$z) = "double" then elements of raster$z are assumed to represent a single layer type. If typeof(raster$z) = "integer" then each unique integer is assumed to represent a unique landcover type. If the mask already contains covariates whose names are identical to those supplied in names then these will be overwritten.

Value

Returns a new mask object.

Examples

# load mask and raster data
data(Boland.leopards1)
data(Boland.landuse.image)
data(Boland.alt.image)

# save par settings
op = par(no.readonly = TRUE)

# single landcover type
par(mfrow = c(1,1), mar = c(2,2,2,2), oma = c(0,0,0,0))
newmask = addcov2mask(Boland.mask1, Boland.alt.image, "altitude")
head(covariates(newmask))

# mutiple landcover types
newmask = addcov2mask(
  Boland.mask1, Boland.landuse.image,
  names = c("Unknown", "Natural", "Cultivated", "Degraded", "Urban", "Water", "Plantation"),
  cols = c("black", "green3", "yellow", "tan", "dimgray", "blue", "gold2"),
)
head(covariates(newmask))

# mutiple landcover types - including distance to water and urbarn
par(mfrow = c(1,3))
newmask = addcov2mask(
  Boland.mask1, Boland.landuse.image,
  names = c("Unknown", "Natural", "Cultivated", "Degraded", "Urban", "Water", "Plantation"),
  cols = c("black", "green3", "yellow", "tan", "dimgray", "blue", "gold2"),
  distance.to = c("Urban","Water")
)
head(covariates(newmask))

# mutiple landcover types - including distance to water and urbarn (and deleting these mask points)
newmask = addcov2mask(
  Boland.mask1, Boland.landuse.image,
  names = c("Unknown", "Natural", "Cultivated", "Degraded", "Urban", "Water", "Plantation"),
  cols = c("black", "green3", "yellow", "tan", "dimgray", "blue", "gold2"),
  distance.to = c("Urban","Water"),
  drop = c("Urban","Water")
)
head(covariates(newmask))

# reset par settings
par(op)

david-borchers/secrgam documentation built on May 14, 2019, 9:30 p.m.