linearmask.object: Description of Linear Mask Objects

In secrlinear: Spatially Explicit Capture-Recapture for Linear Habitats

Description of Linear Mask Objects

Description

An object of class linearmask maps linear habitat, such as a river, in a way that is useful for spatially explicit capture–recapture analysis. The object contains two representations of the linear habitat: as features in a SpatialLinesDataFrame, and discretized as a set of equally spaced points, as in a mask object from secr.

Usually the object was created in a call to read.linearmask, which performs the discretization given linear input data and a chosen spacing.

A linearmask object does not explicitly include topological information i.e. the branching pattern of river networks. Instead, a network is constructed on-the-fly in asgraph by joining adjacent mask points (those within a certain small Euclidean distance). This results in 'shortcut' edges where lines approach at an acute angle (see the Examples for a demonstration).

A more complete, but also more cumbersome, solution for river networks might be to adapt the SpatialStreamNetwork object class of the SSN package (ver Hoef et al. 2014). Data input for SSN requires pre-processing in the proprietary software ArcGIS.

Value

The object itself is a dataframe of coordinates for the points comprising the mask, as for a 2-dimensional habitat mask (mask).

Additional information is stored as attributes (these usually remain hidden, but are essential except for ‘graph’):

spacingfactor	The spacing factor determines which points in a linear mask are joined to form a linear network, as happens automatically when computing distances with networkdistance. Points are joined when their Euclidean separation is less than spacingfactor x spacing. The default value is 1.1.
SLDF	The SLDF attribute is a SpatialLinesDataFrame (defined in sp) that provides the underlying description of the linear habitat.
graph	igraph network. Optionally computed by read.linearmask and saved as attribute, or computed on the fly by networkdistance

.

covariates

A required dataframe of covariate values, including at least the column ‘LineID’

Other attributes are common to masks and linearmasks:

spacing	In this context, the length of habitat in metres that is associated with each mask point (segment length).
meanSD	A matrix with mean and standard deviation of coordinates.
boundingbox	In this context, a dataframe giving the coordinate ranges spanned by the SLDF. Differs from 2-D mask in not being 'buffered' by spacing/2.

Further attributes of 2-dimensional masks (area, polygon etc.) are not relevant but may be present. The object has class c("linearmask", "mask", "data.frame") or c("list", "linearmask", "mask", "data.frame") for a multi-session linear mask.

References

Ver Hoef, J. M., Peterson, E. E., Clifford, D. and Shah, R. (2014) SSN : an R package for spatial statistical modeling on stream networks. Journal of Statistical Software 56(3), 1–45.

See Also

read.linearmask, mask, showpath, networkdistance

Examples

## make a complex linear mask
x <- seq(0, 4*pi, length = 200)
xy <- data.frame(x = x*100, y = sin(x)*300)
xy2 <- data.frame(x = x*100, y = cos(x)*300)
test <- read.linearmask(data = xy, spacing = 10)
test2 <- read.linearmask(data = xy2, spacing = 10)
test3 <- rbind(test, test2)

## visualize the igraph network used by networkdistance()
tmp <- asgraph(test3)
if(require('igraph')) {
  i2 <- get.edges(tmp,1:ecount(tmp))
  plot(test3, linecol = 'white', col = 'grey')
  segments(test3[i2[,1],1], test3[i2[,1],2], test3[i2[,2],1],
      test3[i2[,2],2], lwd=2, col='blue')
}

secrlinear documentation built on Oct. 17, 2023, 9:07 a.m.