rangemodelR (version 1.0.4)

rangemod2d: Range Cohesion Models for Spatial Polygon Grids

Description

rangemod2d takes observed site by species matrix and returns expected species richness values of each site based on user defined neighbour relationships.

Usage

rangemod2d(spmat, shp, field, nb, rsize = c("observed", "unif"), var = NULL,
  reps, degen = FALSE, first = FALSE)

Arguments

spmat

a site by species matrix or data frame with species in columns

shp

shapefile of sites where species occurences are recorded

field

a number or character vector indicating which column in the dbf of shapefile is the unique id

nb

a neighbour object similar to one generated from poly2nb of spdep If 'nb' is NA then result is range scatter

rsize

which rangesizes to use for simulation, can be an integer vector of same length as number of species(collumns) or either 'observed' or'unif'. See details for explanations

var

an optional vector containing explanatory variable for constraining the randomization

reps

number of replicates

degen

If true, each randomized site by species matrix is saved and provided in output

first

If true, 'var' is used while choosing the first occurence as well.if 'var' is null, first is always set 'FALSE'

Value

A list containing following elements:

• "out.df" a data frame with four columns for mean, standard deviation, lower and upper limits of confidence intervals of predicted species richness

• "out.shp" same as the input shapefile with additional the four colums of 'out.df' in attribute table

• "degenerate.matrices" a list of all the randomized matrices(only present if 'degen' is TRUE)

Details

rangemod2d impliments simulations used by Rahbeck et.al. (2007) to species distribution data on a continuous grid. In 'spmat' the sites (rows) represent each cell in the grid.The species occurences across sites are randomly spread maintaining strict range cohesion. A neighbour object is used to limit the choice of cells during random selctions to immidiate neighbours. Options for creating four cell (rook) or eight cell (queen) neighbours can be accessed while creating the 'nb' object, (typically from package spdep). The randomisation proceeds by selecting a single site,(weighted by 'var' if provided, and first is TRUE), and then continues selecting one site at a time from a vector of available neighbours taken from 'nb' and weighted by 'var' if provided. The vector of available sites is updated after each site is selected.

References

Rahbek, C., Gotelli, N., Colwell, R., Entsminger, G., Rangel, T. & Graves, G. (2007) Predicting continental-scale patterns of bird species richness with spatially explicit models. Proceedings of the Royal Society B: Biological Sciences, 274, 165.

Gotelli, N.J., Anderson, M.J., Arita, H.T., Chao, A., Colwell, R.K., Connolly, S.R., Currie, D.J., Dunn, R.R., Graves, G.R. & Green, J.L. (2009) Patterns and causes of species richness: a general simulation model for macroecology. Ecology Letters, 12, 873-886.

Examples

# NOT RUN {
if(require(rgdal)&&require(maptools)&&require(rgeos)&&require(ggplot2)){
data(shp)
data(neigh_ob)
data(spmat)
mod.out <- rangemod2d(spmat,shp,"ID",nb = neigh_ob,rsize = "observed",
                       var = NULL,reps = 5)
shp.out <- mod.out$out.shp
shp.out.df <- shp.out@data
shp.out.fort <- fortify(shp.out,region = "ID")
seq <- match(shp.out.fort$id,shp.out.df$ID)
shp.out.gg <- data.frame(shp.out.fort,shp.out.df[seq,])
ggplot(shp.out.gg)+
  geom_map(map=shp.out.gg,aes_string(map_id="id",
                                     fill = "mod.rich"))+
  geom_path(aes(x = long,y = lat,group = group),colour = "white")+
  coord_equal() + theme_bw()+
  scale_fill_continuous(low = "white",high = "black")
}
# }