offsetExpert: Build a raster of expert knowledge prior information
In ReseauBiodiversiteQuebec/mapSpecies: Map species distribution

Description Usage Arguments Details Value References

Rasterize an expert maps and weight its importance using the samples and their distance to the expert map with a five parameters logistic curve.

offsetExpert(
  expert,
  sPoints,
  raster,
  family,
  link,
  iniParam = c(upper = 1, lower = 1, rate = 1, shift = 1, skew = 1),
  ...
)

`expert`	A `SpatialPolygons` or a `SpatialPolygonsDataFrame` defining the expert map.
`sPoints`	A `SpatialPointsDataFrame` defining the sampled location with a single variable.
`raster`	An object of class `RasterLayer`.
`family`	This argument defines which reference distribution should be used to estimate the parameter of the logistic curve.
`link`	The link function to use if `family = "binomial"`. This argument is ignored if `family` is not `"binomial"`.
`iniParam`	A named vector defining the initial parameters of the logistc curve used to weight the importance of the expert map. The name of the parameters need to be `upper`, `lower`, `rate`, `shift` and `skew`. As default, the values of 1 are given to each parameters.
`...`	arguments pass to `link[gnlm]{bnlr}` if `family = "binomial"` or `link[gnlm]{gnlr}` otherwise.

This function uses the five parameters logistic curve proposed by Richards (1959) and suggested by Merow et al. (2017) as a way to weight the expert knowledge. The equation associated to this logistic curve is

W(x) = u - (u - l)/((1 + exp(-r(x-k)))^(1/s))

where eqnu and eqnl are the upper and lower asymptotes of the logistic curve, eqnr is a rate that gives flexibility to the curve from a sharpe step to a flat surface and eqns is a measure of skewness that adjust the symmetry of the decay on the edge of the expert map. As for eqnk, it shifts the curve inside or outside the expert map. Finally, eqnx is the distance to the expert map.

The detailed mathematics of the logistics curve is presented in the uniSpace vignette.

The five parameters logistic equation are estimated using nonlinear modelling with the help of the gnlm R package.

This function is designed to be with presence-absence, abundance and continuous data, but not presence-only. To estimate the parameter of the logistic curve using presence-only data use the bossMaps R package.

The function tends to send warnings message that stems from the calls to either bnlr or gnlr. They essentially inform users of different choices automatically made in the code.

An object of class RasterLayer that includes expert maps in its object.

Merow, C., A. M. Wilson, and W. Jetz. 2017. Integrating occurrence data and expert maps for improved species range predictions, Global Ecology and Biogeography 26:243–258.

Richards, F. J. 1959. A flexible growth function for empirical use. Journal of Experimental Botany 10:290–301

ReseauBiodiversiteQuebec/mapSpecies documentation built on Dec. 18, 2021, 9:57 a.m.