RooksGradient: Calculate Gradient Statistics in the Rook's Neighbourhood

In ICvectorfields: Vector Fields from Spatial Time Series of Population Abundance

Calculate Gradient Statistics in the Rook's Neighbourhood

Description

The movement of populations into adjacent cells may sometimes be influenced by the gradient of some predictive variable. This function enables the calculation of a simple gradient statistic in the rook's neighbourhood of each cell in a dataset. The statistic must first be computed for each grid cell using SubgridStats. Then for each grid, the RooksGradient function computes the arithmetic average of the difference between the statistic at the focal grid cell and the statistic in the four (or fewer) adjacent neighbours in its Rook's neibourhood. This arithmetically averaged difference is then returned under the column header of 'Gradient'. A negative gradient estimate indicates that the statistic in the central cell is higher than that in neighbouring cells whereas a positive gradient estimate indicates the opposite.

Usage

RooksGradient(vfdf, statistic = "mean")

Arguments

vfdf	A data frame as returned by SubgridStats
statistic	desired output statistic: It should be one of "mean", "var", or "sum". Default setting is mean.

Value

A data frame similar to vfdf except that it includes an additional column called Gradient as described above.

Examples

# creating pattern patterns
Mat1 <- matrix(rep(0,9*9), nrow = 9)
Mat1[c(4:6), c(4:6)] <- 2
Mat1[c(4:6), c(1:3)] <- 1
Mat1[c(1:3), c(4:6)] <- 1
Mat1[c(7:9), c(4:6)] <- 1
Mat1[c(4:6), c(7:9)] <- 1
Mat1

Rast1 <- terra::rast(Mat1)
terra::plot(Rast1)

# calculating the mean in 9 subgrids
(statsdf1 <- SubgridStats(Rast1, factv1 = 3, facth1 = 3, statistic = "mean"))

# computing the gradient statistic on the mean
(graddf1 <- RooksGradient(statsdf1, statistic = "mean"))
# the Gradient statistic in the central grid in row 5 should
# be equal to negative one

ICvectorfields documentation built on March 18, 2022, 7:34 p.m.