wavelet.bivariate: Bivariate wavelet variance using furier transforms.
In forestgeo/ctfs: Tools for the Analysis of Forest Dynamics

Description Usage Arguments Details Value Warning Author(s) Examples

Function to calculate the wavelet variance to evaluate the association between two point patterns using furier transforms.

1
2
3

wavelet.bivariate(raster1 = NULL, raster2 = NULL, coords1, coords2,
  gridsize = 1, plotdim = c(1000, 500), FUN = NULL, k0 = 8,
  dj = 0.15, graph = TRUE)

`coords1`	A matrix with raster data OR a table with two or three columns that can be used to calculate a raster, the first two columns are the coordinates and the third is the mark. Type must be specified.
`coords2`	Second dataset for the bivariate analyisis.
`gridsize`	Side of the square quadrate.
`plotdim`	The x and y dimensions of the plot.
`FUN`	Function to apply to the marked point pattern, by default it sums the values as would be used for sum of basal areas or sum of above graound biomass
`k0`	Numeric. Smoothing parameter of the wavelet filter. (k0 between 5.5-15), lower values of k0 produce a smoother wavelet variance.
`dj`	Numeric. Discretization of the scale axis.
`graph`	Logical. If TRUE plot the wavelet variace.
`type`	Type of data entered, 'raster'if a raster matrix, 'point'for a point pattern, or 'marked'for a marked point pattern;
`gridsize`	The quadrat size of the rasterization.

It accepts a raster data or a point pattern, but the type of data entered has to be specified in the argument type (xxx see section Warning).

The wavelet variance describes the spatial autocorrelation or aggregation of point distribution.

A wavelet variance greater than 1 indicates scales at which individuals are aggregated. A wavelet variance less than 1, indicates scales at which individuals are dis-aggregated. A wavelet variance equal to 1, indicates scales at which individuals are randomply distribuited (as Poisson process).

A graphical test is implemented on the null hypothesis of comple randomness.

If the wavelet variance is out of the conf bounds the point distribution is significantly different from a random process.

Dependencies: needs the package 'spatstat'and the CTFSRpackage

A list containing vectors for the wavelet variance, the scale of the wavelet variance, and the normalized variance.

If the argument type is ignored the function works. But one issue with this function is that the description mentions the argument type, but type is not part of the function definition not is passed to any other function. Confusingly, this function calls plot(), which has argument type but seems to have nothing to do with the tpye referred to in the description of this function. In the example, type is kept as a reminder of this issue but is with a comment.

Matteo Detto and Tania Brenes

## Not run: 
sp.one = subset(bciex::bci12t7mini, sp == "quaras")[, c("gx", "gy")]
sp.two = subset(bciex::bci12t7mini, sp == "cordal")[, c("gx", "gy")]
wv = wavelet.bivariate(
  coords = sp.one,
  coords2 = sp.two,
  #' type = 'point',  # dissabled because it errs, see section Warning
  k0 = 8,
  dj = 0.15,
  graph = TRUE
)
# plots the scale of aggregation

## End(Not run)