kfunc {riverdist}

Plotting K-functions for a Set of Surveys

Description

Plots K-functions for locations in each of a set of surveys. In this implementation, this can be interpreted as the proportion of additional fish within a given distance. This will increase as a function of distance, and may provide evidence of clustering or dispersion features, particularly if the envelope is used.

Usage

kfunc(
  seg,
  vert,
  survey = NULL,
  rivers,
  lwd = 2,
  envelope = TRUE,
  envreps = 1000,
  envcol = "grey80",
  envborder = NA,
  maxdist = NULL,
  xlab = "Distance",
  ylab = "% within",
  showN = TRUE,
  whichplots = NULL,
  returnoutput = FALSE,
  ...
)

Arguments

seg

A vector of river locations (segment)

vert

A vector of river locations (vertex)

survey

A vector of survey IDs corresponding to the values of seg and vert. Defaults to NULL. If this argument is used, K-functions will be calculated for each unique survey, and separate plots will be produced.

rivers

The river network object to use

lwd

Line width used for plotting. Defaults to 2.

envelope

Whether to construct and display a 95 percent confidence envelope (see note.) Defaults to TRUE if survey is specified, and is automatically FALSE otherwise.

envreps

Number of bootstrap replicates to use for envelope calculation. Defaults to 1000.

envcol

Color to use for envelope plotting. Defaults to "grey80".

envborder

Border color to use for envelope plotting. Defaults to NA, which will result in no border being plotted.

maxdist

Maximum distance (x-axis value) for plotting. The default value (NULL) will result in an appropriate value being chosen.

xlab

X-coordinate label for plotting

ylab

Y-coordinate label for plotting

showN

Whether to show the sample size for each survey in each plot title. Defaults to TRUE.

whichplots

A vector of plots to produce, if multiple plots are produced. For example, specifying whichplot=c(2,3,4) will result in only the second, third, and fourth plots of the sequence being produced. Accepting the default (NULL) will result in all plots being produced.

returnoutput

Whether to return output instead of producing a plot. Defaults to FALSE.

...

Additional plotting parameters.

Note

K-function envelopes for each survey are constructed by bootstrapping all within-survey distances, that is, the distances between all individuals within each survey, for all surveys. This results in a confidence envelope under the assumption that spacing is independent of survey; therefore a survey K-function outside the envelope provides evidence of clustering or dispersal in that survey that is outside the typical range. An envelope is not available if only one survey is plotted.

A K-function above the envelope for a given distance range provides evidence of a greater number of individuals than expected at that distance range (clustering); A K-function below the envelope for a given distance range provides evidence of a smaller number of individuals than expected at that distance range (dispersal).

This function is distance-computation intensive, and will be extremely slow-running if a river network is used that does not have segment routes and/or distance lookup tables for fast distance computation. See buildsegroutes and/or buildlookup for more information.

Author(s)

Matt Tyers

Examples

data(Gulk, fakefish)

# # 10 plots will be created - recommend calling
# # par(mfrow=c(3,4))

kfunc(seg=fakefish$seg, vert=fakefish$vert, rivers=Gulk, survey=fakefish$flight,
envreps=100, maxdist=200000)

# # This shows relatively high amounts of clustering for surveys 1 and 8,
# # and relatively high amounts of dispersal in surveys 5 and 6.


# # plotting the survey locations that led to this calculation, for comparison

# # 10 plots will be created - recommend calling
# # par(mfrow=c(3,4))
for(i in 1:10) {
  plot(x=Gulk, segmentnum=FALSE, color=FALSE, main=i)
  riverpoints(seg=fakefish$seg[fakefish$flight==i], 
  vert=fakefish$vert[fakefish$flight==i], rivers=Gulk, col=2, pch=15)
}