plot.smnet: Visualise a Fitted Stream Network Model

Description

Plot linear, univariate and bivariate smooth effects and network smooth terms that resulting from a call to smnet().

Usage

# S3 method for smnet
plot(
  x,
  type = "covariates",
  se = FALSE,
  res = FALSE,
  weight = NULL,
  sites = FALSE,
  sites.col = NULL,
  sites.cex = 1,
  network.col = NULL,
  shadow = 0,
  key = TRUE,
  legend.text = NULL,
  legend.range = NULL,
  ...
)

Arguments

An object of class smnet.

type

Character string identifying the type of plot to be produced. The default, type = 'covariates', produces plots of all linear and smooth components (the latter corresponding to each appearance of m() in the model formula). type = 'full' plots the stream network fitted mean using the full set of spatial points contained in the associated SSN object. type = 'segments' plots the stream network fitted mean using a set of connected line segments to represent the spatial network, this can be faster for large networks than type = 'full'.

Logical. When TRUE and type = 'covariates', standard errors are shown on plots of linear and smooth terms. When type = 'segment' or type = 'full' spatial standard errors are plotted.

res

Logical. When TRUE, partial residuals are shown on plots of linear and smooth component. Ignored when type = 'full' or type = 'segments'.

weight

Positive real number that scales stream segment widths (as determined using Shreve order) to indicate relative size of stream segments. Ignored when type = 'covariate'. Currently only 'autoShreve' is supported, defaults to NULL in which all streams segments are plotted with lines with identical widths.

sites

Logical indicating whether locations of observation sites should be added to spatial plots. Ignored when type = 'covariate' and defaults to FALSE.

sites.col

Single colour to plot observation locations. If not provided, points will be coloured according to the default legend and average observation at each location.

sites.cex

Expansion factor for the size of plotted observation points.

network.col

Single colour to represent all stream segments. By default, network is coloured according to fitted values from model. Ignored when type = 'covariate'.

shadow

Positive scalar that adds a black outline to spatial stream segment plots, useful if the colour scale has many light colours. Ignored when type = 'covariate' and defaults to 0 (no shadow is drawn).

key

Logical. Plots a colour legend for network plots. Ignored when type = 'covariates'.

legend.text

Character annotation to add to color scale. Ignored if key = FALSE or type = 'covariates'.

legend.range

Range of values represented by the color scale. Ignored if key = FALSE or type = 'covariates'.

...

Other arguments passed to plot().