plot_anisotropy: Plot geometric anisotropy

In James-Thorson/VAST: Vector-Autoregressive Spatio-Temporal (VAST) Model

Plot geometric anisotropy

Description

plot_anisotropy plots the degree of geometric anisotropy

Usage

plot_anisotropy(
  Obj,
  FileName,
  ControlList = list(Width = 4, Height = 5, Res = 200, Units = "in"),
  type = "ellipse",
  Report = Obj$report(),
  TmbData = Obj$env$data
)

Details

VAST estimates a matrix H, representing a linear transformation of coordinates (typically eastings-northings) when computing a distance measure used to calculate correlations between sites. Geometric anisotropy represents the tendancy for correlations to decline faster in some direction than others, as represented by H. plot_anistropy visualizes the distance needed to achieve a correlation of approximately 10 centered at coordinates (0,0). Therefore, an ellipse with a major (long) axis pointed northwest-southeast will have correlations that decline slower along this axis then movement northeast-southwest. This ellipse is shown for both linear predictors, which share the same H estimate but differ in their overall estimated decorrelation rate.

James-Thorson/VAST documentation built on Feb. 9, 2025, 9:05 a.m.