PC2grid: Principal Component grid

In SantanderMetGroup/transformeR: A climate4R package for general climate data manipulation and transformation

Principal Component grid

Description

Convert a Principal Component to a 'time-series' grid.

Usage

PC2grid(
  prinCompObj,
  var = names(prinCompObj)[1],
  pc.idx = 1,
  scale = FALSE,
  opp = FALSE
)

Arguments

prinCompObj	A PCA object as returned by prinComp
var	Character string. Name of the variable for which to extract the PC. Defaults to the first variable in prinCompObj
pc.idx	Index position of the PC to be converted to a grid. Default to the first PC (pc.idx = 1).
scale	Logical. Should the time series values be scaled?. Default to FALSE. If TRUE, the default scale method is applied to each time series.
opp	Logical. Default to FALSE and rarely used. The data are multiplied by -1 so the opposite is returned.

Details

The XY coordinates correspond to the centroid of the window used for the PCA analysis. Details on the extent of this window are provided by the attributes domX and domY within the xyCoords component.

Currently, only the start days are returned. The end dates are not returned by the prinComp output.

Value

A 'time series' grid (i.e., a grid with dimensions lon and lat of length 1). It can be a multimember.

Author(s)

J. Bedia

See Also

Other pca: grid2PCs(), gridFromPCA(), prinComp()

Examples

require(climate4R.datasets)
data("CFS_Iberia_psl")
pc <- prinComp(CFS_Iberia_psl, v.exp = .95)
# Convert to grid the PC of the leading EOF (pc.idx = 1):
psl.index <- PC2grid(prinCompObj = pc, pc.idx = 1)
aggr.fun <- list(FUN = "mean", na.rm = TRUE)
psl.index <- aggregateGrid(psl.index, aggr.m = aggr.fun, aggr.y = aggr.fun,
                           aggr.mem = aggr.fun)
plot(as.Date(getRefDates(psl.index)), psl.index$Data, ty = "o",
     xlab = "year", ylab = "SLP index")
abline(h=0, lty = 2)
title(main = "1st PC of multimember mean DJF SLP over Iberia")
# opp can be used to obtain the inverse:
psl.index.inv <- PC2grid(prinCompObj = pc, pc.idx = 1, opp = TRUE)
psl.index.inv <- aggregateGrid(psl.index.inv, aggr.m = aggr.fun, aggr.y = aggr.fun,
                           aggr.mem = aggr.fun)
lines(as.Date(getRefDates(psl.index.inv)), psl.index.inv$Data, ty = "o", col = "red")

SantanderMetGroup/transformeR documentation built on Nov. 25, 2024, 1:25 p.m.