rtsa.eot: EOT (Empirical Orthogonal Teleconnections analysis)

In ffilipponi/rtsa: Raster Time Series Analysis

Description

This function conducts an Empirical Orthogonal Teleconnection analysis (EOT) of a raster time series predictor using "remote" package especially designed to identify locations of enhanced potential and explain spatio-temporal variability within the same geographic field.

Usage

rtsa.eot(rasterts, rastermask = NULL, nu = NULL, gapfill = "none",
  predictor = "all", standardised = TRUE, reduce.both = FALSE,
  type = "rsq", verbose = FALSE)

Arguments

rasterts	Input raster time series as RasterStackTS or RasterBrickTS object.
rastermask	Either a RasterLayer or "compute". Raster layer to use as a mask. When "compute" is set raster mask is computed to remove all pixels with incomplete time series.
nu	Numeric. Defines the number of EOTs to return. Defaults to return the first 2 EOT modes.
gapfill	Character. Defines the algorithm to be used to interpolate pixels with incomplete temporal profiles Accepts argument supported as method in function rtsa.gapfill.
predictor	Character. Defines the predictor components to export from those available in Value section of eot. Supports one or more of the following: 'all', 'r_predictor', 'rsq_predictor', 'rsq_sums_predictor', 'int_predictor', 'slp_predictor', 'p_predictor'.
...	Additional arguments to be passed through to function eot.

Value

Object of class EOTstack-class containing the following components:

	eot	EOT temporal profiles corresponding to base point coordinates as xts object
	total_variance	Numeric. Total explained variance of input raster time series by the entire set of computed EOTs
	explained_variance	Numeric vector. Percentage of variance explained by each EOT mode with respect to the total variance of input raster time series
	r_predictor	RasterBrick. Correlation coefficients between the base point and each pixel of the predictor domain as RasterBrick object (only exported if predictor = "all" or predictor = "r_predictor")
	rsq_predictor	RasterBrick. Coefficient of determination between the base point and each pixel of the predictor domain as RasterBrick object (only exported if predictor = "all" or predictor = "rsq_predictor")
	rsq_sums_predictor	RasterBrick. Sums of correlation coefficients between the base point and each pixel of the predictor domain as RasterBrick object (only exported if predictor = "all" or predictor = "rsq_sums_predictor")
	int_predictor	RasterBrick. Intercept of the regression equation for each pixel of the predictor domain as RasterBrick object (only exported if predictor = "all" or predictor = "int_predictor")
	slp_predictor	RasterBrick. Slope of the regression equation for each pixel of the predictor domain as RasterBrick object (only exported if predictor = "all" or predictor = "slp_predictor")
	r_predictor	RasterBrick. The significance (p-value) of the the regression equation for each pixel of the predictor domain as RasterBrick object (only exported if predictor = "all" or predictor = "p_predictor")

Author(s)

Federico Filipponi

References

van den Dool H.M., Saha S., Johansson A. (2000). Empirical Orthogonal Teleconnections. Journal of Climate, Volume 13, Issue 8, pp. 1421-1435. http://journals.ametsoc.org/doi/abs/10.1175/1520-0442%282000%29013%3C1421%3AEOT%3E2.0.CO%3B2

See Also

eot, rtsa.eof, rtsa.gapfill, EOTstack

Examples

## Not run: 
## create raster time series using the 'pacificSST' data from 'remote' package
require(remote)

data(pacificSST)
pacificSST[which(getValues(pacificSST == 0))] <- NA # set NA values
# create rts object
rasterts <- rts(pacificSST, seq(as.Date('1982-01-15'), as.Date('2010-12-15'), 'months'))

## generate raster mask
rmk <- pacificSST[[1]] # create raster mask
values(rmk) <- 1 # set raster mask values
rmk[which(is.na(getValues(pacificSST[[1]])))] <- 0 # set raster mask values

## compute EOT
# compute the first 2 EOTs
eot_result <- rtsa.eot(rasterts=rasterts, rastermask=rmk, nu=2)
# compute the first 2 EOTs and export only the compontents 'r_predict' and 'p_predictor'
eot_rp <- rtsa.eot(rasterts=rasterts, rastermask="compute", nu=2, predictor=c("r_predictor", "p_predictor"))
# compute the first 2 EOTs using the index of agreement
eot_ioa <- rtsa.eot(rasterts=rasterts, rastermask=rmk, nu=2, type="ioa", predictor="all", verbose=T)

## End(Not run)

ffilipponi/rtsa documentation built on Oct. 18, 2019, 12:37 a.m.