CorQuantilesNullHyp: CorQuantilesNullHyp

In EarthSystemDiagnostics/corit: time scale dependent correlation of irregular time series

Description

Monte Carlo based procedure to estimate the quantiles of the correlation of independent noise which belong to the methods and applied filter to estimate the time scale dependent correlation of irregularly sampled time series

Usage

CorQuantilesNullHyp(timser1, timser2, beta.noise1, beta.noise2, detr, rep,
  quant, method = c("InterpolationMethod", "DirectFiltering",
  "IntegrandInterpolationMethod"), appliedFilter = c("gauss", "runmean",
  "lowpass"), fc, tn = seq(from = 10, to = max(c(index(timser1),
  index(timser2))), by = 10), dt, int.method = c("linear", "nearest"),
  k = 5)

Arguments

timser1, timser2	time series (zoo-objects) for which the Null Hypothesis of the correlation is to be applied
beta.noise1, beta.noise2	fitted slopes from the power spectrum of timser1 and timser2
detr	TRUE for removing a linear trend, else FALSE
rep	number of repetitions during the Monte Carlo procedure
quant	quantiles to be estimated
method	method to handle irregularity of sampling (DirectFiltering, IntegrandInterpolationMethod, InterpolationMethod)
appliedFilter	time domain filter (gauss, runmean, lowpass)
fc	cut-off frequency of the applied filter
tn	output vector (time) of the filtered data (only used in case of DirectFiltering and IntegrandInterpolationMethod)
dt	regular inter-observation time step of the interpolation (only used in case of InterpolationMethod)
int.method	kind of interpolation (linear, nearest neighbor) (only used in case of InterpolationMethod)
k	scaling factor to define the sharpness of the lowpass

Value

$corPair: estimated correlations during the Monte Carlo procedure, $Quantile: quantiles to be estimated

Examples

timeseries1 <- zoo(rnorm(100), order.by=sort(runif(100,min=1,max=1000)))
timeseries2 <- zoo(rnorm(100), order.by=sort(runif(100,min=1,max=1000)))
slopes <- estimateTimserSlopes(timeseries1,timeseries2,1)
CorQuantilesNullHyp(timser1=timeseries1, timser2=timeseries2, beta.noise1=slopes$s1, beta.noise2=slopes$s2, detr=FALSE, rep=1000, quant=c(0.05,0.95), method="InterpolationMethod", appliedFilter="gauss", fc=1/200, tn=NA, dt=10, int.method="linear", k=NA)

EarthSystemDiagnostics/corit documentation built on May 29, 2019, 1:39 p.m.