Package ‘earlywarnings’

February 27, 2013

Type Package

Title Early Warning Signals Toolbox for Detecting Critical Transitions in Timeseries

Version 1.0.32

Date 2013-02-18

Author Vasilis Dakos <[email protected]>, with contributions from S.R.

Carpenter, T. Cline, L. Lahti

Maintainer Vasilis Dakos <[email protected]>

Description The Early-Warning-Signals Toolbox provides methods for estimating

statistical changes in timeseries that can be used for identifying nearby
critical transitions. Based on Dakos et al (2012) Methods for Detecting
Early Warnings of Critical Transitions in Time Series Illustrated Using
Simulated Ecological Data. PLoS ONE 7(7):e41010

Depends R (>= 2.14.0), lmtest, nortest, stats, som, Kendall,KernSmooth, mo-

ments, fields, spam, tseries, quadprog, akima,ggplot2

LazyLoad yes

URL
http://www.early-warning-signals.org, http://www.vasilisdakos.net

License GPL (>= 2)

Collate 'data.R' 'BDSboot.R' 'bdstest_ews.R' 'ch_ews.R'

'ddjnonparam_ews.R' 'earlywarnings-internal.R' 'generic_ews.R'
'sensitivity_ews.R' 'surrogates_ews.R' 'potential_ews.R'

R topics documented:
bdstest_ews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
ch_ews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
ddjnonparam_ews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
foldbif . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
generic_ews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
livpotential_ews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
movpotential_ews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1
2 bdstest_ews

sensitivity_ews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
surrogates_ews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
YD2PB_grayscale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

bdstest_ews Description: BDS test Early Warning Signals

Description
bdstest_ews is used to estimate the BDS statistic to detect nonlinearity in the residuals of a
timeseries after first-difference detrending, fitting an ARMA(p,q) model, and fitting a GARCH(0,1)
model. The function is making use of bds.test from the tseries package.

Usage
bdstest_ews(timeseries, ARMAoptim = TRUE,
ARMAorder = c(1, 0), GARCHorder = c(0, 1), embdim = 3,
epsilon = c(0.5, 0.75, 1), boots = 1000,
logtransform = FALSE, interpolate = FALSE)

Arguments
timeseries a numeric vector of the observed univariate timeseries values or a numeric ma-
trix where the first column represents the time index and the second the observed
timeseries values. Use vectors/matrices with headings.
ARMAoptim is the order of the ARMA(p,q) model to be fitted on the original timeseries. If
TRUE the best ARMA model based on AIC is applied. If FALSE the ARMAorder
is used.
ARMAorder is the order of the AR(p) and MA(q) process to be fitted on the original time-
series. Default is p=1 q=0.
GARCHorder fits a GARCH model on the original timeseries where GARCHorder[1] is the
GARCH part and GARCHorder[2] is the ARCH part.
embdim is the embedding dimension (2, 3,... embdim) up to which the BDS test will be
estimated (must be numeric). Default value is 3.
epsilon is a numeric vector that is used to scale the standard deviation of the timeseries.
The BDS test is computed for each element of epsilon. Default is 0.5, 0.75 and
1.
boots is the number of bootstraps performed to estimate significance p values for the
BDS test. Default is 1000.
logtransform logical. If TRUE data are logtransformed prior to analysis as log(X+1). Default
is FALSE.
interpolate logical. If TRUE linear interpolation is applied to produce a timeseries of equal
length as the original. Default is FALSE (assumes there are no gaps in the
timeseries).

Details
See also bds.test{tseries} for more details. The function requires the installation of pack-
ages tseries and quadprog that are not available under Linux and need to be manually installed
under Windows.
ch_ews 3

Value

bdstest_ews returns output on the R console that summarizes the BDS test statistic for all em-
bedding dimensions and epsilon values used, and for first-differenced data, ARMA(p.q) resid-
uals, and GARCH(0,1) residuals). Also the significance p values are returned estimated both by
comparing to a standard normal distribution and by bootstrapping.
In addition, bdstest_ews returns a plot with the original timeseries, the residuals after first-
differencing, and fitting the ARMA(p,q) and GARCH(0,1) models. Also the autocorrelation acf
and partial autocorrelation pacf functions are estimated serving as guides for the choice of lags of
the linear models fitted to the data.

Author(s)

S. R. Carpenter, modified by V. Dakos

References

J. B. Cromwell, W. C. Labys and M. Terraza (1994): Univariate Tests for Time Series Models, Sage,
Thousand Oaks, CA, pages 32-36.
Dakos, V., et al (2012)."Methods for Detecting Early Warnings of Critical Transitions in Time Series
Illustrated Using Simulated Ecological Data." PLoS ONE 7(7): e41010. doi:10.1371/journal.pone.0041010

See Also

generic_ews; ddjnonparam_ews; bdstest_ews; sensitivity_ews; surrogates_ews;

ch_ews; movpotential_ews; livpotential_ews

Examples
#
data(foldbif)
bdstest_ews(foldbif,ARMAoptim=FALSE,ARMAorder=c(1,0),embdim=3,epsilon=0.5, boots=200,logt

ch_ews Description: Conditional Heteroskedasticity

Description

ch_ews is used to estimate changes in conditional heteroskedasticity within rolling windows along
a timeseries

Usage

ch_ews(timeseries, winsize = 10, alpha = 0.1,

optim = TRUE, lags = 4, logtransform = FALSE,
interpolate = FALSE)
4 ch_ews

Arguments
timeseries a numeric vector of the observed timeseries values or a numeric matrix where
the first column represents the time index and the second the observed timeseries
values. Use vectors/matrices with headings.
winsize is length of the rolling window expressed as percentage of the timeseries length
(must be numeric between 0 and 100). Default is 10%.
alpha is the significance threshold (must be numeric). Default is 0.1.
optim logical. If TRUE an autoregressive model is fit to the data within the rolling
window using AIC optimization. Otherwise an autoregressive model of specific
order lags is selected.
lags is a parameter that determines the specific order of an autoregressive model to
fit the data. Default is 4.
logtransform logical. If TRUE data are logtransformed prior to analysis as log(X+1). Default
is FALSE.
interpolate logical. If TRUE linear interpolation is applied to produce a timeseries of equal
length as the original. Default is FALSE (assumes there are no gaps in the
timeseries).

Details
see ref below

Value
ch_ews returns a matrix that contains:

time the time index.

r.squared the R2 values of the regressed residuals.
critical.value
the chi-square critical value based on the desired alpha level for 1 degree of
freedom divided by the number of residuals used in the regression.
test.result logical. It indicates whether conditional heteroskedasticity was significant.
ar.fit.order the order of the specified autoregressive model- only informative if optim
FALSE was selected.

In addition, ch_ews plots the original timeseries and the R2 where the level of significance is also
indicated.

Author(s)
T. Cline, modified by V. Dakos

References
Seekell, D. A., et al (2011). "Conditional heteroscedasticity as a leading indicator of ecological
regime shifts." American Naturalist 178(4): 442-451
Dakos, V., et al (2012)."Methods for Detecting Early Warnings of Critical Transitions in Time Series
Illustrated Using Simulated Ecological Data." PLoS ONE 7(7): e41010. doi:10.1371/journal.pone.0041010
circulation 5

See Also
generic_ews; ddjnonparam_ews; bdstest_ews; sensitivity_ews; surrogates_ews;
ch_ews; movpotential_ews; livpotential_ews

Examples
data(foldbif)
out=ch_ews(foldbif, winsize=50, alpha=0.05, optim=TRUE, lags)

circulation circulation data set

Description
circulation data set

Format
TBA

Source
TBA

References
See citation("earlywarnings")

Examples
#

ddjnonparam_ews Description: Drift Diffusion Jump Nonparametrics Early Warning

Signals

Description
ddjnonparam_ews is used to compute nonparametrically conditional variance, drift, diffusion
and jump intensity in a timeseries. It also interpolates to obtain the evolution of the nonparametric
statistics in time.

Usage
ddjnonparam_ews(timeseries, bandwidth = 0.6, na = 500,
logtransform = TRUE, interpolate = FALSE)
6 ddjnonparam_ews

Arguments

timeseries a numeric vector of the observed univariate timeseries values or a numeric ma-
trix where the first column represents the time index and the second the observed
timeseries values. Use vectors/matrices with headings.
bandwidth is the bandwidht of the kernel regressor (must be numeric). Default is 0.6.
na is the number of points for computing the kernel (must be numeric). Default is
500.
logtransform logical. If TRUE data are logtransformed prior to analysis as log(X+1). Default
is FALSE.
interpolate logical. If TRUE linear interpolation is applied to produce a timeseries of equal
length as the original. Default is FALSE (assumes there are no gaps in the
timeseries).

Details

The approach is based on estimating terms of a drift-diffusion-jump model as a surrogate for the
unknown true data generating process: [1] dx = f (x, θ)dt + g(x, θ)dW + dJ Here x is the state
variable, f() and g() are nonlinear functions, dW is a Wiener process and dJ is a jump process.
Jumps are large, one-step, positive or negative shocks that are uncorrelated in time.

Value

ddjnonparam_ews returns an object with elements:

avec is the mesh for which values of the nonparametric statistics are estimated.
S2.vec is the conditional variance of the timeseries x over avec.
TotVar.dx.vec
is the total variance of dx over avec.
Diff2.vec is the diffusion estimated as total variance - jumping intensity
vs avec.
LamdaZ.vec is the jump intensity over avec.
Tvec1 is the timeindex.
S2.t is the conditional variance of the timeseries x data over Tvec1.
TotVar.t is the total variance of dx over Tvec1.
Diff2.t is the diffusion over Tvec1.
Lamda.t is the jump intensity over Tvec1.

In addition, ddjnonparam_ews returns a first plot with the original timeseries and the residuals
after first-differencing. A second plot shows the nonparametric conditional variance, total variance,
diffusion and jump intensity over the data, and a third plot the same nonparametric statistics over
time.

Author(s)

S. R. Carpenter, modified by V. Dakos

foldbif 7

References

Carpenter, S. R. and W. A. Brock (2011). "Early warnings of unknown nonlinear shifts: a nonpara-
metric approach." Ecology 92(12): 2196-2201
Dakos, V., et al (2012)."Methods for Detecting Early Warnings of Critical Transitions in Time Series
Illustrated Using Simulated Ecological Data." PLoS ONE 7(7): e41010. doi:10.1371/journal.pone.0041010

See Also

generic_ews; ddjnonparam_ews; bdstest_ews; sensitivity_ews;surrogates_ews;

ch_ews; movpotential_ews; livpotential_ews

Examples

data(foldbif)
output<-ddjnonparam_ews(foldbif,bandwidth=0.6,na=500,
logtransform=TRUE,interpolate=FALSE)

foldbif foldbif data set

Description

foldbif data set

Format

TBA

Source

TBA

References

See citation("earlywarnings")

Examples

#
8 generic_ews

generic_ews Description: Generic Early Warning Signals

Description

generic_ews is used to estimate statistical moments within rolling windows along a timeserie

Usage

generic_ews(timeseries, winsize = 50,

detrending = c("no", "gaussian", "linear", "first-diff"),
bandwidth = NULL, logtransform = FALSE,
interpolate = FALSE, AR_n = FALSE,
powerspectrum = FALSE)

Arguments

timeseries a numeric vector of the observed univariate timeseries values or a numeric ma-
trix where the first column represents the time index and the second the observed
timeseries values. Use vectors/matrices with headings. If the powerspectrum is
to be plotted as well, the timeseries lenght should be even number.
winsize is the size of the rolling window expressed as percentage of the timeseries length
(must be numeric between 0 and 100). Default is 50%.
bandwidth is the bandwidth used for the Gaussian kernel when gaussian filtering is applied.
It is expressed as percentage of the timeseries length (must be numeric between
0 and 100). Alternatively it can be given by the bandwidth selector bw.nrd0
(Default).
detrending the timeseries can be detrended/filtered prior to analysis. There are four options:
gaussian filtering, linear detrending and first-differencing. De-
fault is no detrending.
logtransform logical. If TRUE data are logtransformed prior to analysis as log(X+1). Default
is FALSE.
interpolate logical. If TRUE linear interpolation is applied to produce a timeseries of equal
length as the original. Default is FALSE (assumes there are no gaps in the
timeseries).
AR_n logical. If TRUE the best fitted AR(n) model is fitted to the data. Default is
FALSE.
powerspectrum
logical. If TRUE the power spectrum within each rolling window is plotted.
Default is FALSE.

Details

see ref below

generic_ews 9

Value
generic_ews returns a matrix that contains:

tim the time index.

ar1 the autoregressive coefficient ar(1) of a first order AR model
fitted on the data within the rolling window.
sd the standard deviation of the data estimated within each rolling window.
sk the skewness of the data estimated within each rolling window.
kurt the kurtosis of the data estimated within each rolling window.
cv the coefficient of variation of the data estimated within each rolling
window.
returnrate the return rate of the data estimated as 1-ar(1) cofficient within each rolling
window.
densratio the density ratio of the power spectrum of the data estimated as the ratio
of low frequencies over high frequencies within each rolling window.
acf1 the autocorrelation at first lag of the data estimated within each
rolling window.

In addition, generic_ews returns three plots. The first plot contains the original data, the de-
trending/filtering applied and the residuals (if selected), and all the moment statistics. For each
statistic trends are estimated by the nonparametric Kendall tau correlation. The second plot, if
asked, quantifies resilience indicators fitting AR(n) selected by the Akaike Information Criterion.
The third plot, if asked, is the power spectrum estimated by spec.ar for all frequencies within
each rolling window.

Author(s)
Vasilis Dakos <[email protected]>

References
Ives, A. R. (1995). "Measuring resilience in stochastic systems." Ecological Monographs 65: 217-
233
Dakos, V., et al (2008). "Slowing down as an early warning signal for abrupt climate change."
Proceedings of the National Academy of Sciences 105(38): 14308-14312
Dakos, V., et al (2012)."Methods for Detecting Early Warnings of Critical Transitions in Time Series
Illustrated Using Simulated Ecological Data." PLoS ONE 7(7): e41010. doi:10.1371/journal.pone.0041010

See Also
generic_ews; ddjnonparam_ews; bdstest_ews; sensitivity_ews; surrogates_ews;
ch_ews; movpotential_ews; livpotential_ews;

Examples
data(foldbif)
out=generic_ews(foldbif,winsize=50,detrending="gaussian",
bandwidth=5,logtransform=FALSE,interpolate=FALSE)
10 livpotential_ews

livpotential_ews Description: Potential Analysis

Description
livpotential_ews performs one-dimensional potential estimation derived from a uni-variate
timeseries

Usage
livpotential_ews(x, std = 1, bw = -1, xi = NULL,
weights = c(), grid.size = 200)

Arguments
x data vector
std the standard deviation of the noise (defaults to 1, so then you use scaled poten-
tials
bw bandwidth for kernel estimation
xi x values at which the potential is estimated
weights optional weights in ksdensity (used by movpotentials).
grid.size grid size

Details
see ref below

Value
livpotential returns a list with the following elements:

xi the grid of points on which the potential is estimated

pot the actual value of the potential
minima the grid points at which the potential has minimum values
maxima the grid points at which the potential has maximum values
bw bandwidth of kernel used

Author(s)
Based on Matlab code from Egbert van Nes modified by Leo Lahti. Implemented in early warnings
package by V. Dakos.

References
Livina, VN, F Kwasniok, and TM Lenton, 2010. Potential analysis reveals changing number of
climate states during the last 60 kyr . Climate of the Past, 6, 77-82.
Dakos, V., et al (2012)."Methods for Detecting Early Warnings of Critical Transitions in Time Series
Illustrated Using Simulated Ecological Data." PLoS ONE 7(7): e41010. doi:10.1371/journal.pone.0041010
movpotential_ews 11

See Also
generic_ews; ddjnonparam_ews; bdstest_ews; sensitivity_ews;surrogates_ews;
ch_ews;movpotential_ews

Examples
data(foldbif)
res <- livpotential_ews(foldbif)
plot(res$xi, res$pot)

movpotential_ews Description: Moving Average Potential

Description
movpotential_ews reconstructs a potential derived from data along a gradient of a given pa-
rameter the movpotential_ews calculates the potential for values that correspond to a particular
parameter. see ref below

Usage
movpotential_ews(X, param, sdwindow = NULL, bw = -1,
minparam = NULL, maxparam = NULL, npoints = 50,
thres = 0.002, std = 1, grid.size = 200, cutoff = 0.5)

Arguments
X a vector of the X observations of the state variable of interest
param parameter values that correspond to the X observations
sdwindow window for smoothing kernels (over the param axis)
bw bandwidth used for smoothing kernels
minparam minimum value of parameter on which to estimate potential
maxparam maximum value of parameter on which to estimate potential
npoints number of potentials
thres threshold for local minima to be discarded
std std
grid.size number of evaluation points
cutoff the cuttof value to estimate minima and maxima in the potential
Returns:

Value
A list with the following elements:
pars values of the covariate parameter as matrix
xis values of the x as matrix
pots smoothed potentials
mins minima in the densities (-potentials; neglecting local optima)
maxs maxima in densities (-potentials; neglecting local optima)
plot an object that displays the potential estimated in 2D
12 sensitivity_ews

Author(s)
Based on Matlab code from Egbert van Nes modified by Leo Lahti. Implemented in early warnings
package by V. Dakos.

References
Hirota, M., Holmgren, M., van Nes, E.H. & Scheffer, M. (2011). Global resilience of tropical forest
and savanna to critical transitions. Science, 334, 232-235.

See Also
generic_ews; ddjnonparam_ews; bdstest_ews; sensitivity_ews;surrogates_ews;
ch_ews; livpotential_ews

Examples
X = c(rnorm(1000, mean = 0), rnorm(1000, mean = -2), rnorm(1000, mean = 2))
param = seq(0,5,length=3000)
res <- movpotential_ews(X, param, npoints = 100, thres = 0.003)

sensitivity_ews Description: Sensitivity Early Warning Signals

Description
sensitivity_ews is used to estimate trends in statistical moments for different sizes of rolling
windows along a timeseries. The trends are estimated by the nonparametric Kendall tau correlation
coefficient.

Usage
sensitivity_ews(timeseries,
indicator = c("ar1", "sd", "acf1", "sk", "kurt", "cv", "returnrate", "densra
winsizerange = c(25, 75), incrwinsize = 25,
detrending = c("no", "gaussian", "linear", "first-diff"),
bandwidthrange = c(5, 100), incrbandwidth = 20,
logtransform = FALSE, interpolate = FALSE)

Arguments
timeseries a numeric vector of the observed univariate timeseries values or a numeric ma-
trix where the first column represents the time index and the second the observed
timeseries values. Use vectors/matrices with headings.
indicator is the statistic (leading indicator) selected for which the sensitivity analysis
is perfomed. Currently, the indicators supported are: ar1 autoregressive co-
efficient of a first order AR model, sd standard deviation, acf1 autocorre-
lation at first lag, sk skewness, kurt kurtosis, cv coeffcient of variation,
returnrate, and densratio density ratio of the power spectrum at low
frequencies over high frequencies.
winsizerange is the range of the rolling window sizes expressed as percentage of the timeseries
length (must be numeric between 0 and 100). Default is 25% - 75%.
sensitivity_ews 13

incrwinsize increments the rolling window size (must be numeric between 0 and 100). De-
fault is 25.
detrending the timeseries can be detrended/filtered. There are three options: gaussian
filtering, linear detrending and first-differencing. Default is no
detrending.
bandwidthrange
is the range of the bandwidth used for the Gaussian kernel when gaussian filter-
ing is selected. It is expressed as percentage of the timeseries length (must be
numeric between 0 and 100). Default is 5% - 100%.
incrbandwidth
is the size to increment the bandwidth used for the Gaussian kernel when gaus-
sian filtering is applied. It is expressed as percentage of the timeseries length
(must be numeric between 0 and 100). Default is 20.
logtransform logical. If TRUE data are logtransformed prior to analysis as log(X+1). Default
is FALSE.
interpolate logical. If TRUE linear interpolation is applied to produce a timeseries of equal
length as the original. Default is FALSE (assumes there are no gaps in the
timeseries).

Details
see ref below

Value
sensitivity_ews returns a matrix that contains the Kendall tau rank correlation estimates for
the rolling window sizes (rows) and bandwidths (columns), if gaussian filtering is se-
lected.
In addition, sensitivity_ews returns a plot with the Kendall tau estimates and their p-values
for the range of rolling window sizes used, together with a histogram of the distributions of the
statistic and its significance. When gaussian filtering is chosen, a contour plot is produced
for the Kendall tau estimates and their p-values for the range of both rolling window sizes and
bandwidth used. A reverse triangle indicates the combination of the two parameters for which the
Kendall tau was the highest

Author(s)
Vasilis Dakos <[email protected]>

References
Dakos, V., et al (2008). "Slowing down as an early warning signal for abrupt climate change."
Proceedings of the National Academy of Sciences 105(38): 14308-14312
Dakos, V., et al (2012)."Methods for Detecting Early Warnings of Critical Transitions in Time Series
Illustrated Using Simulated Ecological Data." PLoS ONE 7(7): e41010. doi:10.1371/journal.pone.0041010

See Also
generic_ews; ddjnonparam_ews; bdstest_ews; sensitivity_ews; surrogates_ews;
ch_ews; movpotential_ews; livpotential_ews
14 surrogates_ews

Examples
data(foldbif)
output=sensitivity_ews(foldbif,indicator="sd",detrending="gaussian",
incrwinsize=25,incrbandwidth=20)

surrogates_ews Description: Surrogates Early Warning Signals

Description
surrogates_ews is used to estimate distributions of trends in statistical moments from different
surrogate timeseries generated after fitting an ARMA(p,q) model on the data. The trends are esti-
mated by the nonparametric Kendall tau correlation coefficient and can be compared to the trends
estimated in the original timeseries to produce probabilities of false positives.

Usage
surrogates_ews(timeseries,
indicator = c("ar1", "sd", "acf1", "sk", "kurt", "cv", "returnrate", "densra
winsize = 50,
detrending = c("no", "gaussian", "linear", "first-diff"),
bandwidth = NULL, boots = 100, logtransform = FALSE,
interpolate = FALSE)

Arguments
timeseries a numeric vector of the observed univariate timeseries values or a numeric ma-
trix where the first column represents the time index and the second the observed
timeseries values. Use vectors/matrices with headings.
indicator is the statistic (leading indicator) selected for which the surrogate timeseries
are produced. Currently, the indicators supported are: ar1 autoregressive co-
efficient of a first order AR model, sd standard deviation, acf1 autocorre-
lation at first lag, sk skewness, kurt kurtosis, cv coeffcient of variation,
returnrate, and densratio density ratio of the power spectrum at low
frequencies over high frequencies.
winsize is the size of the rolling window expressed as percentage of the timeseries length
(must be numeric between 0 and 100). Default valuise 50%.
detrending the timeseries can be detrended/filtered prior to analysis. There are three op-
tions: gaussian filtering, linear detrending and first-differencing.
Default is no detrending.
bandwidth is the bandwidth used for the Gaussian kernel when gaussian filtering is selected.
It is expressed as percentage of the timeseries length (must be numeric between
0 and 100). Alternatively it can be given by the bandwidth selector bw.nrd0
(Default).
boots the number of surrogate data. Default is 100.
logtransform logical. If TRUE data are logtransformed prior to analysis as log(X+1). Default
is FALSE.
interpolate logical. If TRUE linear interpolation is applied to produce a timeseries of equal
length as the original. Default is FALSE (assumes there are no gaps in the
timeseries).
surrogates_ews 15

Details

see ref below

Value

surrogates_ews returns a matrix that contains:

Kendall tau estimate original

the trends of the original timeseries.
Kendall tau p-value original
the p-values of the trends of the original timeseries.
Kendall tau estimate surrogates
the trends of the surrogate timeseries.
Kendall tau p-value surrogates
the associated p-values of the trends of the surrogate timeseries.
significance p
the p-value for the original Kendall tau rank correlation estimate compared to
the surrogates.

In addition, surrogates_ews returns a plot with the distribution of the surrogate Kendall tau
estimates and the Kendall tau estimate of the original series. Vertical lines indicate the 5% and 95%
significance levels.

Author(s)

Vasilis Dakos <[email protected]>

References

Dakos, V., et al (2008). "Slowing down as an early warning signal for abrupt climate change."
Proceedings of the National Academy of Sciences 105(38): 14308-14312
Dakos, V., et al (2012)."Methods for Detecting Early Warnings of Critical Transitions in Time Series
Illustrated Using Simulated Ecological Data." PLoS ONE 7(7): e41010. doi:10.1371/journal.pone.0041010

See Also

generic_ews; ddjnonparam_ews; bdstest_ews; sensitivity_ews; surrogates_ews;

ch_ews; movpotential_ews; livpotential_ews

Examples

data(foldbif)
output=surrogates_ews(foldbif,indicator="sd",winsize=50,detrending="gaussian",
bandwidth=10,boots=200,logtransform=FALSE,interpolate=FALSE)
16 YD2PB_grayscale

YD2PB_grayscale YD2PB_grayscale data set

Description
YD2PB_grayscale data set

Format
TBA

Source
TBA

References
See citation("earlywarnings")

Examples
#