net function

Calculate NET

Computes the $NET$ parameter for a set of tree-ring records or other time-series data. UTF-8

net(x, weights = c(v = 1, g = 1))

Arguments

• x: A matrix or data.frame with at least two rows and two columns containing numeric data. The rows should represent a sequence of sampling points with uniform intervals (e.g. a range of years), but this is not checked. Each column is a time-series spanning either the whole time range or a part of it.

• weights: A numeric vector with two elements. Normally, variation ("v") and

  ("g") contribute to NET with equal weight. It is possible to use different weights by setting them here. The names of the vector are matched to c("v", "g") (see Examples ). If no names are given, the first element is the weight of variation.

Details

This function computes the $NET$ parameter (Esper et al., 2001). The overall $NET$ is an average of all (non-NA) yearly values $NET[j]$, which are computed as follows:

The yearly variation $v[j]$ is the standard deviation of the measurements of a single year divided by their mean. $1-G[j]$ is based on one definition of

$G[j]$, similar to but not the same as what glk

computes. Particularly, in the formula used by this function (Esper et al., 2001), simultaneous zero differences in two series are not counted as a synchronous change.

The weights of $v[j]$ and $1-G[j]$ in the sum can be adjusted with the argument weights (see above). As a rather extreme example, it is possible to isolate variation or by setting one of the weights to zero (see Examples ).

Returns

A list with the following components, in the same order as described here:

• all: a numeric vector containing $NET[j]$. Row names of x (if any) are copied here.

• average: a numeric value $NET$, the average of the "all" vector (NA values removed).

References

Esper, J., Neuwirth, B., and Treydte, K. (2001) A new parameter to evaluate temporal signal strength of tree-ring chronologies. Dendrochronologia, 19 (1), 93 102.

Author(s)

Mikko Korpela

Examples

library(utils)
data(ca533)
ca533.rwi <- detrend(rwl = ca533, method = "ModNegExp")
ca533.net <- net(ca533.rwi)
tail(ca533.net$all)
ca533.net$average
## Not run:

## Isolate the components of NET
ca533.v <- net(ca533.rwi, weights=c(v=1,0))
ca533.g <- net(ca533.rwi, weights=c(g=1,0))
## End(Not run)