Intersite: Modeling intersite correlation
In martindurocher/floodStat: FloodNet Regional Frequency Analysis

Description Usage Arguments Value Examples

Returns an intersite correlation model. The function computes the sample L-moments and estimates the empirical intersite correlation matrix between sites. A power exponential model can be fitted to smooth the intersite correlations in respect with the distance.

Intersite(x, ...)

## S3 method for class 'data.frame'
Intersite(x, distance = NULL, nmin = 20,
  smooth = FALSE, smooth.p = 1, smooth.hmax = Inf, nmom = 4)

## S3 method for class 'formula'
Intersite(form, x, ...)

`x`	Data. If not specify by a formula, the columns must be respectively : value, site and time.
`distance`	Matrix of distances. Must be of the same order as the sites appear in x.
`nmin`	Minimal number of paired observations for computing the empirical correlation, otherwise no correlation is assumed.
`smooth`	Logical. Should the intersite correlation matrix be smoothed by a power exponential model.
`smooth.p`	Exponent of the power exponential model.
`smooth.hmax`	Maximal distance used for fitting the power exponential model.
`nmom`	Number of L-moments to evaluate.
`form`	Formula that describes the response value, sites and time variables. Must have the form `value ~ site + time`.

`lmom`	Matrix of L-moments.
`nrec`	Record lengths.
`distance`	Distance matrix.
`corr`	Intersite correlation matrix.
`para`	Parameters of the power exponential model.

h <- flowAtlantic$distance
a <- flowAtlantic$ams

Intersite(ams ~ id + year, a)

Intersite(ams ~ id + year, a, distance = h,
          smooth = TRUE, smooth.hmax = 500)