simAtSite: Simulation of at-site data
In martindurocher/floodRFA: A R-package for regional frequency of floods

Description Usage Arguments Details See Also Examples

Returns a matrix of the simulation of at-site data based on the choice of a marginal distribution and a copula model.

1
2
3

simAtSite(n, distr = "gev", marg = c(0, 1, 0), cop = "norm", h = NULL,
  type = "exp", p = NULL, sigma = NULL, demi = NULL, nu = Inf,
  unpivot = FALSE)

`n`	Number of simulation.
`distr`	The marginal distribution. See vec2par.
`marg`	The parameter of the marginal distribution. Either a vector or a matrix with parameters in rows.
`cop`	A string representing the copula familly. Possible choice: Normal ('norm'), t-copula ('t'), Chi-square ('chisq'), squared t-copula ('tsq'), reverse chi-square ('rchisq') and reverse squared t-copula ('rtsq').
`h`	A matrix of distances. If `model == 'ind'` and `distr == 'unif'`, a number of sites must be provided instead.
`type`	A string representing a correlation model. See `corModel`
`p`	The parameters of the correlation model .
`sigma, demi`	Covariance matrix or its square root.
`nu`	Degree of freedom of a t-copula (if necessary).
`unpivot`	If `TRUE` the results is returns in the form of a table with columns : id, time, value.

The Chi-squared and the squared t-copula are the copulas of the squared variables coming from a multivariate Normal and Student distribution respectively. The "squared" copulas introduce radial asymmetry and their reverse have asymmetry in the opposite direction.

corModel

xx <- simAtSite(100, distr = 'gev', marg = c(100, 10, .1),
                cop = 'ind', h = 10, unpivot = TRUE)

coord <- replicate(2,runif(200,0,500))
h <- as.matrix(dist(coord))

# normal copula
xx <- simAtSite(1000, marg = c(100,20,-.1), h = h, p = c(400,0))

hist(xx[,2])

# t-copula

demi <- chol(corModel(c(400,0),h))
xx <- simAtSite(100, distr = 'unif',
                cop = 't', demi = demi, nu = 5)

hist(xx)