covariog.model.env: Envelope for Empirical Covariogram for the Poisson-log normal model

Description

Computes envelope for empirical covariogram by simulating data for given model parameters. This function is for the Poisson-log normal model.

Usage

covariog.model.env(geodata, coords = geodata$coords, units.m = "default",
     obj.covariog, model.pars, nsim = 500, prob = c(0.025, 0.975),
     messages)

Arguments

geodata

a list containing element coords as described next. Typically an object of the class "geodata" - a geoR data set. If not provided the argument coords must be given instead. The list may also contain an argument units.m as described below.

coords

an $n \times 2$ matrix, each row containing Euclidean coordinates of the $n$ data locations. By default it takes the element coords of the argument geodata.

units.m

$n$-dimensional vector of observation times for the data. By default (units.m = "default"), it takes geodata$units.m in case this exist and else the value 1 for every observation.

obj.covariog

an object of the class "covariogram", typically an output of the function covariog.

model.pars

a list with model specification and parameter values. The required components of the list are:

beta, the mean parameter. Default is beta=0.
cov.model, the covariance model. Default is "exponential".
cov.pars, the covariance parameters $\sigma^2$ and $\phi$.
kappa, the extra covariance parameters for some of the covariance models. Default is kappa = 0.5.
nugget, the relative nugget variance. Default is nugget = 0.

nsim

number of simulations used to compute the envelope. Default is nsim = 500.

prob

the quantiles used for constructing the envelopes. Default is 2.5% and 97.5%

messages

logical. If TRUE, the default, status messages are printed while the function is running.

Value

An object of the class "covariogram.envelope" which is a list with the components:

a vector with distances.

v.lower

a vector with the upper-quantile covariogram values for each distance in u.

v.upper

a vector with the lower-quantile covariogram values for each distance in u.

Details

The envelope is computed assuming a Poisson-log normal model. Simulated values are generated at the data locations, given the model parameters. The empirical covariogram is computed for each simulation using the same binning as for the original covariogram of the data. The envelope is computed by taking, at each lag, the quantile-values of the covariograms for the simulated data.

Examples

Run this code

# NOT RUN {
data(p50)
covar <- covariog(p50, uvec = c(1:10))
parmval <- list(cov.model = "exponential", cov.pars = c(0.8,0.1), 
                beta = 1)
class(parmval) <- "covariomodel"
konvol <- covariog.model.env(p50, obj.covariog = covar, 
                             model.pars = parmval)
plot(covar, envelope.obj = konvol)
lines(parmval, max.dist = 10, lty = 1)
# }