Computes the unbiased sample L-moments and L-moment ratios of multiple sets of data stored in a list or matrix. Following the paradigm of regional frequency analysis, we regard the data sets as coming from different measurement sites.
regsamlmu(x, nmom =5, sort.data =TRUE, lcv =TRUE)
Arguments
x: A list of numeric vectors, or a numeric matrix.
nmom: Number of L-moments to be computed.
sort.data: Logical: whether each data set should be sorted.
lcv: Logical. If TRUE, the second L-moment will be expressed as a fraction of the mean, i.e. the computed value will be the sample L-CV t=l2/l1. If FALSE, the second L-moment will simply be the sample L-scale value l2.
Details
Sample L-moments are computed for each data set. The calculations use samlmu internally. If x is a list, each list element should contain data for one site and the names of the list elements should be the site names. If x is a matrix, each column should contain data for one site and the column names should be the site names.
Returns
An object of class regdata. It is a data frame with columns "name" and "n", containing respectively the site names and the number of non-missing data values at each site, and further columns containing the L-moments and L-moment ratios, in the order l1, t (or l2), t3, t4, etc.
Note
The default parameter values are chosen to be convenient for the regional frequency analysis methods described by Hosking and Wallis (1997). Note that the number of L-moments and the choice of whether to return L-CV or L-scale are different from the defaults for samlmu.
Users of the LMOMENTS Fortran package, version 3.04, should note that its PROGRAM XFIT by default uses plotting-position estimators of L-moment ratios, which give different results from the unbiased estimators used by regsamlmu
(and by all other functions in package lmomRFA).
References
Hosking, J. R. M., and Wallis, J. R. (1997). Regional frequency analysis: an approach based on L-moments. Cambridge University Press.