Heterogeneity measure (H2) for pooling groups.

Quantifies the heterogeneity of a pooled group

Usage

H2(x, H1 = FALSE)

Arguments

x

pooling group derived from the Pool() function

H1

logical with a default of FALSE. If TRUE, the function applies the 'H1' version of the test (see Hosking & Wallis 1997 reference). If FALSE, the default H2 version is applied.

Value

A vector of two characters; the first representing the H2 score and the second stating a qualitative measure of heterogeneity.

Details

The H2 measure was developed by Hosking & Wallis and can be found in their book 'Regional Frequency Analysis: An Approach Based on L-Moments' (1997). It was also adopted for use by the Flood Estimation Handbook (1999) and is described in volume 3. It works by recreating 500 pooling groups with the same sample sizes, assuming a four parameter Kappa distribution (parameters from the pooled L-moments). L-moment ratios are calculated for each of the 500 simulated pooling groups. The heterogeneity is determined by comparing the variance of L-moment ratios in the observed pooling group with the variance of the L-moment ratios across the simulated pooling groups. The simulations are homogeneous, therefore if the observed pooling group is homogeneous the expectation is that the variance will be similar to the average of the simulated variance.

Author

Anthony Hammond

Examples

# Get CDs, form a pooling group, and calculate H2
cds_203018 <- GetCDs(203018)
pool_203018 <- Pool(cds_203018)
H2(pool_203018)
#> [1] "7.01"                                                              
#> [2] "Group is strongly heterogenous: a review of the group is essential"