The screen that appears after clicking the ‘Site Analysis’ button within the ‘Data Screening’ window is called the ‘Site Analysis’ window. The ‘Site Analysis’ window is organized into four parts that allow the user to select a valid site (determined during data screening) and view pertinent statistics associated with that site, to display any of a variety of plots pertaining to the selected site or to all sites as a whole, to select one or more standard distributions to display on the site pdf or growth curve plots, and to view a table of exceedance intensities (frequencies) for various frequencies (intensities) using a user-selected distribution.
All sites that meet the user-supplied criteria in the ‘Data Screening’ window are listed within the site drop-box. When the user selects a site from the list, all associated statistics, plots, and exceedance table values are updated. All site statistics, which include the number of years used in the analysis/total number of years available, L-moments, and the discordance, are listed here for the site selected in the site drop-box. A site may also be deleted from the analysis; reasons for deleting a site include the site having a high discordance value, the site frequency distribution not fitting the same standard frequency distribution as a majority of the others sites, among other reasons.
L-moments include the L-mean (L1), L-Scale (L2), L-Coefficient of Variation (t1), L-Skew (t3), and L-Kurtosis (t4). These L-moments help to describe the location, dispersion, skewness, and peakedness, respectively, of a site’s data distribution. Discordancy represents a single statistic measured from the L-moments of a site’s data that determines whether that site is or is not an outlier. If the L-moments for that site are similar to the average L-moments for all sites in the study, then the discordancy will be low. If the L-moments for that site are significantly different, then the discordancy will be high and the site may be labeled as discordant. The critical value of discordancy that determines whether a site is considered discordant is dependent on the number of sites present in the study region.