A Comparison Of Plotting Formulas For The Pearson Type Iii Distribution

Abstract

Unbiased plotting position formulas are discussed to fit the Pearson Type 3 distribution (PIII). The best quantile estimate made from the plotting position should be unbiased and should have the smallest root means square error among all such estimates. Probability plot correlation coefficient (PPCC) is used to evaluate goodness of fit to test the PIII distribution hypothesis. Results obtained using the annual maximum flow data from Peninsular of Malaysia based on PPCC show the plotting position formulas consistently produced linear probability plots with correlation coefficient near to one. Based on root mean square error (RMSE) and root mean absolute error, the Weibull formula performs better than the other formulas.