Frequency analysis

Chapter 6 introduces basic probability concepts and presents the different steps involved in frequency analysis, with focus on parametric methods. Special attention is given to extreme value analysis of time series of low flow and drought characteristics, such as duration and deficit volume. It discusses how to select data for extreme value analysis, including how to deal with seasonality, non-stationarity, zero values and outliers. The selection of extreme events is related to two limiting distributions, the Generalised Extreme Value (GEV) distribution for Annual Maximum/Minimum Series (AMS) and the Generalised Pareto (GP) distribution for Partial Duration Series (PDS). Emphasis is on their applicability for extreme value analysis of time series of minimum values (low flow) and maximum values (deficit characteristics). Several distribution functions are presented along with their applicability for the data at hand. Both at-site and regional frequency analysis are covered, and the procedures are demonstrated using time series from the example datasets.