Development of a kriging based regional flood frequency analysis technique in Australia

Abstract

Flood is a natural disaster that causes widespread damage and inconveniences. Flood-related economic losses and number of fatalities have increased dramatically over the past decades around the globe. Flood damage can be minimised by ensuring optimum design of drainage infrastructure, which depends largely on reliable estimation of design floods or flood quantiles. Design flood is defined as a flood discharge associated with a given annual exceedance probability (AEP) or average recurrence interval (ARI). This thesis focuses on design flood estimation problem in ungauged catchments using a geostatistical based regional flood frequency analysis (RFFA) approach. The geo-statistical approach has been applied in RFFA in a number of countries due to its ability of exploiting directly the spatial correlations of the observed flood data either in the common geographical space or along the stream network. One of the simplest and most widely applied geo-statistical models in hydrology is ordinary kriging, which is capable of combining flood quantile at multiple sites for making regional prediction. In this thesis, ordinary kriging based RFFA models are developed and tested for Australia. No previous study in Australia focused on kriging based RFFA model development and testing using a comprehensive flood database. A total of 586 catchments have been selected from eastern Australia (Queensland (QLD), New South Wales (NSW), Victoria (VIC)) and South Australia (SA) as the study catchments. The reason for selecting these states is that there are a good number of high quality stream gauging stations in these states that can be used in the development and testing of a kriging based RFFA model. The research and development here are centred on flood quantile estimation in ungauged catchments in the range of frequent to rare flood quantiles (2, 5, 10, 20, 50 and 100 years of ARIs).

Date of Award	2020
Original language	English