Application of quality control charts for early detection of flood hazards

Abstract

Flooding is a serious and devastating hazard that many countries face regularly. Timely detection of these changes can help us manage rainwater and hence flooding. Numerous statistical tools are available for the early detection of these changes. In this regard, application of control charts is an effective choice toward monitoring natural events. In this study, we used self-proposed and existing control charts as a means of early detection of changes in the rainfall data of Pakistan as a case study. The proposed methodology covered two aspects: one to enhance the ability of Shewhart control charts toward detection of small or moderate changes in the behavior of natural events, and another to offer skewness correction based on individual control chart under runs rules for managing these natural events – especially when collected data follows unknown skewed distribution. Results elucidate that control charts structures were efficient in early detection and prediction of floods during 2010. Our results are in accordance with the theoretical results of existing studies, and we propose similar methods to be used for meteorological purposes.