GENERAL PERSPECTIVES ON APPLICATION OF DECISION SUPPORT SYSTEM (DSS) FOR METROPOLITAN INTELLIGENT FLOOD MANAGEMENT IN NIGERIA

Abstract

Climate change has had a great impact on man and his environment. In recent years, climate change has aggravated levels of extreme events such as floods, drought, heat waves, cyclones, hurricanes, tornadoes and wildfires. Though the nature and extent of this impact vary across the globe, no region is immune from the effect of the change. Attempts have been made over the years to address or mitigate the effects. To this end, flood defense has been approached in terms of preparedness, mitigation, response and adaptation to various degrees of success. The measures taken are not only limited to structural but also institutional (knowledge-based and collaboration). The failure of some of these methods is highly attributed to the stochastic nature of the flood phenomenon. To address this, different methods have been adopted to address the menace, among which is the Decision Support System (DSS). The application of DSS entails the decomposition of the distinct the nature of flood phenomenon. In view of this, an attempt was made in this paper to put the application of DSS in perspective considering its general application and robustness for flood study since it is derived from the concept of Artificial Intelligence (AI) System. Generalization of the application of DSS is founded on the concept of intuitive knowledge-based science and engineering, hence its potential for real-time application. The problem associated with the DSS is entrenched in the difficulties in quantifying data, network failure, collection of required data, lack of technological knowledge by the user, information overload, subjectivity, cost of development as well as fear of implementation of new technology by users. Thus, there is a need for a new approach based on AI-DSS that encompasses a language system (problem processing and knowledge base), a functional approach which includes the data, a model dialogue module and a tool-based approach. This basically requires modelling by taking into account the non-crispy nature of the variables.