A prominent element within this program has been the development of operational flood forecasting systems. These systems have evolved through advances in mathematical modeling (Wood and O’Connell, 1985; O’Connell, 1991; Lambert’ and Pilate, 1996), the installation of telemetry and field monitoring equipment at critical sites in drainage networks (Alexander, 1991), through satellite and radar sensing of extreme rainfalls (Collier, 1991), and through the coupling of precipitation and runoff models (Grosbeaks and Foulard-Georgia, 1991; Franchise et al. , 1996).
However, in practice, successful real- time flood forecasting often depends on the efficient integration of all these separate activities (Douglas and Dobson, 1987). Under the auspices of the World Meteorological Organization (1992) a series of projects were implemented to compare ten counterblasts Ana performance AT various operational models Ana t updating procedures. A major conclusion of the most recent inter-comparison exercise was the need for robust simulation models in order to achieve consistently teeter results for longer lead times even when accompanied by an efficient updating procedure.
The attractiveness of Artificial Neural Networks (Ann.’s) to flood forecasting is threefold. First, Ann.’s can represent any arbitrary non-linear function given sufficient complexity of the trained network. Second, Ann.’s can find relationships between different input samples and, if necessary, can group samples in analogous fashion to cluster analysis. Finally, and perhaps most importantly, Ann.’s are able to generalize a relationship from small subsets of data while remaining elatedly robust in the presence of noisy or missing inputs, and can adapt or learn in response to changing environments.
However, despite these potential advantages, Ann.’s have found rather limited application in hydrology and related disciplines. For example, French et al. (1992) used a neural network to forecast rainfall intensity fields in space and time, while Raman and Summarily (1995) used an ANN to synthesize reservoir inflow series for two sites in the Breakthrough basin, S. Linda. The use of artificial neural networks for flood forecasting is an area which has yet to be fully explored.
Up until now the majority of work in this area has been mainly theoretical; concentrating on neural network performance with artificially generated rainfall- runoff data; for example Minis and Hall (1996). However, these theoretical approaches tend to overlook the difficulty in converting and applying actual data to artificial neural network topologies. Hall and Minis (1993) go some way to address this criticism by applying neural networks to a small urban catchments area. However, their discussion is limited to the performance of a neural network on a small number of events. Problem Statement
Research question Does artificial intelligence provide an optimum solution to the prediction of rainfall- runoff? Aim of the research The aim of this project is to evaluate and develop artificial intelligence technique for modeling weather patterns. Research Objectives To identify and evaluate existing weather pattern modeling techniques. To identify artificial intelligence approaches suitable for modeling weather patterns. To design and develop an artificial intelligence system for modeling weather patterns. To compare and contrast the developed system with existing approaches.
Research methods First, the researcher will conduct a summation evaluation on techniques for developing ANN in the domain of hydrology. Secondly, an algorithm for the ANN slung ten prototyping process model will De developed Prototyping allows ten system developer/researcher to explore the requirements and technical feasibility of the system with the user. The researcher will then carry out laboratory experiments at the Ghana Meteorological Services Department to identify the relationships between weather patterns and rainfall-runoff using simulations.
The Ghana Meteorological Services Department is relevant in that it is directly involved in the scientific study of the earth’s atmosphere, especially, its patterns of climate and weather. Laboratory experiments permit the researcher to identify precise relationships between a small number of variables that are studied intensively via a designed laboratory situation using quantitative analytical techniques with a view to making generalize statements applicable to real-life situations.
Simulation involves copying the behavior of a system. Simulation is used in situations where it would be official normally to solve problems analytically and typically involves the introduction of random variables. Forecasting techniques such as regression analysis and time series analysis (quantitative techniques) will be used to predict future rainfall-runoff. Forecasting involves the use of techniques such as regression analysis and time series analysis to make predictions about likely future events.
It is a useful form of research method in that it attempts to cope with the rapid changes that are taking place in IT and predict the impacts of these changes on individuals, organizations or society. Interviews will be conducted at the premises of the Ghana Meteorological Services Department with the view to illicit system requirements for the development of the ANN. Responses will be recorded for later transcription and analysis. Data analysis will use the SPAS software.
Secondary data will be collected from the same location for simulation. Expected Outcomes/deliverables An artificial intelligence system for predicting flood, as well as a report detailing the following: I. A discussion of the problems that need to be addressed when applying neural networks to rainfall-runoff modeling I’. An explanation of the effectiveness of artificial neural networks in the domain of hydrology. Iii. An evaluation of the techniques and methods for developing Ann.’s in the domain of hydrology lb.
An evaluation of the prospects for the use of the developed ANN in real-time flood forecasting by applying the developed system to flood simulation in two catchments in Ghana. V. A discussion and evaluation of alternative tools and methods for weather pattern modeling, including suggestions concerning necessary refinements to the developed ANN prior to transfer to operational use.
