Hourly runoff forecasting for flood risk management: Application of various computational intelligence models
MetadataShow full item record
© 2015 Elsevier B.V. Reliable river flow forecasts play a key role in flood risk mitigation. Among different approaches of river flow forecasting, data driven approaches have become increasingly popular in recent years due to their minimum information requirements and ability to simulate nonlinear and non-stationary characteristics of hydrological processes. In this study, attempts are made to apply four different types of data driven approaches, namely traditional artificial neural networks (ANN), adaptive neuro-fuzzy inference systems (ANFIS), wavelet neural networks (WNN), and, hybrid ANFIS with multi resolution analysis using wavelets (WNF). Developed models applied for real time flood forecasting at Casino station on Richmond River, Australia which is highly prone to flooding. Hourly rainfall and runoff data were used to drive the models which have been used for forecasting with 1, 6, 12, 24, 36 and 48. h lead-time. The performance of models further improved by adding an upstream river flow data (Wiangaree station), as another effective input. All models perform satisfactorily up to 12. h lead-time. However, the hybrid wavelet-based models significantly outperforming the ANFIS and ANN models in the longer lead-time forecasting. The results confirm the robustness of the proposed structure of the hybrid models for real time runoff forecasting in the study area.
Showing items related by title, author, creator and subject.
LUCICAT Model as a river flow forecasting tool: an experiment with Fitzroy River catchment of Western AustraliaIslam, S.A.; Anwar, Faisal; Ezzy, G.; Bari, M. (2013)Early warnings of river flow, particularly high flow, allow individuals, communities and industries to respond in an appropriate manner to reduce the probability of suffering, personal injury, death and economic loss. To ...
Sarukkalige, Priyantha Ranjan; Badrzadeh, Honey (2014)Accurate river flow forecasts play a key role in sustainable water resources and environmental management. Recently, computational intelligence approaches have become increasingly popular due to minimum information ...
Impact of multi-resolution analysis of artificial intelligence models inputs on multi-step ahead river flow forecastingBadrzadeh, Honey; Sarukkalige, Priyantha Ranjan; Jayawardena, A. (2013)In this paper an attempt is made to show that the performance of daily river flow forecasting is improved when data-preprocessing techniques are used with computational intelligence methods. Especially for forecasting ...