Effect of lag time in rainfall-runoff modeling using ANFIS

• Main Author: Sim, Hui Ni.
• Other Authors: Chua Hock Chye Lloyd
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Civil engineering::Water resources
• Online Access: http://hdl.handle.net/10356/40131
• Institution: Nanyang Technological University

The effect of shifting lag time in forecasting rainfall runoff using the Artificial Neural Fuzzy Inference System (ANFIS) will be compared in this paper using a total of 63 rainfall events from the period of 16 Dec 2004 to 3 Nov 2006. The rainfall events were then grouped accordingly to their correlated rainfall antecedents. Out of the 63 rainfall events, 47 events which occurred in the most correlated rainfall antecedents were then further assembled into their respective training sets and training groups to be used for the ANFIS model. To determine the ANFIS model capabilities in modeling runoff forecasts, the Coefficient of Efficiency (CE) and Relative Peak Error (PE) were used as defining parameters to gauge the ANFIS model’s adequacy in predicting runoff discharge for Q(t+6), Q(t+8), and Q(t+10). The ANFIS model developed for this study made use of two rainfall inputs and one target rainfall output, which is the discharge forecast. A total of 78 rainfall inputs combinations were selected for the runoff forecasting of Q(t), whereas 12 rainfall inputs combinations will be used for Q(t+6), Q(t+8), and Q(t+10). From the analysis, it is shown that ANFIS is potentially qualified in modeling forecasts for up to Q(t+6) with generally good results in terms of CE and PE. However, the ANFIS model proved to be insufficient in discharge forecasting of Q(t+8) and Q(t+10).