- 97 views
- 168 downloads
River Ice Breakup Forecasting with Fuzzy and Neuro-fuzzy Models
-
- Author / Creator
- Mahabir, Chandra Lilian
-
Spring river ice breakup on northern rivers can quickly result in ice jams which present severe flood risk to which have little or no advanced warning. Despite the serious threat posed, there are no reliable means to predict the severity of breakup with a significant lead time. Many of the previous studies regarding ice jam flood forecasting methods cite the lack of a comprehensive database as an obstacle to modeling. The ability to transfer a model between river basins is highly desirable but has not previously been achieved due to site specific nature of most river breakup models.
As a foundation, this thesis documents the development of an extensive database containing 106 variables, and covering the period from 1972 to 2004, that was created for ice jam forecasting on the Athabasca River for the community of Fort McMurray, Alberta. The number of historical years of data, rather than the scope of variables was found to be the major limitation for ice modeling at this site.
The potential for short and long time predictive models was evaluated. A short lead time model was achieved though multiple linear regression analysis, equations were developed to model the maximum water level. The optimal model contained a combination of hydrological and meteorological data collected from early fall until the day before river ice breakup. Soft computing including fuzzy logic and artificial neural networks was used to model the maximum water level. It was found that a simple fuzzy expert system based exclusively on expert experience could qualitatively distinguish years when flooding occurred but produced poor quantitative results. A neuro-fuzzy model with fewer variables was able to simulate water levels equally as well as a multiple linear regression model with fewer input variables which provided a longer lead time.
Basin transferability was evaluated at Hay River in northern Canada. Qualitative results showed that the fuzzy model was transferred between basins because extreme events could be distinguished from years when flooding did not occur. The high quantitative accuracy of the neuro-fuzzy model was not reproduced. Climate change scenarios for the Athabasca River indicated a continuously decreasing risk of severe ice jams while the frequency in the Hay River Basin increased for a period before waning. -
- Graduation date
- Spring 2007
-
- Type of Item
- Thesis
-
- Degree
- Doctor of Philosophy
-
- License
- This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.