Thesis
Regional frequency analysis and copula multivariate statistics for characterizing extreme precipitation and drought
Washington State University
Master of Science (MS), Washington State University
2015
Handle:
https://hdl.handle.net/2376/100570
Abstract
This study coupled regional frequency approach with Bayesian uncertainty quantification to produce more reliable and updated estimate of extreme precipitation patterns for Bangladesh using daily precipitation data from 1961 to 2010. Regional frequency analysis based on L-moments statistics was used on 1-day, 2-day and 5-day annual maximum precipitation series due to its advantages over the commonly applied at-site estimation. Five 3-parameter distribution models (i.e., Generalized Logistic, Generalized Extreme value, Generalized Normal, Pearson Type Three, and Generalized Pareto) were used for a thorough selection of appropriate models that fit the sample data. Uncertainties related to the selection of the models and estimation of their parameters were quantified using the Bayesian approaches. The study also aims at probabilistic evaluation of bivariate drought characteristics to assess drought duration and severity in Bangladesh. The procedures involve: 1) Applying the Standardized precipitation index (SPI) to identify duration and severity of droughts based on precipitation data from 1961-2010; 2) Applying the L-moment based regional frequency analysis along with region of influence approach to fit marginal distributions to both duration and severity. Pearson type III and Wakeby have been selected as best fit marginal distribution for drought severity and duration respectively. Since drought duration and severity are mutually correlated measured by kendall’s tau and they were modeled using different distributions, copulas were used to model their dependence structure. The joint and conditional probabilities and associated recurrence intervals of drought events were derived from the copula based joint distribution. The results from this study is expected to be used for updating the current design and management of hydraulic structures as well as for effective drought management strategies in Bangladesh.
Metrics
18 File views/ downloads
33 Record Views
Details
- Title
- Regional frequency analysis and copula multivariate statistics for characterizing extreme precipitation and drought
- Creators
- Md Rubayet Mortuza
- Contributors
- Yonas K. Demissie (Degree Supervisor)
- Awarding Institution
- Washington State University
- Academic Unit
- Civil and Environmental Engineering, Department of
- Theses and Dissertations
- Master of Science (MS), Washington State University
- Publisher
- Washington State University; [Pullman, Washington] :
- Identifiers
- 99900525139401842
- Language
- English
- Resource Type
- Thesis