A nonparametric approach to flood frequency analysis / Charles W. Labatiuk.
Material type:
TextSeries: Thesis. (M.A.Sc.)Publication details: Ottawa, ON : University of Ottawa, 1985.Description: xv, 187 leaves : ill. ; 28 cmOnline resources: Abstract: Parametric statistical techniques are generally used by hydrologists in flood frequency analysis. The assumption is made that the underlying distribution of the flood data is known, for example, it is a log-Pearson type III distribution. Such an assumption is not always justified, and can sometimes lead to considerable variability in the estimation of design floods. Nonparametric density estimation provides an alternative method of analysis which does not require any a priori distributional assumption. This thesis presents an application of nonparametric techniques to the computation of design floods. In particular, various methods of obtaining the smoothing factor required by the kernel estimator are investigated. Two sets of flood data are analyzed by both nonparametric and traditional parametric techniques. A simulation study numerically compares the relative merits and suitability of the nonparametric methods investigated. The principle finding is that design floods up to the Q100
| Item type | Current library | Collection | Call number | Status | Date due | Barcode |
|---|---|---|---|---|---|---|
Electronic Report
|
Electronic Library | Non-fiction | LAB (Browse shelf(Opens below)) | Available | 1520 |
Thesis(M.A.Sc.)--University of Ottawa, 1985.
Includes bibliographic references (leaves 148-154).
Parametric statistical techniques are generally used by hydrologists in flood frequency analysis. The assumption is made that the underlying distribution of the flood data is known, for example, it is a log-Pearson type III distribution. Such an assumption is not always justified, and can sometimes lead to considerable variability in the estimation of design floods. Nonparametric density estimation provides an alternative method of analysis which does not require any a priori distributional assumption. This thesis presents an application of nonparametric techniques to the computation of design floods. In particular, various methods of obtaining the smoothing factor required by the kernel estimator are investigated. Two sets of flood data are analyzed by both nonparametric and traditional parametric techniques. A simulation study numerically compares the relative merits and suitability of the nonparametric methods investigated. The principle finding is that design floods up to the Q100