Abstract
In this paper, the Salt River basin possessing paleoflood study has been selected for flood estimations. By plotting the flood data on log-normal probability paper, it is found that the trend of high floods, whose peaks are greater than 2,400 m3/s, deviates from the straight line fitted to the gauged flood data and the trend of these high floods is similar to that of the paleofloods. Thus, it is reasonable to assume that these high floods belong to different probability distributions than that of the rest. The distributions for the high values of gauged flood data and the paleoflood data are assessed by linear regression using their plotting positions on probability paper. Then, a Bayesian statistical approach is applied to combine the gauged flood data and paleoflood data. The floods estimated by the combined distribution are much lower than those obtained by the conventional methods. For example, the 100 year flood estimated by the combined distribution is about 60% of that obtained by a Pearson type III distribution.
| Original language | English (US) |
|---|---|
| Pages | 425-430 |
| Number of pages | 6 |
| State | Published - 1998 |
| Event | Proceedings of the 1998 25th Annual Conference on Water Resources Planning and Management - Chicago, IL, USA Duration: Jun 7 1998 → Jun 10 1998 |
Other
| Other | Proceedings of the 1998 25th Annual Conference on Water Resources Planning and Management |
|---|---|
| City | Chicago, IL, USA |
| Period | 6/7/98 → 6/10/98 |
ASJC Scopus subject areas
- Civil and Structural Engineering
- Water Science and Technology