Estimations of the extreme flow distributions by stochastic models (Q1966376): Difference between revisions
From MaRDI portal
Created a new Item |
Added link to MaRDI item. |
||
links / mardi / name | links / mardi / name | ||
Revision as of 17:37, 1 February 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Estimations of the extreme flow distributions by stochastic models |
scientific article |
Statements
Estimations of the extreme flow distributions by stochastic models (English)
0 references
1 March 2000
0 references
This paper deals with the problem of estimation of extreme value flood peaks or the risk of exceeding which is an important task in hydrological designs. The authors focused on following stochastic approaches used for this purpose: the annual maximum series (AMS) method which takes into account only annual maximum floods; the partial duration series method (PDH) which considers all floods above a given threshold; and the multi-maximum peak model which uses two or more annual maximums (MAMS). Several models within these approaches are considered. In these models the magnitude of the flood peak is a random variable with density function \(f(x,\theta)\) and distribution function \(F(x,\theta)\) independent of the time interval, where \(\theta\) is a parameter vector (dependent on the threshold); and the number of flood peaks \(N(t)\) in the time interval \([0,t)\) is a homogeneous Poisson process with intensity \(\lambda\). The joint distributions of two or more maximum extremes from historical records in the proposed models are derived; the explicit formulas include \(\lambda\), \(f(\cdot)\) and \(F(\cdot)\). Thus, statistical procedures for estimating \(\lambda\) and \(F\) are discussed and a list of distributions appropriate for modelling the flood magnitude is presented. Also a multi-variate shock model is proposed to estimate the distribution of the extremes for a multi-model stream flow. To investigate the performance of the estimators, based on the AMS, MAMS and PDS approaches, a statistical analysis of both real and simulated data was carried out. The results show that all the proposed stochastic models and related methods are reliable and useful. Multi-maximum models have some advantages and multi-shock models need further investigations.
0 references
distribution of extremes
0 references
tail distributions
0 references
Poisson processes
0 references
annual maximum series method
0 references
partial duration series method
0 references
multi-maximum peak model
0 references
multivariate-shock model
0 references