- Author
- Lechner, J. A. | Simiu, E. | Heckert, N. A.
- Title
- Assessment of 'Peaks Over Threshold' Methods for Estimating Extreme Value Distribution Tails.
- Coporate
- National Institute of Standards and Technology, Gaithersburg, MD
- Journal
- Structural Safety, Vol. 12, 305-314, 1993
- Keywords
- distribution tails | extreme values | peaks over threshold methods | wind engineering
- Abstract
- In the past twenty years a vast new body of extreme value theory was developed, referred to as 'peaks over threshold modeling'. This theory allows the use in the analysis of all data exceeding a sufficiently high threshold, a feature that may result in improved extreme value estimates. The application of the theory depends upon the performance of methods for estimating the distribution parameters corresponding to any given set of extreme data. We present a comparative assessment of the performance of three such methods. The assessment is based on Monte Carlo simulations from populations with four distributions: Gumbel, Weibull, generalized Pareto, and normal. The simulation results showed that the de Haan and the Conditional Mean Exceedance (CME) methods performed consistently better than the Pickands method (NIST implementation). For the distributions, parameter values, and mean recurrence intervals assumed in this work, the CME method outperformed the de Haan method only when the percent estimation errors were about one percent or smaller, a case unlikely to be encountered in wind engineering practice.