Comparison of precipitation extremes estimation using parametric and nonparametric methods

Holešovský, Jan; Fusek, Michal; Blachut, Vít; Michálek, Jaroslav

doi:10.1080/02626667.2015.1111517

Cited by 10 publications

(6 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Parametric resampling randomly generates samples from a parametric model (distribution) fitted to the data, and is the preferred bootstrap method for extreme value analysis [33,34]. The advantage of parametric resampling is that data can be generated beyond the values already observed, and the method is rather computationally undemanding; however, it only considers the uncertainty associated with the estimation of the GEV parameters and sampling errors [35].…”

Section: Modified Maximum Likelihood Estimationmentioning

confidence: 99%

Estimating Rainfall Design Values for the City of Oslo, Norway—Comparison of Methods and Quantification of Uncertainty

Lutz

Grinde

Dyrrdal

2020

Water

View full text Add to dashboard Cite

Due to its location, its old sewage system, and the channelling of rivers, Oslo is highly exposed to urban flooding. Thus, it is crucial to provide relevant and reliable information on extreme precipitation in the planning and design of infrastructure. Intensity-Duration-Frequency (IDF) curves are a frequently used tool for that purpose. However, the computational method for IDF curves in Norway was established over 45 years ago, and has not been further developed since. In our study, we show that the current method of fitting a Gumbel distribution to the highest precipitation events is not able to reflect the return values for the long return periods. Instead, we introduce the fitting of a Generalised Extreme Value (GEV) distribution for annual maximum precipitation in two different ways, using (a) a modified Maximum Likelihood estimation and (b) Bayesian inference. The comparison of the two methods for 14 stations in and around Oslo reveals that the estimated median return values are very similar, but the Bayesian method provides upper credible interval boundaries that are considerably higher. Two different goodness-of-fit tests favour the Bayesian method; thus, we suggest using the Bayesian inference for estimating IDF curves for the Oslo area.

show abstract

Section: Modified Maximum Likelihood Estimationmentioning

confidence: 99%

Estimating Rainfall Design Values for the City of Oslo, Norway—Comparison of Methods and Quantification of Uncertainty

Lutz

Grinde

Dyrrdal

2020

Water

View full text Add to dashboard Cite

show abstract

“…This is directly related to POT method applied in [10], whereby X n−k,n plays a role of a threshold (for POT see e.g. [11]). From relation (2) arise two main types of semiparametric endpoint estimators.…”

Section: Endpoint Estimators and Their Propertiesmentioning

confidence: 99%

“…For this purpose was used either the maximum likelihood estimator (suitable for γ > − 1 2 , see [1,17]), or another moment estimator of γ (little different from γ MOM ) [1], whereby the optimal k 0 was chosen via double bootstrap methodology (see e.g. [4,11]). Hence, we consider the following models:…”

Section: Simulation Studymentioning

confidence: 99%

“…In practical situations is the optimal k 0 being determined mostly by visual inspection of the parameter dependencies (see e.g. [11] for so-called mean-residual-life plot in POT analysis), or by several rules of thumb (as in [10] where they choose k such that X n−k,n is equivalent to the upper 90% quantile). Lately, attention was paid to adaptive methods for selection of k 0 , follow e.g.…”

Section: Simulation Studymentioning

confidence: 99%

See 1 more Smart Citation

Distribution Endpoint Estimation Assessment for the use in Metaheuristic Optimization Procedure

Holesovky¹

2018

mendel

View full text Add to dashboard Cite

Metaheuristic algorithms are often applied to numerous optimization problems, involving large-scale and mixed-integer instances, specifically. In this contribution we discuss some refinements from the extreme value theory to the lately proposed modification of partition-based random search. The partition-based approach performs iterative random sampling at given feasible subspaces in order to exclude the less favourable regions. The quality of particular regions is evaluated according to the promising index of a region. From statistical perspective, determining the promising index is equivalent to the endpoint estimation of a probability distribution induced by the objective function at the sampling subspace. In the following paper, we give a short review of the recent endpoint estimators derived on the basis of extreme value theory, and compare them by simulations. We discuss also the difficulties in their application and suitability of the estimators for various optimization instances.

show abstract

“…These methods have a problem with the correct choice of threshold, as described in detail e.g. in Holešovský, Fusek and Michálek (2015). In case of very short time series of observed values it is preferred to use the POT method supplemented by the bootstrapping methods (described e.g.…”

Section: Introductionmentioning

confidence: 99%

Determination of N-year Design Precipitation in the Czech Republic by Annual Maximum Series Method

Kozlovská¹,

Šácha²,

Toman³

2019

Acta Univ. Agric. Silvic. Mendelianae Brun.

View full text Add to dashboard Cite

The sum of design precipitation of a selected repetition period, provided that it is evenly distributed over the river basin area, is a basic input for the calculation of the direct outflow volume by the curve number method. It is necessary to determine the design precipitation for each location using the statistical methods and the longest available data series on daily precipitation sums, or more specifically their annual maximums. This paper deals with the determination of design precipitation from data of eight stations of the Czech Hydrometeorological Institute for the period 1961 - 2013. From a series of annual maximum values of daily precipitation sums, N-year design precipitations were calculated using two methods (Gumbel and generalized extreme value distributions). The conformity of both models with empirical distribution of values was statistically tested to evaluate which of the models gave more accurate results. In these cases, it was more appropriate to use the generalized extreme value distribution. Finally, the newly calculated characteristics were compared with the design values used by Šamaj et al. (1985), where significant differences were found.

show abstract

Comparison of precipitation extremes estimation using parametric and nonparametric methods

Cited by 10 publications

References 38 publications

Estimating Rainfall Design Values for the City of Oslo, Norway—Comparison of Methods and Quantification of Uncertainty

Estimating Rainfall Design Values for the City of Oslo, Norway—Comparison of Methods and Quantification of Uncertainty

Distribution Endpoint Estimation Assessment for the use in Metaheuristic Optimization Procedure

Determination of N-year Design Precipitation in the Czech Republic by Annual Maximum Series Method

Contact Info

Product

Resources

About