The effect of serial correlation in environmental conditions on estimates of extreme events

Mackay, Ed; Hauteclocque, Guillaume de; Vanem, Erik; Jonathan, Philip

doi:10.1016/j.oceaneng.2021.110092

“…The extreme value analysis presented in this paper is a way of accounting for serial correlation by simulating from a time series model that preserves the marginal distribution and the autocorrelation in the observed data. It is demonstrated to give lower extreme value estimates than what one would get if serial correlation is ignored and hence to reduce the positive bias known to occur [11]. It also has an advantage compared to standard extreme value analysis techniques in that there is no need for sub-sampling and de-clustering, making it less wasteful.…”

Section: Discussionmentioning

confidence: 99%

“…Indeed, a second benchmark exercise was announced at OMAE 2021 calling for estimates of extreme environmental conditions based on specific datasets of significant wave height that were made available [9]. This is a follow-up of a previous benchmarking exercise that, inter alia, highlighted the effect of serial correlation on extreme value estimates [10], as elaborated further in [11].…”

Section: Introductionmentioning

confidence: 96%

Analyzing Extreme Sea State Conditions by Time-Series Simulation Accounting for Seasonality

Vanem

¹

2023

Journal of Offshore Mechanics and Arctic Engineering

Self Cite

View full text Add to dashboard Cite

This paper presents an extreme value analysis on data of significant wave height based on time-series simulation. A method to simulate time series with given marginal distribution and preserving the autocorrelation structure in the data is applied to significant wave height data. Then, extreme value analysis is performed by simulating from the fitted time-series model that preserves both the marginal probability distribution and the autocorrelation. In this way, the effect of serial correlation on the extreme values can be taken into account, without subsampling and de-clustering of the data. The effect of serial correlation on estimating extreme wave conditions have previously been highlighted, and failure to account for this effect will typically lead to an overestimation of extreme conditions. This is demonstrated by this study, that compares extreme value estimates from the simulated times-series model with estimates obtained directly from the marginal distribution assuming that 3-hourly significant wave heights are independent and identically distributed. A dataset of significant wave height provided as part of a second benchmark exercise on environmental extremes that was presented at OMAE 2021, has been analysed. This paper is an extension of a study presented at OMAE 2022 (OMAE2022-78795), and includes additional pre-processing of the data to account for seasonality and new results.

show abstract

“…The extreme value analysis presented above is a way of accounting for serial correlation by simulating from a time series model that preserves the marginal distribution and the autocorrelation in the observed data. It is demonstrated to give lower extreme value estimates than what one would get if serial correlation is ignored and hence to reduce the positive bias known to occur [11]. It also has an advantage compared to standard extreme value analysis techniques in that there is no need for subsampling and de-clustering, making it less wasteful.…”

Section: Discussionmentioning

confidence: 99%

“…Indeed, a second benchmark exercise was announced at OMAE 2021 calling for estimates of extreme environmental conditions based on specific datasets of significant wave height that were made available [9]. This is a follow-up of a previous benchmarking exercise that, inter alia, highlighted the effect of serial correlation on extreme value estimates [10], as elaborated further in [11]. This paper presents extreme value estimates of significant wave height from the datasets presented in the second benchmark exercise, based on a time-series model that preserves the marginal distribution and the autocorrelation of the data.…”

Section: Introductionmentioning

confidence: 96%

“…This gives an empirical distribution of the maximum H S in a T -year period, and the corresponding T -year return value can be estimated as the 1/e-quantile of this distribution (see e.g. [9,11]). However, the approach presented in this paper gives an estimate of the full T -year maximum distribution and other measures such as the mean and standard deviation of the T -year maximum can also be estimated.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Analyzing Extreme Sea State Conditions by Time-Series Simulation

Vanem

¹

2022

Volume 2: Structures, Safety, and Reliability

View full text Add to dashboard Cite

This paper presents an extreme value analysis on data of significant wave height based on time-series simulation. A method to simulate time series with given marginal distribution and preserving the autocorrelation structure in the data is applied to significant wave height data. Then, extreme value analysis is performed by simulating from the fitted time-series model that preserves both the marginal probability distribution and the autocorrelation. In this way, the effect of serial correlation on the extreme values can be taken into account, without subsampling and de-clustering of the data. The effect of serial correlation on estimating extreme wave conditions have previously been highlighted, and failure to account for this effect will typically lead to an overestimation of extreme conditions. This is demonstrated by this study, that compares extreme value estimates from the simulated times-series model with estimates obtained directly from the marginal distribution assuming that 3-hourly significant wave heights are independent and identically distributed. A dataset of significant wave height provided as part of a second benchmark exercise on environmental extremes that was presented at OMAE 2021, has been analysed.

show abstract