• Journal of Korea Water Resources Association
• /
• v.50 no.4
• /
• pp.253-261
• /
• 2017

A lot of nonstationary frequency analyses have been studied in recent years as the nonstationarity occurs in hydrologic time series data. In nonstationary frequency analysis, various forms of probability distributions have been proposed to consider the time-dependent statistical characteristics of nonstationary data, and various methods for parameter estimation also have been studied. In this study, we aim to introduce a parameter estimation method for nonstationary Gumbel distribution using ensemble empirical mode decomposition (EEMD); and to compare the results with the method of maximum likelihood. Annual maximum rainfall data with a trend observed by Korea Meteorological Administration (KMA) was applied. As a result, both EEMD and the method of maximum likelihood selected an appropriate nonstationary Gumbel distribution for linear trend data, while the EEMD selected more appropriate nonstationary Gumbel distribution than the method of maximum likelihood for quadratic trend data.

• Journal of Korea Water Resources Association
• /
• v.48 no.5
• /
• pp.331-343
• /
• 2015

Most nonstationary frequency models are defined as the probability models containing the time-dependent parameters. For frequency analysis of annual maximum rainfall data, the Gumbel distribution is generally recommended in Korea. For the nonstationary Gumbel models, the time-dependent location and scale parameters are defined as linear and exponential relationship, respectively. The exponentially time-varying scale parameter of nonstationary Gumbel model is generally used because the scale parameter should be positive. However, the exponential form of scale parameter occasionally provides overestimated quantiles. In this study, various forms of time-varying scale parameters such as exponential, linear, and logarithmic forms were proposed and compared. The parameters were estimated based on the method of maximum likelihood. To compare the accuracy of each scale parameter, Monte Carlo simulation was performed for various conditions. Additionally, nonstationary frequency analysis was conducted for the sites which have more than 30 years data with a trend in rainfall data. As a result, nonstationary Gumbel model with exponentially time-varying scale parameter generally has the smallest root mean square error comparing with another forms.

• Journal of Korea Water Resources Association
• /
• v.52 no.11
• /
• pp.895-904
• /
• 2019

It has been well recognized that extreme rainfall process often features a nonstationary behavior, which may not be effectively modeled within a stationary frequency modeling framework. Moreover, extreme rainfall events are often described by a two (or more)-component mixture distribution which can be attributed to the distinct rainfall patterns associated with summer monsoons and tropical cyclones. In this perspective, this study explores a Mixture Distribution based Nonstationary Frequency (MDNF) model in a changing rainfall patterns within a Bayesian framework. Subsequently, the MDNF model can effectively account for the time-varying moments (e.g. location parameter) of the Gumbel distribution in a two (or more)-component mixture distribution. The performance of the MDNF model was evaluated by various statistical measures, compared with frequency model based on both stationary and nonstationary mixture distributions. A comparison of the results highlighted that the MDNF model substantially improved the overall performance, confirming the assumption that the extreme rainfall patterns might have a distinct nonstationarity.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.30 no.4B
• /
• pp.389-397
• /
• 2010

This study introduced a Bayesian based frequency analysis in which the statistical trend analysis for hydrologic extreme series is incorporated. The proposed model employed Gumbel extreme distribution to characterize extreme events and a fully coupled bayesian frequency model was finally utilized to estimate design rainfalls in Seoul. Posterior distributions of the model parameters in both Gumbel distribution and trend analysis were updated through Markov Chain Monte Carlo Simulation mainly utilizing Gibbs sampler. This study proposed a way to make use of nonstationary frequency model for dynamic risk analysis, and showed an increase of hydrologic risk with time varying probability density functions. The proposed study showed advantage in assessing statistical significance of parameters associated with trend analysis through statistical inference utilizing derived posterior distributions.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2010.05a
• /
• pp.581-585
• /
• 2010

This study introduced a Bayesian based frequency analysis in which the statistical trend seasonal analysis for hydrologic extreme series is incorporated. The proposed model employed Gumbel and GEV extreme distribution to characterize extreme events and a fully coupled bayesian frequency model was finally utilized to estimate design rainfalls in Seoul. Posterior distributions of the model parameters in both trend and seasonal analysis were updated through Markov Chain Monte Carlo Simulation mainly utilizing Gibbs sampler. This study proposed a way to make use of nonstationary frequency model for dynamic risk analysis, and showed an increase of hydrologic risk with time varying probability density functions. In addition, full annual cycle of the design rainfall through seasonal model could be applied to annual control such as dam operation, flood control, irrigation water management, and so on. The proposed study showed advantage in assessing statistical significance of parameters associated with trend analysis through statistical inference utilizing derived posterior distributions.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2016.05a
• /
• pp.41-41
• /
• 2016

최근 다변량 확률모형을 이용한 빈도해석이 수문자료 등에 적용되면서 다양하게 연구되고 있으며 다변량 확률모형 중 copula 모형은 주변분포형에 대한 제약이 없어 여러 분야에 걸쳐 활발히 연구되고 있다. 강우자료는 기존 일변량 빈도해석을 수행하기 위하여 사용하던 block maxima 방법 대신 최소무강우시간(inter event time)을 통하여 강우사상을 추출하여 표본으로 사용한다. 또한 기후변화로 인한 강우량의 변화등에 대응하기 위하여 비정상성 Generalized Extreme Value(GEV)와 Gumbel 등의 확률분포형에 대한 연구도 많은 부분 이루어져 있다. 본 연구에서는, Archimedean copula 모형을 이용하여 이변량 확률모형을 구축하면서 여기에 사용되는 주변분포형에 정상성/비정상성 분포형을 적용하였다. 모형의 매개변수는 inference function for margin 방법을 이용하였으며 주변분포형으로는 정상성/비정상성 GEV, Gumbel 모형을 적용하였다. 결과로 정상성/비정상성 경향을 나타내는 지점을 구분하고 각 지점에 대한 정상성/비정상성 주변분포형을 적용한 이변량 확률분포형을 구하였다.

• Journal of Korea Water Resources Association
• /
• v.48 no.1
• /
• pp.37-44
• /
• 2015

A Bayesian nonstationary probability rainfall estimation model using the Grid method is developed. A hierarchical Bayesian framework is consisted with prior and hyper-prior distributions associated with parameters of the Gumbel distribution which is selected for rainfall extreme data. In this study, the Grid method is adopted instead of the Matropolis Hastings algorithm for random number generation since it has advantage that it can provide a thorough sampling of parameter space. This method is good for situations where the best-fit parameter values are not easily inferred a priori, and where there is a high probability of false minima. The developed model was applied to estimated target year probability rainfall using hourly rainfall data of Seoul station from 1973 to 2012. Results demonstrated that the target year estimate using nonstationary assumption is about 5~8% larger than the estimate using stationary assumption.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2011.05a
• /
• pp.379-379
• /
• 2011

본 연구는 우리나라 전국 기상관측소 중 1973년부터 2009년까지의 시 강수자료가 구축되어 있는 기상관측소 55개 지점에 대하여 비정상성 빈도해석을 수행하였다. 각 지점에 대하여 지속시간 1시간, 24시간에 대한 연 최대 강수량 자료를 구축하여 초기 20년을 기준으로 1년씩 추가한 연 최대 강수량 누적 자료를 생성하고, 생성된 기간별 자료의 평균, 위치매개변수, 축척매개변수를 산정하였으며, 위치매개변수와 축척매개변수는 확률가중모멘트법을 사용하여 산정하였다. 산정된 연 최대 평균 누적 강수량과 연도와의 선형 회귀식을 산정하여 목표연도별(2040, 2070, 2100년) 평균 강수량을 산정하였고, 위치매개변수와 축척매개변수도 평균 누적 강수량과의 선형 회귀식을 산정함으로써, 목표연도에 해당하는 각 매개변수를 산정하였다. 또한 산정된 목표연도별 평균 강수량, 위치매개변수와 축척매개변수를 이용해 확률강수량을 산정하였다. 비정상성 빈도해석을 수행하여 산정된 55개 지점에 대한 목표연도별 확률강수량을 Inverse Distance Weighted(IDW) 보간법을 사용하여 전국의 확률강수량을 공간적으로 표현하였다. 전국단위의 비정상성 빈도해석을 실시한 결과, 전체적으로 각 목표연도별 확률강수량이 증가하는 것으로 나타났으나, 일부 감소하는 지역도 나타났다. 경기도와 강원도 등 중부지역에서 확률강수량의 증가가 큰 것으로 나타났으며, 특히 강원도(강릉, 인재 등) 지역에서 확률강수량의 증가폭이 가장 크게 나타났다. 또한 남해지역에서는 대부분 확률강수량이 감소하는 것으로 나타났고, 그중에서 전라남도 남해안 부근(장흥 등)에 확률강수량의 감소가 가장 크게 나타났으며, 경북지역과 전북지역 부근에서는 증가 또는 감소의 차이가 미비하게 나타났다. 하지만 목표연도 2070년과 2100년에 대하여 산정된 확률강수량으로부터 선형 회귀식을 통해 목표연도별 평균 강수량, 위치매개변수, 축척매개변수를 추정하여 확률강수량을 산정하는 것에 한계가 있음을 보여주었다.