The Study to Estimate the Fitness of Bivariate Rainfall Frequency Analysis Considering the Interdependence Between Rainfall and Wind Speed
Jae Beom Park, Byung Seok Kal, Juri Heo
Abstract
Jae Beom Park, Byung Seok Kal, Juri Heo
Abstract
본 연구에서는 강우와 바람사이의 상호의존성을 고려한 Copula 모형을 기반으로 하는 이변량 강우빈도해석 기법의 활용 가능성을 검토하였다. 이를 위해 기상청 산하 6개 관측소의 연최대강우량과 풍속을 대상으로 4개 주변분포와 Archimedean copula 모형을 적용하고 중간 확률변수에 대하여 산정한 RMSE와 AIC를 기반으로 최적모형을 선정하였다. 또한 조건부 Copula를 이용하여 확률강우량을 산정하고 바람의 영향을 고려하지 않는 기존 일변량 빈도해석 결과와의 차이를 분석하였다. 바람의 영향을 고려하여 산정한 Copula 모형의 확률강우량은 지속시간과 재현기간에 따라 차이는 있지만 기존 확률강우량 보다 크게 산정되는 것으로 분석되었다. We examine the feasibility of a bivariate rainfall frequency analysis technique that is based on the copula model, taking into account the interdependence between rainfall and wind speed. Four marginal and archimedean copula model are applied for annual maximum rainfall and wind speed of six stations under Korea Meteorology Administration(KMA). Optimal copula model is identified based on RMSE(Root Mean Square Error) and AIC(Akaike Information Criterion) using an intermediate random variable. The probability rainfall is estimated by conditional copula and is compared with the result of univariate rainfall frequency analysis. Taking into account the effect of the wind speed, the probability rainfall of copula model showed increasing trend with increment of duration time and return period.
OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
본 연구에서는 강우와 바람사이의 상호의존성을 고려한 Copula 모형을 기반으로 하는 이변량 강우빈도해석 기법의 활용 가능성을 검토하였다. 이를 위해 기상청 산하 6개 관측소의 연최대강우량과 풍속을 대상으로 4개 주변분포와 Archimedean copula 모형을 적용하고 중간 확률변수에 대하여 산정한 RMSE와 AIC를 기반으로 최적모형을 선정하였다. 또한 조건부 Copula를 이용하여 확률강우량을 산정하고 바람의 영향을 고려하지 않는 기존 일변량 빈도해석 결과와의 차이를 분석하였다. 바람의 영향을 고려하여 산정한 Copula 모형의 확률강우량은 지속시간과 재현기간에 따라 차이는 있지만 기존 확률강우량 보다 크게 산정되는 것으로 분석되었다. We examine the feasibility of a bivariate rainfall frequency analysis technique that is based on the copula model, taking into account the interdependence between rainfall and wind speed. Four marginal and archimedean copula model are applied for annual maximum rainfall and wind speed of six stations under Korea Meteorology Administration(KMA). Optimal copula model is identified based on RMSE(Root Mean Square Error) and AIC(Akaike Information Criterion) using an intermediate random variable. The probability rainfall is estimated by conditional copula and is compared with the result of univariate rainfall frequency analysis. Taking into account the effect of the wind speed, the probability rainfall of copula model showed increasing trend with increment of duration time and return period.
Key concepts: Copula (linguistics), Bivariate analysis, Akaike information criterion, Wind speed, Statistics, Univariate, Mathematics, Mean squared error