Return Levels of Maximum Temperature in Busan, Korea

• 발행기관: 한국자료분석학회
• 저자명: Keon Tae Sohn Xianglong Liang
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.15 No.1, 11~20쪽, 전체 10쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2013.02.28

영문 초록

The objectives of this study are to fit the extreme value distribution to maximum temperature and to examine dynamic patterns of distributions and return levels based on moving-windowing. As the extreme value distributions, we considered two extreme value distributions. The one is a generalized extreme value distribution (GEV) for the block maxima method and the other is a generalized Pareto distribution (GPD) for the threshold method. For this study, we used daily maximum temperatures in Busan, Korea, for a period of 51 years (1960 to 2010). Some return levels were obtained from the estimated distributions and compared. To examine the dynamic features, the GEV and the GPD were fitted to 30-year historical data and return levels were obtained from the estimated models. Dynamic features of parameters and return levels were examined based on the 30-year moving-window method with shifting data period separately.

목차

1. Introduction

2. Review on extreme value distributions

3. Data

4. Block maxima method using GEV model

5. Threshold method using GPD model

6. Concluding Remarks

References

키워드

Maximum temperature GEV GPD Return level Moving-window