KSCE Journal of Civil and Environmental Engineering Research (대한토목학회논문집)

Korean Society of Civil Engeneers (대한토목학회)
권호 표지
DOI QR코드

DOI QR Code

Development of Temporal Downscaling under Climate Change using Vine Copula

Vine Copula를 활용한 기후변화 시나리오 시간적 상세화 기법 개발

  • 유재웅 (세종대학교 공학대학 건설환경공학과) ;
  • 권윤정 (세종대학교 공학대학 건설환경공학과) ;
  • 박민우 (세종대학교 공학대학 건설환경공학과) ;
  • 권현한 (세종대학교 공학대학 건설환경공학과)
  • Received : 2023.12.01
  • Accepted : 2024.01.25
  • Published : 2024.04.01
Download PDF
⟨ Previous Next ⟩

Abstract

A Copula approach has the advantage of providing structural dependencies for representing multivariate distributions for the hydrometeorological variable marginal distribution involved, however, copulas are inflexible for extending in high dimensions, and satisfy certain assumptions to make the dependency. In addition, since the process of estimating design rainfall under the future climate associated with durations given a return period is mainly analyzed by 24-hour annual maximum rainfalls, the dependency structure contains information only on the daily and sub-daily extreme precipitation. Methods based on bivariate copula do not provide information for other duration's dependencies, which causes the intensity to be reversed. The vine copula has been proposed to process the multivariate analysis as vines consisting of trees with nodes and edges connecting pair-copula construction. In this study, we aimed to downscale under climate change to produce sub-daily extreme precipitation data considering different durations based on vine copula.

일반적으로 수문기상변량의 상관관계를 파악하기 위해서는 Copula 기법을 활용하여 의존관계를 규명하고 있으나, 단순히 Copula 기법을 다변량으로 확장하는 것은 분석결과가 유연하지 않으며 Copula 기법에 대한 수학적인 가정을 확인하여 만족 여부를 판단해야 하는 등 복잡해지는 단점이 있다. 또한, 기존의 이변량 Copula 기법을 활용하여 기후변화모델의 지속시간에 따른 설계강우량을 추정하는 과정은 주로 일별 자료만을 활용하여 분석하므로 24시간 최대강우량에 대한 정보와의 의존관계를 규명하여 추정하는 방법을 채택하고 있다. 그러나, 24시간 최대강우량만을 활용하여 다른 지속시간에 대한 설계강우량을 산정하는 것은 다른 지속시간에 대한 의존관계 정보는 제공되지 않아 지속시간에 따른 강우 강도가 역전되는 현상이 야기되는 원인이다. 따라서, 본 연구에서는 변수간의 쌍구조 관계를 연결하는 Vine Copula 기법을 활용하여 다른 지속 시간에 대한 정보를 반영하여 미래 강우강도의 변화를 전망하고자 한다.

Keywords

Acknowledgement

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Water Management Program for Drought, funded by Korea Ministry of Environment(MOE)(2022003610003). This paper has been written by modifying and supplementing the KSCE 2023 CONVENTION paper.

References

  1. Bedford, T. and Cooke, R. M. (2001). "Probability density decomposition for conditionally dependent random variables modeled by vines." Annals of Mathematics and Artificial Intelligence, Springer Nature, Vol. 32, pp. 245-268, https://doi.org/10.1023/A:1016725902970. 
  2. Bedford, T. and Cooke, R. M. (2002). "Vines--a new graphical model for dependent random variables." The Annals of Statistics, Institute of Mathematical Statistics, Vol. 30, No. 4, pp. 1031-1068, https://doi.org/10.1214/aos/1031689016. 
  3. Cooke, R. M. (1997). "Markov and entropy properties of tree-and vine-dependent variables." Proceedings of the Section of Bayesian Statistical Science, American Statistical Association, Anaheim, California. 
  4. Czado, C. (2019). Analyzing dependent data with vine copulas, Springer, Cham, Zug. 
  5. Dissmann, J., Brechmann, E. C., Czado, C. and Kurowicka, D. (2013). "Selecting and estimating regular vine copulae and application to financial returns." Computational Statistics & Data Analysis, Elsevier, Vol. 59, pp. 52-69, https://doi.org/10.1016/j.csda.2012.08.010. 
  6. El Hannoun, W., Boukili Makhoukhi, A., Zoglat, A. and El Adlouni, S.-E. (2023). "Intensity-duration-frequency curves for dependent datasets." Water, MDPI, Vol. 15, No. 14, 2641, https://doi.org/10.3390/w15142641. 
  7. Ganguli, P. and Coulibaly, P. (2019). "Assessment of future changes in intensity-duration-frequency curves for Southern Ontario using North American (NA)-CORDEX models with nonstationary methods." Journal of Hydrology: Regional Studies, Elsevier, Vol. 22, 100587, https://doi.org/10.1016/j.ejrh.2018.12.007. 
  8. Joe, H. (1996). "Families of m-variate distributions with given margins and m (m-1)/2 bivariate dependence parameters." Lecture Notes-monograph Series, Vol. 28, pp. 120-141. 
  9. Joo, K.-W., Shin, J.-Y. and Heo, J.-H. (2012). "Bivariate frequency analysis of rainfall using copula model." Journal of Korea Water Resources Association, KWRA, Vol. 45, No. 8, pp. 827-837, https://doi.org/10.3741/JKWRA.2012.45.8.827 (in Korean). 
  10. Khaliq, M. N., Ouarda, T. B., Ondo, J.-C., Gachon, P. and Bobee, B. (2006). "Frequency analysis of a sequence of dependent and/or non-stationary hydro-meteorological observations: A review." Journal of Hydrology, Elsevier, Vol. 329, Nos. 3-4, pp. 534-552, https://doi.org/10.1016/j.jhydrol.2006.03.004. 
  11. Kim, J.-Y., Park, C.-Y. and Kwon, H.-H. (2016a). "A development of downscaling scheme for sub-daily extreme precipitation using conditional copula model." Journal of Korea Water Resources Association, KWRA, Vol. 49, No. 10, pp. 863-876, https://doi.org/10.3741/JKWRA.2016.49.10.863 (in Korean). 
  12. Kim, J.-Y., So, B.-J., Kim, T.-W. and Kwon, H.-H. (2016b). "A development of trivariate drought frequency analysis approach using copula function." Journal of Korea Water Resources Association, KWRA, Vol. 49, No. 10, pp. 823-833, https://doi.org/10.3741/JKWRA.2016.49.10.823 (in Korean). 
  13. Kurowicka, D. and Cooke, R. M. (2006). Uncertainty analysis with high dimensional dependence modelling, John Wiley & Sons, Hoboken, New Jersey. 
  14. Li, J., Evans, J., Johnson, F. and Sharma, A. (2017). "A comparison of methods for estimating climate change impact on design rainfall using a high-resolution RCM." Journal of Hydrology, Elsevier, Vol. 547, pp. 413-427, https://doi.org/10.1016/j.jhydrol.2017.02.019. 
  15. Mailhot, A., Duchesne, S., Caya, D. and Talbot, G. (2007). "Assessment of future change in intensity-duration-frequency (IDF) curves for Southern Quebec using the Canadian Regional Climate Model (CRCM)." Journal of Hydrology, Elsevier, Vol. 347, Nos. 1-2, pp. 197-210, https://doi.org/10.1016/j.jhydrol.2007.09.019. 
  16. Martel, J.-L., Brissette, F. P., Lucas-Picher, P., Troin, M. and Arsenault, R. (2021). "Climate change and rainfall intensityduration-frequency curves: Overview of science and guidelines for adaptation." Journal of Hydrologic Engineering, ASCE, Vol. 26, No. 10, 03121001, https://doi.org/10.1061/(ASCE)HE.1943-5584.0002122. 
  17. Masseran, N. (2021). "Modeling the characteristics of unhealthy air pollution events: A copula approach." International Journal of Environmental Research and Public Health, MDPI, Vol. 18, No. 16, 8751, https://doi.org/10.3390/ijerph18168751. 
  18. Ministry of Environment(ME) (2019). Guideline of design flood estimation (in Korean). 
  19. Sklar, M. (1959). "Fonctions de repartition a n dimensions et leurs marges." Annales de l'ISUP, de l'Institut de Statistique de L'Universite de Paris, Vol. 8, No. 3, pp. 229-231 (in French). 
  20. So, B.-J., Kim, J.-Y., Kwon, H.-H. and Lima, C. H. (2017). "Stochastic extreme downscaling model for an assessment of changes in rainfall intensity-duration-frequency curves over South Korea using multiple regional climate models." Journal of Hydrology, Elsevier, Vol. 553, pp. 321-337, https://doi.org/10.1016/j.jhydrol.2017.07.061. 
  21. Sugahara, S., Da Rocha, R. P. and Silveira, R. (2009). "Nonstationary frequency analysis of extreme daily rainfall in Sao Paulo, Brazil." International Journal of Climatology, Royal Meteorological Society, Vol. 29, No. 9, pp. 1339-1349, https://doi.org/10.1002/joc.1760. 
  22. Tramblay, Y., Neppel, L., Carreau, J. and Najib, K. (2013). "Non-stationary frequency analysis of heavy rainfall events in southern France." Hydrological Sciences Journal, International Association of Hydrological Sciences, Vol. 58, No. 2, pp. 280-294, https://doi.org/10.1080/02626667.2012.754988. 
  23. Zhang, B., Wang, S., Moradkhani, H., Slater, L. and Liu, J. (2022). "A vine copula-based ensemble projection of precipitation intensity-duration-frequency curves at sub-daily to multi-day time scales." Water Resources Research, AGU, Vol. 58, No. 11, https://doi.org/10.1029/2022WR032658.