Journal of Korean Society of Coastal and Ocean Engineers (한국해안·해양공학회논문집)

Korean Society of Coastal and Ocean Engineers (한국해안·해양공학회)
권호 표지
DOI QR코드

DOI QR Code

Outliers and Level Shift Detection of the Mean-sea Level, Extreme Highest and Lowest Tide Level Data

평균 해수면 및 최극조위 자료의 이상자료 및 기준고도 변화(Level Shift) 진단

  • Lee, Gi-Seop (UST, KIOST School, University of Science and Technology) ;
  • Cho, Hong-Yeon (Ocean Big Data Center, Korea Institute of Ocean Science and Technology)
  • 이기섭 (과학기술연합대학원대학교 한국해양과학기술원 스쿨 UST) ;
  • 조홍연 (한국해양과학기술원 해양빅데이터센터)
  • Received : 2020.08.27
  • Accepted : 2020.09.28
  • Published : 2020.10.31
Download PDF
⟨ Previous Next ⟩

Abstract

Modeling for outliers in time series was carried out using the MSL and extreme high, low tide levels (EHL, HLL) data set in the Busan and Mokpo stations. The time-series model is seasonal ARIMA model including the components of the AO (additive outliers) and LS (level shift). The optimal model was selected based on the AIC value and the model parameters were estimated using the 'tso' function (in 'tsoutliers' package of R). The main results by the model application, i.e.. outliers and level shift detections, are as follows. (1) The two AO are detected in the Busan monthly EHL data and the AO magnitudes were estimated to 65.5 cm (by typhoon MAEMI) and 29.5 cm (by typhoon SANBA), respectively. (2) The one level shift in 1983 is detected in Mokpo monthly MSL data, and the LS magnitude was estimated to 21.2 cm by the Youngsan River tidal estuary barrier construction. On the other hand, the RMS errors are computed about 1.95 cm (MSL), 5.11 cm (EHL), and 6.50 cm (ELL) in Busan station, and about 2.10 cm (MSL), 11.80 cm (EHL), and 9.14 cm (ELL) in Mokpo station, respectively.

부산, 목포 지점의 평균해수면(MSL)과 고극조위, 저극조위 자료의 이상자료 시계열 모델링을 수행하였다. 시계열 모델은 계절성분을 포함하는 SARIMA 모형이며, 일시적인 변화에 해당하는 이상자료(Additive Outlier, AO)와 영구적인 변화를 의미하는 기준고도 변화(Level Shift, LS)를 모델에 포함하였으며, AIC 기준에 의거하여 최적 모델을 선정하였다. 이상자료 모형의 매개변수 추정은 R 프로그램 'tsoutliers' 패키지('tso' 함수)를 이용하였다. 선정 모형을 이용하여 이상자료와 기준고도 변화 진단에 적용한 결과, 부산의 월 단위 고극조위 자료에서 2003, 2012년 발생한 태풍 매미(MAEMI), 산바(SANBA)에 의한 일시적인 수위상승을 65.5, 29.5 cm 정도로 추정하였으며, 목포의 월 단위 평균해수면 자료에서는 1983년의 영산강 하굿둑 건설 사업에 의한 기준고도 변화를 21.2 cm 정도로 추정하였다. 한편 본 연구에서 구성한 모형은 모형의 편향을 유발하는 이상자료의 영향을 포함하며, 모형에 의한 RMS 오차는 연간 자료를 사용한 경우, 부산은 MSL 1.95 cm, 고극조위, 저극조위 각각 5.11 cm, 6.50 cm이며, 목포의 경우에는 큰 조차의 영향으로 MSL 2.01 cm, 고극조위, 저극조위 각각 11.80 cm, 9.14 cm로 부산보다 다소 높게 나타났다.

Keywords

References

  1. Box, G.E., Jenkins, G.M. and Reinsel, G.C. (2008). Time Series Analysis Forecasting and Control, Fourth Edition, Chap. 9, John Wiley & Sons.
  2. Chen, C. and Liu, L.M. (1993). Joint estimation of model parameters and outlier effects in time series. J. of American Statistical Association (JASA), 88(Issue. 121), 284-297.
  3. Cho, H.Y., Jeong, S.T. and Lee, U.J. (2020). Spatial correlation analysis of the mean sea level data sets in the coastal seas. Korea. J. of KSCOE, 32(1), 85-93 (in Korean).
  4. Cho, H.Y., Lee, G.S. and Ahn, S.M. (2016). Impact of outliers on the statistical measures of the environmental monitoring data in Busan coastal sea. Ocean and Polar Research, 38(2), 149-159 (in Korean). https://doi.org/10.4217/OPR.2016.38.2.149
  5. Davison, A.C. and Hinkley, D.V. (1997), Bootstrap Methods and Their Application. Cambridge University Press.
  6. Hyndman, R.J. and Khandakar, Y. (2008). Automatic time series forecasting: the forecast package for R. Journal of Statistical Software, 26(3).
  7. Lopez-de-Lacalle, J. (2019). tsoutliers: Detection of Outliers in Time Series. R package version 0.6-8. https://CRAN.R-project.org/package=tsoutliers.
  8. Jung, B.S., Lee, O.J., Kim, K.M. and Kim, S.D. (2018). Non-stationary frequency analysis of extreme sea level using POT approach. J. of Korean Society of Hazard Mitigation, 18(7), 631-638 (in Korean). https://doi.org/10.9798/kosham.2018.18.7.631
  9. Kang, J.W. and Moon, S.R. (2000). Frequency analysis of extreme high water level at Mokpo Harbor considering tidal environment changes. J. of KSCOE, 12(4), 203-209 (in Korean).
  10. Korea Hydrographic and Oceanographic Administration, Korea Ocean Observing And Forcecating System. http://www.khoa.go.kr/oceangrid/koofs/kor/tide/tide.do Accessed 2020-09-22.
  11. Mann, H.B. (1945). Nonparametric tests against trend. Econometrica, 13, 245-259. https://doi.org/10.2307/1907187
  12. McLeod, A.I. (2011). Kendall: Kendall rank correlation and Mann-Kendall trend test. R package version 2.2. https://CRAN.R-project.org/package=Kendall.
  13. Moritz, S. and Bartz-Beielstein, T. (2017). imputeTS: time series missing value imputation in R. The R Journal, 9(1), 207. https://doi.org/10.32614/rj-2017-009
  14. Tsay, R.S. (1988). Outliers, Level shifts, and Variance changes in time series. J. of Forecasting, 7, 1-20. https://doi.org/10.1002/for.3980070102