한국습지학회지 (Journal of Wetlands Research)

한국습지학회 (Korean Wetlands Society)
권호 표지
DOI QR코드

DOI QR Code

장기 기후 변동성을 고려한 인공신경망 앙상블 모형 적용: 한강 유역 댐 유입량 예측을 중심으로

Application of Artificial Neural Network Ensemble Model Considering Long-term Climate Variability: Case Study of Dam Inflow Forecasting in Han-River Basin

  • 김태림 (연세대학교 건설환경공학과) ;
  • 주경원 (연세대학교 건설환경공학과) ;
  • 조완희 (K-Water 통합물관리처) ;
  • 허준행 (연세대학교 건설환경공학과)
  • Kim, Taereem (Department of Civil and Environmental Engineering, Yonsei university) ;
  • Joo, Kyungwon (Department of Civil and Environmental Engineering, Yonsei university) ;
  • Cho, Wanhee (Integrated River Basin Mnagement Division, K-water) ;
  • Heo, Jun-Haeng (Department of Civil and Environmental Engineering, Yonsei university)
  • 투고 : 2019.10.07
  • 심사 : 2019.11.21
  • 발행 : 2019.12.30
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

최근 장기적인 기후 변동성을 고려하기 위하여 대기-해양 순환 패턴을 수치화한 기상인자가 수문 변수 예측에 널리 사용되고 있다. 또한 정확하고 안정적인 예측을 위해 인공신경망 기반의 예측 모형이 꾸준히 발전하고 있다. 기상인자를 활용하여 기후 변동성을 고려한 수문량 예측은 수자원 및 환경 보존의 장기적인 관리에 효율적으로 활용될 수 있으므로 수문 변수에 유의한 인자의 파악과 이를 활용한 예측 모형의 적용은 꾸준한 도전이 될 것이다. 본 연구에서는 우리나라 한강 유역 댐 유입량에 통계적으로 유의한 상관성이 있는 대표 기상인자를 선정하고, 이를 인공신경망 앙상블 모형에 적용하여 댐 유입량 예측을 수행하였다. 이를 위해 앙상블 경험적 모드분해법을 활용하여 댐 유입량과 기상인자간의 통계적 상관성을 확인하였으며, 기존 단일 인공신경망 모형의 한계를 보완한 인공신경망 앙상블 모형을 구축하였다. 예측 수행 결과, 5개 댐 상관계수 평균이 훈련 기간에서 0.88, 검증 기간에서 0.68의 예측력을 보이는 것을 확인하였으며, 본 연구에서의 절차를 토대로 우리나라의 다양한 수문 변수와 기후 변동성간의 관계를 활용한 다양한 적용 사례가 나오길 기대한다.

Recently, climate indices represented by quantifying atmospheric-ocean circulation patterns have been widely used to predict hydrologic variables for considering long-term climate variability. Hydrologic forecasting models based on artificial neural networks have been developed to provide accurate and stable forecasting performance. Forecasts of hydrologic variables considering climate variability can be effectively used for long-term management of water resources and environmental preservation. Therefore, identifying significant indicators for hydrologic variables and applying forecasting models still remains as a challenge. In this study, we selected representative climate indices that have significant relationships with dam inflow time series in the Han-River basin, South Korea for applying the dam inflow forecasting model. For this purpose, the ensemble empirical mode decomposition(EEMD) method was used to identify a significance between dam inflow and climate indices and an artificial neural network(ANN) ensemble model was applied to overcome the limitation of a single ANN model. As a result, the forecasting performances showed that the mean correlation coefficient of the five dams in the training period is 0.88, and the test period is 0.68. It can be expected to come out various applications using the relationship between hydrologic variables and climate variability in South Korea.

키워드

참고문헌

  1. Abbot, J, and Marohasy, J (2014). Input selection and optimisation for monthly rainfall forecasting in Queensland, Australia, using artificial neural networks, Atmospheric Research, 138, pp. 166-178. https://doi.org/10.1016/j.atmosres.2013.11.002
  2. Anctil, F, and Nicolas, L (2004). Generalisation for neural networks through data sampling and training procedures, with applications to streamflow predictions, Hydrology and Earth System Sciences, 8(5), pp. 940-958. https://doi.org/10.5194/hess-8-940-2004
  3. Boucher, MA, Perreault, L, and Anctil, F (2009). Tools for the assessment of hydrological ensemble forecasts obtained by neural networks, J. of Hydroinformatics, 11.3-4: 297-307. https://doi.org/10.2166/hydro.2009.037
  4. Cannon, AJ, and Whitfield, PH (2002). Downscaling recent streamflow conditions in British Columbia, Canada using ensemble neural network models, J. of Hydrology, 259, pp. 136-151. https://doi.org/10.1016/S0022-1694(01)00581-9
  5. Huang, NE, Shen, Z, Long, SR, Wu, MC, Shih, HH, Zheng, Q, Yen, NC, Tung, CC, and Liu, HH (1998). The empirical mode decomposition method and the Hilbert spectrum for non-stationary time series analysis, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 354, pp. 903-995.
  6. Jung, TH, Kim, H, Kim, H, and Heo, JH (2019). Selection of climate indices for nonstationary frequency analysis and estimation of rainfall quantiles, J. of the Korean Society of Civil Engineers, 39(1), pp. 165-174. [Korean Literature] https://doi.org/10.12652/KSCE.2019.39.1.0165
  7. Kim, H, Kim, T, Shin, H, and Heo, JH (2017). A study on a tendency of parameters for nonstationary distribution using ensemble empirical mode decomposition method, J. of Korea Water Resources Association, 50(4), pp. 253-261. [Korean Literature] https://doi.org/10.3741/JKWRA.2017.50.4.253
  8. Kim, SE, and Seo, IW (2015). Artificial neural network ensemble modeling with conjunctive data clustering for water quality prediction in rivers, J. of Hydro-Environment Research, 9 (3), pp. 325-339
  9. Kim, T, Shin, JY, Kim, S, and Heo, JH (2018). Identification of relationships between climate indices and long-term precipitation in South Korea using ensemble empirical mode decomposition, J. of Hydrology, 557, pp. 726-739. https://doi.org/10.1016/j.jhydrol.2017.12.069
  10. Kim, Y, and Cho, K (2016). Sea level rise around Korea: analysis of tide gauge station data with the ensemble empirical mode decomposition method, J. of Hydro-environment Research, 11, pp. 138-145.
  11. Kolen, JF, and Pollack, JB (1991). Back propagation is sensitive to initial conditions, Advances in Neural Information Processing Systems, 3, pp. 860-867.
  12. Lee, T, and Ouarda, TBMJ (2019). Multivariate nonstationary oscillation simulation of climate indices with empirical mode decomposition, Water Resources Research, 55. https://doi.org/10.1029/2018WR023892.
  13. Li, J, Liu, X, and Chen, F (2015). Evaluation of nonstationarity in annual maximum flood series and the associations with large-scale climate patterns and human activities, Water Resources Management, 29(5), pp. 1653-1668. https://doi.org/10.1007/s11269-014-0900-z
  14. Mekanik, F, Imteaz, MA, Gato-Trinidad, S, and Elmahdi, A (2013). Multiple regression and artificial neural network for long-term rainfall forecasting using large scale climate modes, J. of Hydrology, 503, pp. 11-21. https://doi.org/10.1016/j.jhydrol.2013.08.035
  15. Santos-Garcia, G, Varela, G, Novoa, N, and Jimenez, MF (2004). Prediction of postoperative morbidity after lung resection using an artificial neural network ensemble, Artificial Intelligence in Medicine, 30, pp.61-69. https://doi.org/10.1016/S0933-3657(03)00059-9
  16. Schepen, A, Wang, QJ, and Robertson, D (2012). Evidence for using lagged climate indices to forecast Australian seasonal rainfall, J. of Climate, 25, pp. 1230-1246. https://doi.org/10.1175/JCLI-D-11-00156.1
  17. Tiwari, MK, and Adamowski, J (2013). Urban water demand forecasting and uncertainty assessment using ensemble wavelet-bootstrap-neural network models, Water Resources Research, 49, pp. 6486-6507. https://doi.org/10.1002/wrcr.20517
  18. Wu, Z, and Huang, NE (2004). A study of the characteristics of white noise using the empirical mode decomposition method, Proceedings of the Royal Society of London. Series A, 460, pp. 1597-1611. https://doi.org/10.1098/rspa.2003.1221
  19. Wu, Z, Huang, NE, Long, SR, and Peng, CK (2007). On the trend, detrending, and variability of nonlinear and nonstationary time series, PNAS, 104(38), pp. 14889-14894.
  20. Wu, Z, and Huang, NE (2009). Ensemble empirical mode decomposition: a noise-assisted data analysis method, Advances in Adaptive Data Analysis, 1(1), pp. 1-41. https://doi.org/10.1142/S1793536909000047
  21. Zaier, I, Shu, C, Ouarda, TBMJ, Seidou, O, and Chebana, F (2010). Estimation of ice thickness on lakes using artificial neural network ensembles, J. of Hydrology, 383, pp. 330-340. https://doi.org/10.1016/j.jhydrol.2010.01.006