Ensemble streamflow forecasting over a cascade reservoir catchment with integrated hydrometeorological modeling and machine learning

Liu, Junjiang; Yuan, Xing; Zeng, Junhan; Jiao, Yang; Li, Yong; Zhong, Lihua; Yao, Ling

doi:https://doi.org/10.5194/hess-26-265-2022

Articles | Volume 26, issue 2

https://doi.org/10.5194/hess-26-265-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/hess-26-265-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 26, issue 2

Research article

|

18 Jan 2022

Research article |

| 18 Jan 2022

Ensemble streamflow forecasting over a cascade reservoir catchment with integrated hydrometeorological modeling and machine learning

Junjiang Liu, Xing Yuan, Junhan Zeng, Yang Jiao, Yong Li, Lihua Zhong, and Ling Yao

Data sets

Thorpex research and the science of prediction (https://apps.ecmwf.int/datasets/data/tigge/levtype=sfc/type=pf/) D. B. Parsons, M. Beland, D. Burridge, P. Bougeault, G. Brunet, J. Caughey, S. M. Cavallo, M. Charron, H. C. Davies, A. Diongue Niang, V. Ducrocq, P. Gauthier, T. M. Hamill, P. A. Harr, S. C. Jones, R. H. Langland, S. J. Majumdar, B. N. Mills, M. Moncrieff, T. Nakazawa, T. Paccagnella, F. Rabier, J.-L. Redelsperger, C. Riedel, R. W. Saunders, M. A. Shapiro, R. Swinbank, I. Szunyogh, C. Thorncroft, A. J. Thorpe, X. Wang, D. Waliser, H. Wernli, and Z. Toth https://doi.org/10.1175/BAMS-D-14-00025.1

Short summary

Hourly streamflow ensemble forecasts with the CSSPv2 land surface model and ECMWF meteorological forecasts reduce both the probabilistic and deterministic forecast error compared with the ensemble streamflow prediction approach during the first week. The deterministic forecast error can be further reduced in the first 72 h when combined with the long short-term memory (LSTM) deep learning method. The forecast skill for LSTM using only historical observations drops sharply after the first 24 h.