Research on water level forecasting and hydraulic parameter calibration in the 1D open channel hydrodynamic model using data assimilation

Abstract

Data assimilation (DA) is an essential technique for improving the accuracy of state prediction in numerical simulation models. To enhance the effectiveness of DA in state estimation, the state-parameter joint data assimilation strategy has been proposed. However, limited attention has been given to the impact of state-parameter estimation Ensemble Kalman Filter (EnKF) on multi-parameter calibration in hydrodynamic models. In this paper, aiming at meeting the state correction and multi-parameter calibration requirements in a water delivery system under gate control, a state-parameter estimation EnKF framework and a model error setting strategy in the EnKF is proposed. The proposed method is applied to both a synthetic and a real-world open channel water delivery system and compared with the state estimation EnKF and the state-parameter estimation EnKF without model errors. The results demonstrate that the state-parameter estimation EnKF achieves better water level assimilation and forecasting performance than the state estimation EnKF, and that incorporating model errors into the assimilation significantly improves the water level estimation and parameter calibration. Furthermore, adding model error to the state variables other than water level or flow variables can further enhance the water level forecasting accuracy. This study provides valuable insights for guiding the state-parameter assimilation in open channel hydrodynamic models and realizing the state perception of open channel water delivery systems.