Moving Horizon vs. Unscented Kalman Filter for State Estimation in Streamflow Prediction

Divas Karimanzira^1, and Thomas Rauschenbach¹

¹Department of Surface water and maritime systems, Fraunhofer IOSB, 98693, Ilmenau, Germany

Cite this paper:
Divas Karimanzira and Thomas Rauschenbach. Moving Horizon vs. Unscented Kalman Filter for State Estimation in Streamflow Prediction. World Journal of Environmental Engineering. 2022; 7(1):1-10. doi: 10.12691/wjee-7-1-1

Abstract

In this paper, a Moving Horizon Estimator (MHE) and an Unscented Kalman Filter (UKF) are applied and compared for state estimation in flood forecasting. The investigations are based on a conceptual rainfall-runoff model proposed by Lorent/Gevers for streamflow forecasting. Data for the investigations was collected from the region Trusetal in Germany. Streamflow prediction, especially for watersheds with fast response to intense rain, require the knowledge of the current state of the system (e.g., soil moisture content). Firstly, a Moving Horizon Estimator (MHE) was applied for the state estimation, due to our good experience with it in other applications, its ability to deal with non-Gaussian disturbances and the fact that the hydrologic model is nonlinear, and its states satisfy equality and inequality constraints. Due to computational intensity of the MHE, an UKF was also implemented for comparison. Even though theory and most literature conclude the superiority of MHE to UKF, in this application example the results show that the UKF and the MHE produce almost similar results with UKF slightly better, which might be due to several reasons such as problems with the initialization of the hessian matrix, choice of prediction horizon and existence of local optima in MHE. Therefore, comprehensive investigations were performed in this respect.

Keywords:
moving horizon estimation unscented Kalman filter flood prediction

This work is licensed under a Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

References:

[1]	L.Bouwer, in: Observed and Projected Impacts from Extreme Weather Events: Implications for Loss and Damage. Loss and Damage from Climate Change, pp.63-82, (2018).
[2]	Muste, M. et al. Revisiting hysteresis of flow variables in monitoring unsteady streamflows. J. Hydraul. Res. 58, 867–887 (2020).
[3]	Xiang, Z. & Demir, I. Distributed long-term hourly streamflow predictions using deep learning —a case study for State of Iowa. Environ. Model. Softw. 131, 104761 (2020).
[4]	Kadiyala, Sai Prasanth, and Wai Lok Woo. "Flood Prediction and Analysis on the Relevance of Features using Explainable Artificial Intelligence." arXiv preprint arXiv:2201.05046 (2022).
[5]	John M. Quilty, Anna E. Sikorska-Senoner, David Hah, A stochastic conceptual-data-driven approach for improved hydrological simulations, Environmental Modelling & Software, Volume 149, 2022, 105326, ISSN 1364-8152.
[6]	Paulo A. Herrera, Miguel Angel Marazuela, Thilo Hofmann, Parameter estimation and uncertainty analysis in hydrological modeling, WIREs Water, 10.1002/wat2.1569, 9, 1, (2021).
[7]	Liu, Jiandong & Doan, Chi & Liong, Shie-Yui. (2011). Conceptual Rainfall-Runoff Model with Kalman Filter for Parameter and Outflow Updating. Advances in Geosciences, Volume 23: Hydrological Science (HS).
[8]	Lu, F., Zeng, H. Application of Kalman Filter Model in the Landslide Deformation Forecast. Sci Rep 10, 1028 (2020).
[9]	Huang, Y. L. et al. A Novel Adaptive Kalman Filter With Inaccurate Process and Measurement Noise Covariance Matrices. IEEE Transactions on Automatic Control 2, 594-601 (2018).
[10]	Liu, K. et al. Application of discrete Kalman filter in dam deformation monitoring. Northwest Hydropower 3, 95-97 (2017).
[11]	Meng, X., Tong, J. & Hu, B.X. Using an ensemble Kalman filter method to calibrate parameters of a prediction model for chemical transport from soil to surface runoff. Environ Sci Pollut Res 28, 4404-4416 (2021).
[12]	Sun, Y. & Bao, Weikai & Valk, K. & Brauer, C. & Sumihar, J. & Weerts, Albrecht. (2020). Improving Forecast Skill of Lowland Hydrological Models Using Ensemble Kalman Filter and Unscented Kalman Filter. Water Resources Research. 56.
[13]	J. Komma, G. Blöschl, and C. Reszler, “Soil moisture updating by ensemble kalman filtering in real-time flood forecasting,” Journal of Hydrology, vol. 357, pp. 228-242, 2008.
[14]	Fablet, Ronan & Chapron, Bertrand & Drumetz, Lucas & Mémin, Étienne & Pannekoucke, Olivier & Rousseau, François. (2020). Learning Variational Data Assimilation Models and Solvers.
[15]	H.-J. Hoffmeyer-Zlotnik and J. Wernstedt, “Drawing up and testing of new models for an operational water quantity forecast for river basin- test example river basin of the werra,” in 8th IFAC-Congress, Kyoto, August 1981.
[16]	B. Lorent and M. Gevers, “Identification of rainfall-runoff processes,” in 4th IFAC Symposium on Identification and System Parameter Estimation, Tbilisi (USSR), 1976, pp. 735-744.
[17]	D. I. Wilson, M. Agarwal, and D. Rippin, “Experiences implementing the extended kalman filter on an industrial batch reactor,” Computational Chemical Engineering, vol. 22, pp. 1653-1672, 1998.
[18]	S. Julier and J. Uhlmann, “Unscented filtering and nonlinear estimation.” IAHS Publ. no. 147, 2004, (Proceedings of the IEEE).
[19]	Daid, Assia & Busvelle, Eric & Mohamed, Aidene. (2020). On the convergence of the unscented Kalman filter. European Journal of Control. 57.
[20]	C. V. Rao, J. B. Rawlings, and D. Q. Mayne, “Constrained state estimation for nonlinear discrete-time systems: stability and moving horizon approximations,” IEEE Trans. Automat. Control, vol. 48, no. 2, pp. 246-258, 2003.
[21]	Schiller, Julian & Muller, Matthias. (2022). Suboptimal nonlinear moving horizon estimation. IEEE Transactions on Automatic Control. 1-1.
[22]	R. O. Imhoff, C. C. Brauer, K. J. Heeringen, R. Uijlenhoet, A. H. Weerts, Large-Sample Evaluation of Radar Rainfall Nowcasting for Flood Early Warning, Water Resources Research, 58, 3, (2022).