Technical Matters
Self-Adaptive Flow Forecasting
Tuesday, 26 April 2011
Over the past ten years, a lot of research and development work has been carried out at Lancaster in connection with real-time flow forecasting (see e.g. Young, 2002, 2010a,2010b; Romanowicz, 2006; Beven et al, 2011). The basic approach entails the identification and estimation on DBM-type nonlinear TF models for the rainfall-flow and flow-flow (flow routing) elements in a catchment model; the conversion of these models into a state space form; and the incorporation of this stochastic state space model into a modified Kalman Filter-based forecasting and data assimilation engine. State updating is an inherent part of this system but self-adaptive enhancements are included based on the recursive estimation routines in CAPTAIN. These range from simple recursive least squares estimation for gain adaption and noise variance estimation, to full model parameter estimation using the emerging Real-time RIV (RRIV) algorithm.
K.
Figure 10.7 in chapter 10 of Young (2011): one-day-ahead forecasting of daily flow on the Leaf River, USA, using a recursively updated self-adaptive gain
J. Beven, D.T. Leedal, P. J. Smith, and P. C. Young. Identification and representation of state dependent non-linearities in flood forecasting using the DBM methodology. In L.Wang, H.Garnier, editors, System Identification, Environmetric Modelling and Control, Berlin, 2011. Springer-Verlag (in press).
R. J. Romanowicz, P. C. Young, and K. J. Beven. Data assimilation and adaptive forecasting of water levels in the River Severn catchment. Water Resources Research, 42:W06407, doi:10.1029/2005WR004373, 2006.
P. C. Young. Advances in real-time flood forecasting. Philosophical Trans. Royal Society, Physical and Engineering Sciences, 360(9):1433–1450, 2002.
P. C. Young. Real-time updating in flood forecasting and warning. In G.J. Pender and H.Faulkner, editors, Flood Risk Science and Management, pages 163–195, Oxford, UK, 2010a. Wiley-Blackwell.
P. C. Young. Gauss, Kalman and advances in recursive parameter estimation. Journal of Forecasting (special issue celebrating 50 years of the Kalman Filter), 30:104–146, 2010b.
P. C. Young Recursive Estimation and Time Series Analysis: an Introduction for the Student and Practitioner, Springer-Verlag, 2011.