14th IFAC Symposium on System Identification, SYSID 2006

SYSID-2006 Paper Abstract

Close

Paper ThA6.4

Norton, John P (Australian National Univ.)

Identification of Abrupt Changes in Catchments by Optimal-Smoothing Adjoints

Scheduled for presentation during the Invited Session "Identification of Ecological/Environmental Systems" (ThA6), Thursday, March 30, 2006, 11:30−11:50, Newcastle Room

14th IFAC Symposium on System Identification, March 29 - 31, 2006, Newcastle, Australia

This information is tentative and subject to change. Compiled on September 24, 2018

Keywords Recursive Identification, Other, Filtering and Smoothing

Abstract

Land-use changes may significantly affect rainfall-runoff dynamics. Recursive parameter estimation in linear models offers a way to identify changes. Several authors have treated time-varying parameters as state variables, executing random walks or more elaborate dynamics and estimated by optimal smoothing. As such models cannot track both smooth and abrupt changes well, other change-detection methods are also worth investigation. A method due to Weston and Norton generates a statistic testing if the difference between parameter estimates from input-output samples before and after each sample instant is consistent with the specified observation noise variance and parameter-increment covariance. The technique is tested on artificial examples and daily rainfall-runoff records from a research catchment in Oregon, USA.