Estimating model‐error covariances in nonlinear state‐space models using Kalman smoothing and the expectation–maximization algorithm

Specification and tuning of errors from dynamical models are important issues in data assimilation. In this work, we propose an iterative expectation-maximisation (EM) algorithm to estimate the model error covariances using classical extended and ensemble versions of the Kalman smoother. We show that, for additive model errors, the estimate of the error covariance converges. We also investigate other forms of model error, such as parametric or multiplicative errors. We show that additive Gaussian model error is able to compensate for non additive sources of error in the algorithms we propose. We also demonstrate the limitations of the extended version of the algorithm and recommend the use of the more robust and flexible ensemble version. This article is a proof of concept of the methodology with the Lorenz-63 attractor. We developed an open-source Python library to enable future users to apply the algorithm to their own nonlinear dynamical models.

[1]  Didier Auroux,et al.  Smoothing Problems in a Bayesian Framework and Their Linear Gaussian Solutions , 2012 .

[2]  B. Khattatov,et al.  Data assimilation : making sense of observations , 2010 .

[3]  P. Houtekamer,et al.  An Adaptive Ensemble Kalman Filter , 2000 .

[4]  X. Deng,et al.  Model Error Representation in an Operational Ensemble Kalman Filter , 2009 .

[5]  Yudong Tian,et al.  Modeling errors in daily precipitation measurements: Additive or multiplicative? , 2013 .

[6]  François Desbouvries,et al.  On Bayesian Fixed-Interval Smoothing Algorithms , 2008, IEEE Transactions on Automatic Control.

[7]  G. Ueno,et al.  Bayesian estimation of the observation‐error covariance matrix in ensemble‐based filters , 2016 .

[8]  Guanrong Chen,et al.  Kalman Filtering with Real-time Applications , 1987 .

[9]  Dinh-Tuan Pham,et al.  A simplified reduced order Kalman filtering and application to altimetric data assimilation in Tropical Pacific , 2002 .

[10]  Dean S. Oliver,et al.  THE ENSEMBLE KALMAN FILTER IN RESERVOIR ENGINEERING-A REVIEW , 2009 .

[11]  R. Daley.  Estimating Model-Error Covariances for Application to Atmospheric Data Assimilation , 1992 .

[12]  Ibrahim Hoteit,et al.  Mitigating Observation Perturbation Sampling Errors in the Stochastic EnKF , 2015 .

[13]  G. Evensen,et al.  Analysis Scheme in the Ensemble Kalman Filter , 1998 .

[14]  Tomoyuki Higuchi,et al.  Maximum likelihood estimation of error covariances in ensemble‐based filters and its application to a coupled atmosphere–ocean model , 2010 .

[15]  D. Dee.  On-line Estimation of Error Covariance Parameters for Atmospheric Data Assimilation , 1995 .

[16]  E. Lorenz.  Deterministic nonperiodic flow , 1963 .

[17]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[18]  Ibrahim Hoteit,et al.  Using Low-Rank Ensemble Kalman Filters for Data Assimilation with High Dimensional Imperfect Models 1 , 2007 .

[19]  I. Hoteit,et al.  Review of nonlinear Kalman, ensemble and particle filtering with application to the reservoir history matching problem , 2011 .

[20]  Timothy J. Robinson,et al.  Sequential Monte Carlo Methods in Practice , 2003 .

[21]  Matthew Newman,et al.  Multiplicative Noise and Non-Gaussianity: A Paradigm for Atmospheric Regimes? , 2005 .

[22]  E. Hairer,et al.  Solving Ordinary Differential Equations I , 1987 .

[23]  Petros G. Voulgaris,et al.  On optimal ℓ∞ to ℓ∞ filtering , 1995, Autom..

[24]  L. Paninski,et al.  Common-input models for multiple neural spike-train data , 2007, Network.

[25]  Manuel Pulido,et al.  Offline parameter estimation using EnKF and maximum likelihood error covariance estimates: Application to a subgrid‐scale orography parametrization , 2013 .

[26]  G. Evensen.  Data Assimilation: The Ensemble Kalman Filter , 2006 .

[27]  Boris Khattatov,et al.  Data Assimilation and Information , 2010 .

[28]  Tyrus Berry,et al.  Adaptive ensemble Kalman filtering of non-linear systems , 2013 .

[29]  Patrick Nima Raanes,et al.  On the ensemble Rauch‐Tung‐Striebel smoother and its equivalence to the ensemble Kalman smoother , 2016 .

[30]  Craig H. Bishop,et al.  Adaptive sampling with the ensemble transform Kalman filter , 2001 .

[31]  R. Shumway,et al.  AN APPROACH TO TIME SERIES SMOOTHING AND FORECASTING USING THE EM ALGORITHM , 1982 .

[32]  Nagatomo Nakamura,et al.  Iterative algorithm for maximum‐likelihood estimation of the observation‐error covariance matrix for ensemble‐based filters , 2014 .

[33]  C. Striebel,et al.  On the maximum likelihood estimates for linear dynamic systems , 1965 .

[34]  Sirish L. Shah,et al.  A comparison of simultaneous state and parameter estimation schemes for a continuous fermentor reactor , 2010 .