Inferring instantaneous, multivariate and nonlinear sensitivities for the analysis of feedback processes in a dynamical system: Lorenz model case‐study

As an alternative to classical linear feedback analysis, we present a nonlinear approach for the determination of the sensitivities of a dynamical system from observations of its variations. The new methodology consists of statistical estimates of all the pair-wise relationships among the system state variables based on a neural-network modelling of the system dynamics (its time evolution). The model can then be used to estimate the instantaneous, multivariate, nonlinear sensitivities. Classical feedback analysis is re-examined in terms of these sensitivities, which are shown to be more fundamental in the analysis of feedback processes than estimates of feedback factors and to provide a more appropriate representation of the system's behaviour. The method is described and tested on synthetic observations of the time variations of the Lorenz low-order atmospheric model where the correct sensitivities can be evaluated analytically. Copyright © 2003 Royal Meteorological Society

[1]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1992, Math. Control. Signals Syst..

[2]  F. Aires,et al.  A new neural network approach including first guess for retrieval of atmospheric water vapor, cloud liquid water path, surface temperature, and emissivities over land from satellite microwave observations , 2001 .

[3]  Noëlle A. Scott,et al.  The "weight smoothing" regularization of MLP for Jacobian stabilization , 1999, IEEE Trans. Neural Networks.

[4]  Filipe Aires,et al.  Independent component analysis of multivariate time series: Application to the tropical SST variability , 2000 .

[5]  Roger K. Smith.  On the theory of cisk , 1997 .

[6]  Taro Takahashi,et al.  Climate processes and climate sensitivity , 1984 .

[7]  J. Peixoto,et al.  Physics of climate , 1992 .

[8]  A. Tuck,et al.  The Brewer‐Dobson Circulation In the Light of High Altitude In Situ Aircraft Observations , 1997 .

[9]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1989, Math. Control. Signals Syst..

[10]  M. Schlesinger,et al.  The application of cause‐and‐effect analysis to mathematical models of geophysical phenomena: 1. Formulation and sensitivity analysis , 1991 .

[11]  Geoffrey E. Hinton,et al.  Learning representations by back-propagating errors , 1986, Nature.

[12]  Filipe Aires.  Problèmes inverses et réseaux de neurones : application à l'interféromètre haute résolution IASI et à l'analyse de séries temporelles , 1999 .

[13]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[14]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[15]  F. Luther,et al.  The Potential climatic effects of increasing carbon dioxide , 1985 .

[16]  Edward N. Lorenz,et al.  Can chaos and intransitivity lead to interannual variability , 1990 .

[17]  W. Rossow,et al.  Advances in understanding clouds from ISCCP , 1999 .

[18]  Peter J. Webster,et al.  Thermodynamics of Atmospheres and Oceans , 1998 .

[19]  R. Allan,et al.  Water Vapor Feedbacks in the ECMWF Reanalyses and Hadley Centre Climate Model , 2000 .

[20]  James L. McClelland,et al.  Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations , 1986 .

[21]  H. W. Bode,et al.  Network analysis and feedback amplifier design , 1945 .

[22]  Tim Palmer,et al.  A Nonlinear Dynamical Perspective on Climate Prediction , 1999 .

[23]  Edward N. Lorenz,et al.  Irregularity: a fundamental property of the atmosphere* , 1984 .

[24]  Yuval.  Neural Network Training for Prediction of Climatological Time Series, Regularized by Minimization of the Generalized Cross-Validation Function , 2000 .