Abstract:Abstract. Despite the great success of machine learning, its applications in climate dynamics have not been well developed. One concern might be how well the trained neural networks could learn a dynamical system and what can be the potential applications of this kind of learning. Detailed studies show that the coupling relations or dynamics among variables in linear or nonlinear systems can be well learnt by reservoir computer (RC) and long short-term memory (LSTM) machine learning, and these learnt coupling relations can be further applied to reconstruct one series from the other dominated by common coupling dynamics. In order to validate the above conclusions, toy models are applied to address the following three questions: (i) what can be learnt from different dynamical time series by machine learning; (ii) what factors significantly influence machine learning reconstruction; and (iii) how to select suitable explanatory or input variables for the reconstructed variable for machine learning. The results from these toy models show that both of RC and LSTM can indeed learn coupling relations among variables, and the learnt implicit coupling relation can be applied to accurately reconstruct one series from the other. Both of linear and nonlinear coupling relations between variables can influence the quality of the reconstructed series. If there is a strong linear coupling between variables, all of variables can be taken as explanatory variables for the reconstructed variable, and the reconstruction can be bi-directional. However, when the linear coupling among variables is much weaker, but with stronger nonlinear causality among variables, the reconstruction quality is direction-dependent and it may be only uni-directional. We propose using convergent cross mapping causality (CCM) index ρa→b to determine which variable can be taken as the reconstructed one and which can be taken as the explanatory variable. For example, the Pearson correlation between the average Tropical Surface Air Temperature (TSAT) and the average Northern Hemispheric SAT (NHSAT) is as weak as 0.08, but the CCM index of that NHSAT cross maps TSAT is ρN→T = 0.70, it means that NHSAT could be taken as the explanatory variable. Then we find that TSAT can be well reconstructed from NHSAT by means of RC. However, the reconstruction quality in the opposite direction is poor, because the CCM index of that TSAT cross maps NHSAT is only ρT→N = 0.24. These results also provide insights on machine learning approaches for paleoclimate reconstruction, parameterization scheme, and prediction in related climate studies.

Reconstructing Langevin systems from high and low-resolution time series using Euler and Hermite reconstructions

A maximum likelihood estimation of Lévy-driven stochastic systems for univariate and multivariate time series of observations

Small-scale Reconstruction in Three-Dimensional Kolmogorov Flows Using Four-Dimensional Variational Data Assimilation

Assessing the Suitability of the Langevin Equation for Analyzing Measured Data Through Downsampling

Statistical Inference of 1D Persistent Nonlinear Time Series and Application to Predictions

Data2Dynamics: a Modeling Environment Tailored to Parameter Estimation in Dynamical Systems.

Semiparametric maximum likelihood reconstruction of stochastic differential equations driven by white and correlated noise

Reconstruction algorithms for source term recovery from dynamical samples in catalyst models

Nonlinear stochastic modeling with Langevin regression

Reconstruction of Markovian dynamics from untimed data

Learning Interpretable Hierarchical Dynamical Systems Models from Time Series Data

Robust Reconstruction of the Fokker-Planck Equations from Time Series at Different Sampling Rates

Reconstructing Nonlinear Dynamical Systems from Multi-Modal Time Series

Model selection of chaotic systems from data with hidden variables using sparse data assimilation

Learning dynamical models from stochastic trajectories

Optimal Reconstruction of Vector Fields from Data for Prediction and Uncertainty Quantification

Relaxation-based schemes for on-the-fly parameter estimation in dissipative dynamical systems

Combining data assimilation and machine learning to build data‐driven models for unknown long time dynamics—Applications in cardiovascular modeling

Reconstruction, forecasting, and stability of chaotic dynamics from partial data

Adaptive stepsize algorithms for Langevin dynamics

Reconstructing coupled time series in climate systems by machine learning