Data-driven prediction of a multi-scale Lorenz 96 chaotic system using a hierarchy of deep learning methods: Reservoir computing, ANN, and RNN-LSTM
share
In this paper, the performance of three deep learning methods for predicting short-term evolution and reproducing the long-term statistics of a multi-scale spatio-temporal Lorenz 96 system is examined. The three methods are: echo state network (a type of reservoir computing, RC-ESN), deep feed-forward artificial neural network (ANN), and recurrent neural network with long short-term memory (RNN-LSTM). This Lorenz system has three tiers of nonlinearly interacting variables representing slow/large-scale (X), intermediate (Y), and fast/small-scale (Z) processes. For training or testing, only X is available, and Y and Z are never known/used. It is shown that RC-ESN substantially outperforms ANN and RNN-LSTM for short-term prediction, e.g., accurately forecasting the chaotic trajectories for hundreds of numerical solver's time steps, equivalent to several Lyapunov timescales. ANN (RNN-LSTM) shows some (little) prediction skills. It is also shown that even after losing the trajectory, data predicted by RC-ESN have a probability density function (PDF) that closely matches the true PDF, even at the tails. PDFs of the data predicted by ANN or RNN-LSTM do not match the true PDF. Implications of the findings, caveats, and applications to data-driven and inexact, data-assisted surrogate modeling of complex dynamical systems such as weather/climate are discussed.
READ FULL TEXT