Data-driven predictions of a multiscale Lorenz 96 chaotic system using machine-learning methods: reservoir computing, artificial neural network, and long short-term memory network
Прогнозирование хаотической многомасштабной системы Лоренца 96 на основе данных с использованием методов машинного обучения: резервуарные вычисления, искусственная нейронная сеть и сеть с долговременной краткосрочной памятью
2020-07-02
SCID: 54.1/xz7r9vvg
Discuss with AI
Echo state networkLong short-term memoryMultiscale Lorenz 96 systemReservoir computingSurrogate modeling
Figures from the paper
Abstract (AI)
Abstract. In this paper, the performance of three deep learning methods for predicting short-term evolution and for reproducing the long-term statistics of a multi-scale spatio-temporal Lorenz 96 system is examined. The 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 96 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; Y and Z are never known or used. We show 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. The RNN-LSTM and ANN show some prediction skills as well; RNN-LSTM bests ANN. Furthermore, even after losing the trajectory, data predicted by RC-ESN and RNN-LSTM have probability density functions (PDFs) that closely match the true PDF, even at the tails. The PDF of the data predicted using ANN, however, deviates from the true PDF. Implications, caveats, and applications to data-driven and data-assisted surrogate modeling of complex nonlinear dynamical systems such as weather/climate are discussed.
Key Findings
1
ANN-generated data showed noticeable deviations from the true probability density function, limiting its statistical fidelity for surrogate modeling.
2
After trajectory divergence, RC-ESN and RNN-LSTM reproduced the true probability density function closely, including its tails.
3
RNN-LSTM and ANN retained predictive skill, with RNN-LSTM performing better than ANN.
4
Reservoir computing with an echo state network substantially outperformed ANN and RNN-LSTM for short-term prediction, accurately forecasting chaotic trajectories for hundreds of solver time steps.
5
Three machine-learning methods were evaluated for short-term forecasting and long-term statistical reproduction in a multiscale Lorenz 96 system using only slow-scale variables.
Research Object
the multiscale spatiotemporal Lorenz 96 chaotic system
Research Subject
short-term trajectory prediction and long-term statistical reproduction, including probability-density-function fidelity, using machine-learning methods
Publication Details
Publication Date
2020-07-02
Journal
Publisher
ISSN
Cited by
218
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest