Deep spatial transformers for autoregressive data-driven forecasting of geophysical turbulence

A deep spatial transformer based encoder-decoder model has been developed to autoregressively predict the time evolution of the upper layer’s stream function of a two-layered quasi-geostrophic (QG) system without any information about the lower layer’s stream function. The spatio-temporal complexity of QG flow is comparable to the complexity of 500hPa Geopotential Height (Z500) of fully coupled climate models or even the Z500 which is observed in the atmosphere, based on the instantaneous attractor dimension metric. The ability to predict autoregressively, the turbulent dynamics of QG is the first step towards building data-driven surrogates for more complex climate models. We show that the equivariance preserving properties of modern spatial transformers incorporated within a convolutional encoder-decoder module can predict up to 9 days in a QG system (outperforming a baseline persistence model and a standard convolutional encoder decoder with a custom loss function). The proposed data-driven model remains stable for multiple years thus promising us of a stable and physical data-driven climate model.

[1]  Rose Yu,et al.  Incorporating Symmetry into Deep Dynamics Models for Improved Generalization , 2020, ArXiv.

[2]  Sebastian Scher,et al.  Weather and climate forecasting with neural networks: using general circulation models (GCMs) with different complexity as a study ground , 2019, Geoscientific Model Development.

[3]  Peter Bauer,et al.  Challenges and design choices for global weather and climate models based on machine learning , 2018, Geoscientific Model Development.

[4]  Isaac M. Held,et al.  Applying the Fluctuation-Dissipation Theorem to a Two-Layer Model of Quasigeostrophic Turbulence , 2015 .

[5]  Da Yang,et al.  Size of the Atmospheric Blocking Events: Scaling Law and Response to Climate Change , 2019, Geophysical Research Letters.

[6]  Thomas Bolton,et al.  Applications of Deep Learning to Ocean Data Inference and Subgrid Parameterization , 2019, Journal of Advances in Modeling Earth Systems.

[7]  Peer Neubert,et al.  Circular Convolutional Neural Networks for Panoramic Images and Laser Data , 2019, 2019 IEEE Intelligent Vehicles Symposium (IV).

[8]  Noah D. Brenowitz,et al.  Prognostic Validation of a Neural Network Unified Physics Parameterization , 2018, Geophysical Research Letters.

[9]  P. O'Gorman,et al.  Using Machine Learning to Parameterize Moist Convection: Potential for Modeling of Climate, Climate Change, and Extreme Events , 2018, Journal of Advances in Modeling Earth Systems.

[10]  Pierre Gentine,et al.  Deep learning to represent subgrid processes in climate models , 2018, Proceedings of the National Academy of Sciences.

[11]  Jonathan A. Weyn,et al.  Can Machines Learn to Predict Weather? Using Deep Learning to Predict Gridded 500‐hPa Geopotential Height From Historical Weather Data , 2019, Journal of Advances in Modeling Earth Systems.

[12]  Devika Subramanian,et al.  Data-driven prediction of a multi-scale Lorenz 96 chaotic system using a hierarchy of deep learning methods: Reservoir computing, ANN, and RNN-LSTM , 2019, ArXiv.

[13]  Ashesh Chattopadhyay,et al.  Analog forecasting of extreme-causing weather patterns using deep learning , 2019, ArXiv.

[14]  Ashesh Chattopadhyay,et al.  Data-driven super-parameterization using deep learning: Experimentation with multi-scale Lorenz 96 systems and transfer-learning , 2020 .

[15]  S. Vannitsem,et al.  Attractor dimension of time-averaged climate observables: insights from a low-order ocean-atmosphere model , 2019, Tellus A: Dynamic Meteorology and Oceanography.

[16]  Andrew Zisserman,et al.  Spatial Transformer Networks , 2015, NIPS.