A Machine Learning‐Based Global Atmospheric Forecast Model

Abstract

The paper investigates the applicability of machine learning (ML) to weather prediction by building a reservoir computing-based, low-resolution, global prediction model. The model is designed to take advantage of the massively parallel architecture of a modern supercomputer. The forecast performance of the model is assessed by comparing it to that of daily climatology, persistence, and a numerical (physics-based) model of identical prognostic state variables and resolution. Hourly resolution 20-day forecasts with the model predict realistic values of the atmospheric state variables at all forecast times for the entire globe. The ML model outperforms both climatology and persistence for the first three forecast days in the midlatitudes, but not in the tropics. Compared to the numerical model, the ML model performs best for the state variables most affected by parameterized processes in the numerical model. a computationally efficient hybrid modeling approach. The present paper implements the parallel ML technique of Pathak, Wikner, et al. (2018) to build a model that predicts the weather in the same format as a global numerical model. We train and verify the model on hourly ERA5 reanalysis data from the European Centre for Medium-Range Weather Forecasts (Hersbach et al., 2019).The work presented here can also be considered an attempt to develop a ML model that can predict the evolution of the three-dimensional, multivariate, global atmospheric state. To the best of our knowledge, the only similar prior attempts were those by Scher (2018) and Scher and Messori (2019), but they trained their three-dimensional multivariate ML model on data that were produced by low-resolution numerical model simulations. In addition, Dueben and Bauer (2018) and Weyn et al., 2019Weyn et al., (2020 designed ML models to predict two-dimensional, horizontal fields of select atmospheric state variables. Similar to our verification strategy, they also verified the ML forecasts against reanalysis data. Compared to all of the aforementioned studies, an important new aspect of our work is that we employ reservoir computing (RC) (Jaeger, 2001;Lukoševičius & Jaeger, 2009;Lukoševičius, 2012;Maass et al., 2002) rather than deep learning (e.g., Goodfellow et al., 2016), which is primarily motivated by the significantly lower computer wall clock time required to train an RC-based model. This difference in training efficiency would allow for a larger number of experiments to tune the ML model at higher resolutions.The structure of the paper is as follows. Section 2 describes the ML model, while section 3 presents the results of the forecast experiments, using as benchmarks persistence of the atmospheric state, daily climatology, and numerical forecasts from a physics-based model of identical prognostic state variables and resolution. Section 4 summarizes our conclusions.

Deep learning for bias correction of MJO prediction

“…A recent study has shown an increase of MJO prediction skill by correcting model biases with a linear statistical model 34 . Deep learning (DL) techniques have been proven to be a powerful tool for identifying weather and climate patterns 35 – 37 , sub-grid scale parameterizations 38 , 39 , weather and climate predictions 40 – 44 , and post-processing of numerical weather forecasts (shorter than 7 days 43 , 44 ). However, post-processing with DL methods has not yet been applied to MJO forecasts.…”

Online learning of both state and dynamics using ensemble Kalman filters

“…The ML techniques developed in these references are evaluated on chaotic low-order models such as the Lorenz-96 model (L96, [42]) or the two-scale Lorenz model (L05III, [41]). Moreover, the NN and reservoir models in [22,55,56,3] were tested on genuine meteorological fields.…”

State, global and local parameter estimation using local ensemble Kalman filters: applications to online machine learning of chaotic dynamics

“…There are by mid 2021, dozens of ML papers in the literature dealing with the problem of estimating the dynamics of a system from observation, even when only focusing on typical low-order models used in the field of geoscience. The problem can be addressed by typical ML techniques, such as the projection on a regressor frame or basis, random forests, analogs, diffusion maps, echo state networks, LSTM and other neural network approaches (Brunton et al, 2016;Lguensat et al, 2017;Harlim, 2018;Pathak et al, 2018;Dueben and Bauer, 2018;Fablet et al, 2018;Scher and Messori, 2019;Weyn et al, 2019;Arcomano et al, 2020). It can also be solved using a conjunction of ML and data assimilation (DA) techniques to exploit noisy and incomplete observations such as those met in realistic geoscience systems Brajard et al, 2020;Bocquet et al, 2020a;Arcucci et al, 2021).…”