Producing realistic climate data with generative adversarial networks

Abstract

Abstract. This paper investigates the potential of a Wasserstein generative adversarial network to produce realistic weather situations when trained from the climate of a general circulation model (GCM). To do so, a convolutional neural network architecture is proposed for the generator and trained on a synthetic climate database, computed using a simple three dimensional climate model: PLASIM. The generator transforms a “latent space”, defined by a 64-dimensional Gaussian distribution, into spatially defined anomalies on the same output grid as PLASIM. The analysis of the statistics in the leading empirical orthogonal functions shows that the generator is able to reproduce many aspects of the multivariate distribution of the synthetic climate. Moreover, generated states reproduce the leading geostrophic balance present in the atmosphere. The ability to represent the climate state in a compact, dense and potentially nonlinear latent space opens new perspectives in the analysis and handling of the climate. This contribution discusses the exploration of the extremes close to a given state and how to connect two realistic weather situations with this approach.

Spherical Sliced-Wasserstein

“…Once we have (U ) L =1 ∼ σ, we project the samples on the circle S 1 by applying Lemma 1 and we compute the coordinates on the circle using the atan2 function. Finally, we can compute the Wasserstein distance on the circle by either applying the binary search algorithm of [30] or the level median formulation (11) for SSW 1 . In the particular case in which we want to compute SSW 2 between a measure µ and the uniform measure on the sphere ν = Unif(S d−1 ), we can use the appealing fact that the projection of ν on the circle is uniform, i.e.…”

“…Using Pytorch [81], we implemented the binary search algorithm of [30] and used it with = 10 −6 . We also implemented SSW 1 using the level median formula (11) and SSW 2 against a uniform measure (12). All experiments are conducted on GPU.…”

“…A simple manifold, but with lots of practical applications, is the hypersphere S d−1 . Several types of data are by essence spherical: a good example is found in directional data [71,85] for which dedicated machine learning solutions are being developed [96], but other applications concern for instance geophysical data [32], meteorology [11], cosmology [84] or extreme value theory for the estimation of spectral measures [44]. Remarkably, in a more abstract setting, considering hyperspherical latent representations of data is becoming more and more common (e.g.…”

Artificial Intelligence for Earth System Predictability (AI4ESP) (2021 Workshop Report)

“…These are recurrent NNs (RNNs) that are built to handle temporal predictions and forecasts (Vandal et al 2017;Shen, Liu, and Wang 2021;Yu et al 2021) and convolutional NNs (CNNs) that efficiently target spatial relationships and patterns (Ise and Oba 2019;Chattopadhyay, Hassanzadeh, and Pasha 2020;Steininger et al 2020;Baño-Medina, Manzanas, and Gutiérrez 2021). Finally, generative modeling approaches, such as GANs (Besombes et al 2021;Klemmer et al 2021;Wang, Tang, and Gentine 2021) and VAEs (Tibau Alberdi et al 2018;Scher 2018;Mooers et al 2020;Zadrozny et al 2021) have demonstrated early promise in the Earth sciences (e.g., ). The VAEs, in particular, show significant potential via their ability to compactly represent and reproduce the climate state via careful construction of a nonlinear latent space.…”

Extreme heat wave sampling and prediction with analog Markov chain and comparisons with deep learning

“…Consequently, for our problem of heat wave prediction, we will consider linear and nonlinear dimensionality reduction techniques and evaluate the performance of SWG in real versus latent space. This approach is partially motivated by the emergence of generative modeling for climate and weather applications, for example, studies combining deep learning architectures with extreme value theory (EVT) for generating extremes (Bhatia et al, 2021; Boulaguiem et al, 2022) and realistic climate situations (Besombes et al, 2021).…”