A wavelet-based approach to detect climate change on the coherent and turbulent component  of the atmospheric circulation

Faranda, Davide; Defrance, Dimitri

doi:10.5194/esd-7-517-2016

Cited by 6 publications

(4 citation statements)

References 41 publications

(41 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead, τ must be a multiple of this quantity but should be smaller than the residence time in the basin of attraction. In previous works, some of the authors of this paper have shown the validity of the Υ indicator to study financial [22] and climate time series [23].…”

Section: An Indicator Of Stability Based On the Stochastic Modelsmentioning

confidence: 78%

Statistical-mechanical approach to study the hydrodynamic stability of the stably stratified atmospheric boundary layer

et al. 2017

View full text Add to dashboard Cite

We study the hydrodynamic equilibrium properties of the stably stratified atmospheric boundary layer from measurements obtained in the Snow-Horizontal Array Turbulence Study (SnoHATS) campaign at the Plaine Morte Glacier in the Swiss Alps. Our approach is based on a combination of dynamical systems techniques and statistical analysis. The main idea is to measure the deviations from the behavior expected by a turbulent observable when it is close to a transition between different metastable states. We first assess the performance of our method on the Lorenz attractor, then on a turbulent flow. The results show that the method recognizes subtle differences among different stable boundary layers turbulence regimes and may be used to help characterise their transitions.

show abstract

Section: An Indicator Of Stability Based On the Stochastic Modelsmentioning

confidence: 78%

Statistical-mechanical approach to study the hydrodynamic stability of the stably stratified atmospheric boundary layer

et al. 2017

View full text Add to dashboard Cite

show abstract

Section: Stability Indicatormentioning

confidence: 79%

On reversals in 2D turbulent Rayleigh-Bénard convection: Insights from embedding theory and comparison with proper orthogonal decomposition analysis

Faranda

Podvin

Sergent

2019

Chaos: An Interdisciplinary Journal of Nonlinear Science

Self Cite

View full text Add to dashboard Cite

Turbulent Rayleigh-Bénard convection in a 2D square cell is characterized by the existence of a large-scale circulation which varies intermittently. We focus on a range of Rayleigh numbers where the large-scale circulation experiences rapid non-trivial reversals from one quasi-steady (or metastable) state to another. In previous work (Podvin and Sergent JFM 2015, Podvin and Sergent PRE 2017), we applied Proper Orthogonal Decomposition (POD) to the joint temperature and velocity elds at a given Rayleigh number and the dynamics of the ow were characterized in a multi-dimensional POD space. Here, we show that several of those ndings, which required extensive data processing over a wide range of both spatial and temporal scales, can be reproduced, and possibly extended, by application of embedding theory to a single time series of the global angular momentum, which is equivalent here to the most energetic POD mode. Specically, embedding theory conrms that the switches among meta-stable states are uncorrelated. It also shows that, despite the large number of degrees of freedom of the turbulent Rayleigh Benard ow, a low dimensional description of its physics can be derived with low computational eorts, providing that a single global observable reecting the symmetry of the system is identied. A strong connection between the local stability properties of the reconstructed attractor and the characteristics of the reversals can also be established. We use Rayleigh-Bénard turbulent convection simulations to compare the properties of quasistationary states and transitions (reversals) previously identied via Principal Orthogonal Decomposition (POD) to those obtained via embedding the global angular momentum of the system. Our main result is to show that we can map POD properties on the attractor reconstructed via embedding techniques. Specically, the embedding technique conrms that the switches among dierent metastable states are uncorrelated. Moreover, a local stability indicator can be used to distinguish and classify the dierent metastable states. The low computational costs of embedding analysis suggests to use this procedure whenever a global observable reecting the symmetry of the system can be identied, while the POD should be preferred when such information is not available.

show abstract

“…However, pinning down this concept of coherent structures with rigorous and principled definitions or heuristics which produce consistent results across a wide class of physical systems is a challenging and open problem [4]. Climate practitioners are left with more ad hoc approaches [5], [6], [7] which can make it difficult to draw meaningful conclusions from analysis [8].…”

Section: Motivationmentioning

confidence: 99%

DisCo: Physics-Based Unsupervised Discovery of Coherent Structures in Spatiotemporal Systems

Rupe

Prabhat

Crutchfield

et al. 2019

2019 IEEE/ACM Workshop on Machine Learning in High Performance Computing Environments (MLHPC)

View full text Add to dashboard Cite

Extracting actionable insight from complex unlabeled scientific data is an open challenge and key to unlocking data-driven discovery in science. Complementary and alternative to supervised machine learning approaches, unsupervised physics-based methods based on behavior-driven theories hold great promise. Due to computational limitations, practical application on real-world domain science problems has lagged far behind theoretical development. However, powerful modern supercomputers provide the opportunity to narrow the gap between theory and practical application. We present our first step towards bridging this divide -DisCo -a high-performance distributed workflow for the behavior-driven local causal state theory. DisCo provides a scalable unsupervised physics-based representation learning method that decomposes spatiotemporal systems into their structurally relevant components, which are captured by the latent local causal state variables. Complex spatiotemporal systems are generally highly structured and organize around a lower-dimensional skeleton of coherent structures, and in several firsts we demonstrate the efficacy of DisCo in capturing such structures from observational and simulated scientific data. To the best of our knowledge, DisCo is also the first application software developed entirely in Python to scale to over 1000 machine nodes, providing good performance along with ensuring domain scientists' productivity. We developed scalable, performant methods optimized for Intel many-core processors that will be upstreamed to open-source Python library packages. Our capstone experiment, using newly developed DisCo workflow and libraries, performs unsupervised spacetime segmentation analysis of CAM5.1 climate simulation data, processing an unprecedented 89.5 TB in 6.6 minutes end-to-end using 1024 Intel Haswell nodes on the Cori supercomputer obtaining 91% weak-scaling and 64% strong-scaling efficiency. This enables us to achieve state-of-the-art unsupervised segmentation of coherent spatiotemporal structures in complex fluid flows.Recently, supervised DL techniques have been applied to address this problem [24], [25], [26] including one of the 2018 Gordon Bell award winners [27]. However, there is an immediate and daunting challenge for these supervised approaches: ground-truth labels do not exist for pixel-level identification of extreme weather events [21]. The DL models used in the above studies are trained using the automated heuristics of TECA [20] for proximate labels. While the results in [24] qualitatively show that DL can improve upon TECA, the results in [26] reach accuracy rates over 97%, essentially reproducing the output of TECA. The supervised learning paradigm of optimizing objective metrics (e.g. training and generalization error) breaks down here [8] since TECA is not ground truth and we do not know how to train a DL model to disagree with TECA in just the right way to get closer to "ground truth".

show abstract

A wavelet-based approach to detect climate change on the coherent and turbulent component of the atmospheric circulation

Cited by 6 publications

References 41 publications

Statistical-mechanical approach to study the hydrodynamic stability of the stably stratified atmospheric boundary layer

Statistical-mechanical approach to study the hydrodynamic stability of the stably stratified atmospheric boundary layer

On reversals in 2D turbulent Rayleigh-Bénard convection: Insights from embedding theory and comparison with proper orthogonal decomposition analysis

DisCo: Physics-Based Unsupervised Discovery of Coherent Structures in Spatiotemporal Systems

Contact Info

Product

Resources

About