Spatio-Temporal Signatures of Intermittency in Helically Rotating Turbulence through Topological Data Analysis
Snigdhashree Mallick, Yashwanth Ramamurthi, Shiva Kumar Malapaka* · 2026
Read Spatio-Temporal Signatures of Intermittency in Helically Rotating Turbulence through Topological Data Analysis by Snigdhashree Mallick, Yashwanth Ramamurthi, Shiva Kumar Malapaka* (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-17560.txt.
Introduction
Mallick and colleagues apply topological data analysis to space-time signs of intermittency in helically rotating turbulence. Vorticity magnitude and eddy-size fields serve as scalar readouts, with shifting shape tracked through persistence diagrams and Wasserstein-distance metrics. The frame proves more sensitive than standard stats to local, short-lived flow changes. Sharp changes in heatmaps give direct signs of strong turbulent fluctuations across space and time.
Catalog record
| CREATOR | Snigdhashree Mallick, Yashwanth Ramamurthi, Shiva Kumar Malapaka* |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62775 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.17560
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Spatio-Temporal Signatures of Intermittency in Helically Rotating Turbulence through Topological Data Analysis Snigdhashree Mallick, Yashwanth Ramamurthi, Shiva Kumar Malapaka* Abstract A central challenge in hydrodynamic turbulence is identifying precisely when, and at which length scales, strong turbulent fluctuations (STFs) emerge and develop into intermittent events, which are often obscured by conventional statistical diagnostics. We address this problem by applying a Topological Data Analysis (TDA) framework to reveal the spatiotemporal signatures of intermittency in low-resolution ( 128 3 128^{3} ) helically rotating turbulent flows. Vorticity magnitude and length-scale (eddy size) fields are used as scalar observables for TDA: vorticity characterizes rotational dynamics that generate multiscale flow structures, while length-scale fields encode the scales at which intermittent activity arises. Their evolving topology is quantified using persistence diagrams and Wasserstein-distance metrics. Compared with traditional statistical approaches, this framework is more sensitive to localized and short-lived flow variations, enabling clearer detection of intermittent behavior. Pronounced variations in Wasserstein-distance heatmaps provide direct signatures of STFs across space and time. Together, these results demonstrate that TDA offers an effective complementary tool for detecting STFs that lead to intermittency within turbulent regime. Keywords: Intermittency, Strong turbulent fluctuations(STFs), Topological data analysis (TDA), Wasserstein distance 1 Introduction Intermittency remains one of the foremost challenges in turbulence research. It manifests as irregular, burst-like fluctuations in velocity, dissipation, and other flow quantities that strongly deviate from Gaussian and self-similar statistics. These rare, extreme events dominate high-order moments and play a disproportionate role in shaping the turbulent energy cascade. It is the non-linear nature of these intermittent events and their interactions with the flow, which create and destroy various structures and drive the turbulent flow. Traditionally, intermittency has been quantified using statistical tools such as structure functions, probability density functions, kurtosis, and correlation functions, which characterize fluctuations through moments of physical observables in real space [ 47 , 20 ] . These approaches primarily describe the degree of intermittency—for example, whether a flow is highly intermittent, weakly intermittent, or nearly Gaussian—while the occurrence and intensity of intermittent events depend sensitively on control parameters such as Reynolds number, rotation, and helicity. Several phenomenological frameworks have been proposed to interpret intermittency in terms of these parameters, including the log-Poisson model [ 40 ] and the Politano–Pouquet model [ 35 ] . More broadly, classical models such as the Kolmogorov–Obukhov (1962) log-normal model [ 22 , 32 ] , the \beta -model [ 19 ] , the multifractal formalism [ 33 , 20 ] , and the She–Leveque model [ 40 ] have advanced the understanding of turbulence by capturing scale-dependent statistics and deviations from Gaussianity. Despite these developments, existing approaches remain fundamentally statistical, relying on global scaling laws and ensemble-averaged quantities. Consequently, a comprehensive understanding of intermittency in turbulent flows—and a first-principles description of turbulence—remains elusive [ 42 ] . Since, existing statistical models of turbulence do not explicitly incorporate the topology of coherent structures; we propose a topology-based framework, built upon existing TDA tools, to provide a more detailed characterization of intermittency. This approach captures the evolution of topological features in both space and time, moving beyond global statistical descriptors to directly track the emergence, interaction, and disappearance of coherent structures associated with strong fluctuations leading to intermittent events. Our approach primarily addresses the questions of when and at which length scales STFs emerge and evolve, enabling a clearer detection of the spatiotemporal signatures associated with intermittency in three-dimensional turbulent flows. By addressing these questions, the proposed framework offers a pathway toward understanding how intermittent events arise from the underlying turbulent dynamics. Motivation to use TDA: TDA employs tools such as persistence diagrams, Wasserstein-distance metrics, and contour trees to extract global structural information from complex, high-dimensional data [ 7 , 15 ] . These methods provide a data-driven framework for identifying and quantifying physically interpretable topological features—such as clusters, loops, and voids—that may be overlooked by conventional statistical approaches. Unlike traditional methods that characterize intermittency through global moments or scaling laws, TDA captures the geometry, connectivity, and temporal evolution of coherent structures across multiple scales. Furthermore, the topological invariance of these features under continuous deformations makes TDA robust to noise and well suited for analyzing nonlinear, multiscale systems such as three-dimensional turbulent flows [ 25 , 50 , 10 ] . Importantly, TDA is applied to the entire simulated dataset, from the beginning to the end of the simulation, enabling each time frame to be analyzed without omission; an approach rarely adopted in conventional turbulence analyses. Although computationally intensive, this comprehensive approach provides deeper insight into the evolving topology of turbulent structures and establishes a unified framework for uncovering structural signatures of intermittency. Previous Work: In fluid dynamics, TDA has been applied extensively to two-dimensional turbulent flows [ 49 , 31 , 5 , 3 ] , whereas only a limited number of studies extend these techniques to three-dimensional turbulence [ 23 , 44 ] , primarily demonstrating feasibility. Lian et al. [ 23 ] analyzed 2D time-resolved particle image velocimetry data in a “box of turbulence” and showed that rate-of-strain tensor extraction requires median filtering at a 3\times 3 vector spacing. Makarenko et al. [ 25 ] applied Betti numbers and persistence diagrams to time-varying multiphase MHD ISM simulations, capturing magnetic-field-induced structural changes beyond autocorrelation methods. Nauleau et al. [ 31 ] used persistent homology on an ensemble of 180 Kelvin–Helmholtz simulations and showed that Wasserstein distance outperforms the L_{2} norm in distinguishing solver configurations and detecting vortex formation. Bridel-Bertomeu et al. [ 5 ] combined merge trees with topological persistence to track eddy evolution in cylindrical pipe flow, while Vivodtzev et al. [ 49 ] employed persistence-based Morse–Smale segmentation to identify vortices in early-stage instabilities. Banesh et al. [ 3 ] used contour trees and persistence filtering in 2D kinetic plasma simulations to detect multiscale magnetic reconnection dynamics, including nested magnetic islands and X-points, representing one of the earliest applications of TDA to MHD reconnection, albeit restricted to two dimensions. Turbulent flow simulations are mostly three-dimensional, and a complete characterization of intermittency requires information from all spatial directions. Since the motivation of this work is to understand the spatio-temporal origins of STFs using TDA, we perform our analysis on the full three-dimensional dataset, in contrast to many previous studies that rely on two-dimensional representations. The main contributions of this study are summarized as follows: • Identification of STFs leading to intermittent events via tracking the continuous evolution of topological features in time, by applying TDA techniques to seven low-resolution ( 128^{3} ) helically rotating hydrodynamic turbulent flow datasets generated by systematically varying helicity and rotation. • Computation of persistence diagrams from physically meaningful scalar fields such as vorticity and local length scale (eddy size) at each time frame, followed by quantitative comparison of the persistence diagrams using Wasserstein-distance metric, and the computation of contour trees in order to show the emergence and evolution of STFs leading to intermittent events in temporal frame. • Computation and literature-based classification of local length scales, together with identification of the most strongly affected scales and the physical quantities showing the largest deviations from their mean behaviour, thereby supporting the onset of STFs leading to intermittency in spatial frame. The proposed workflow clearly localizes STFs, which