arxiv: 2604.25386 · v1 · submitted 2026-04-28 · ⚛️ physics.soc-ph

Recognition: unknown

Universal Features in Atmospheric Particulate Matter Dynamics

Suchismita Banerjee , Koyena Ghosh , Urna Basu , Banasri Basu

Authors on Pith no claims yet

Pith reviewed 2026-05-07 14:34 UTC · model grok-4.3

classification ⚛️ physics.soc-ph

keywords PM2.5particulate matteruniversal fluctuationsexponentially modified Gaussianstochastic modelpower spectrumair quality

0 comments

The pith

PM2.5 fluctuations display universal properties across diverse Indian cities after trend and seasonal removal.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors examine six years of daily PM2.5 data from fifty-four cities under varied conditions. They demonstrate that residual fluctuations, after subtracting slow trends and seasonal cycles, have probability densities that collapse to one curve fitted by an exponentially modified Gaussian distribution. The residuals also share similar autocorrelation decays and power spectra showing 1/f behavior at the tails. They introduce a minimal stochastic model that accounts for the distribution, correlations, and spectral features.

Core claim

The rescaled probability density functions of the residual fluctuations collapse onto a single curve and are well described by an exponentially modified Gaussian distribution. The rescaled residual time-series for all the cities further exhibit certain robust dynamical features, with similar decay of auto-correlation functions, and power spectral densities displaying a similar 1/f decay at the tails. A minimal stochastic model for the residual dynamics explains the observed universal features.

What carries the argument

The minimal stochastic model for residual dynamics that produces the exponentially modified Gaussian stationary distribution and the observed temporal correlations and spectral scaling.

If this is right

Universal collapse implies that local factors do not affect the core fluctuation statistics in the studied cities.
The exponentially modified Gaussian provides a standard form for modeling PM2.5 residuals.
Shared 1/f spectral decay indicates common long-memory properties in the dynamics.
The stochastic model serves as a simple generator for realistic fluctuation series across locations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Similar analysis on data from other regions could reveal whether this universality is specific to India or more general.
The model might be extended to forecast short-term pollution spikes using the same parameters everywhere.
Connections to other environmental time series with 1/f noise could be explored to find shared mechanisms.
Health impact studies could use the universal distribution to estimate exposure variability without city-specific data.

Load-bearing premise

That the procedure for removing slow trends and seasonal components successfully isolates the residual fluctuations without creating artificial similarities between cities.

What would settle it

If the rescaled PDFs from the fifty-four cities do not all collapse onto the same curve when using the same removal method, the claimed universality would not hold.

Figures

Figures reproduced from arXiv: 2604.25386 by Banasri Basu, Koyena Ghosh, Suchismita Banerjee, Urna Basu.

**Figure 1.** Figure 1: FIG. 1. Map of India indicating the locations of the cities view at source ↗

**Figure 3.** Figure 3: FIG. 3. Probability density functions of daily PM view at source ↗

**Figure 4.** Figure 4: FIG. 4. Data collapse of the probability density function of view at source ↗

**Figure 7.** Figure 7: FIG. 7. Plot of probability distribution function view at source ↗

**Figure 8.** Figure 8: FIG. 8. Plot of auto-correlation functions view at source ↗

**Figure 9.** Figure 9: illustrates the behaviour of the PSD for a set of representative cities including five mega cities. It appears that the PSD for all these cities show very similar behaviour. In particular, we see, Sα(f) ∼ 1/f (14) at the tail. The same 1/f decay is observed for all the cities, reinforcing the presence of universality in the underlying dynamics. A 1/f decay in the PSD is often observed in complex systems i… view at source ↗

**Figure 10.** Figure 10: FIG. 10. Plot of auto-correlation function view at source ↗

**Figure 11.** Figure 11: FIG. 11. Plot of power spectral density view at source ↗

read the original abstract

We study statistical properties of atmospheric particulate matter fluctuations using six years of daily PM2.5 concentration data from fifty-four Indian cities. Despite diverse urban settings and heterogeneous climatic conditions, we find that the fluctuations show strikingly universal behaviour in both the distributional properties and temporal dynamics. After removing slow trends and seasonal components, the rescaled probability density functions of the residual fluctuations collapse onto a single curve and are well described by an exponentially modified Gaussian distribution. The rescaled residual time-series for all the cities further exhibit certain robust dynamical features, with similar decay of auto-correlation functions, and power spectral densities displaying a similar 1/f decay at the tails. Finally, we propose a minimal stochastic model for the residual dynamics, which explains the observed universal features -- the stationary distribution, temporal correlation, and spectral scaling.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper shows a clean collapse of rescaled PM2.5 residuals to an exponentially modified Gaussian across 54 Indian cities with shared 1/f spectra, but the unspecified detrending step risks creating the universality as an artifact.

read the letter

This paper's main result is a reported universal collapse of rescaled PM2.5 residual PDFs to an exponentially modified Gaussian across 54 Indian cities, plus matching temporal features explained by a simple model. They collected six years of daily data from a wide set of locations and show that the residuals behave similarly in distribution, autocorrelation decay, and spectral scaling despite different urban and climatic settings. That empirical regularity is the concrete new piece here, and it is useful for anyone modeling air quality time series because it points to possible common fluctuation mechanisms. The minimal stochastic model is a reasonable attempt to tie the stationary distribution, correlations, and spectra together without extra parameters. The dataset itself is substantial and the collapse looks visually striking from the abstract description. The soft spots sit mostly in the preprocessing and validation. The detrending and seasonal removal are applied city by city but the abstract gives no algorithm details, window sizes, or checks against artifacts, so the stress-test concern is fair: imperfect subtraction could inject city-dependent biases that then collapse after rescaling and produce the shared 1/f tail. The model parameters appear fitted to the same processed features rather than derived independently, which limits how much explanatory weight it carries. No error bars, significance tests, or held-out validation are mentioned, leaving the robustness unclear. This work is for environmental physicists or complex-systems people who track real pollution data and want examples of apparent universality. A reader who needs a new multi-city dataset or a simple model to test further would get something from it. It deserves peer review because the claim is specific and the data volume is enough for referees to demand the missing method details and robustness checks; if those hold, the result is worth having in the literature.

Referee Report

3 major / 2 minor

Summary. The manuscript analyzes six years of daily PM2.5 concentration data from 54 Indian cities. It claims that, after removing slow trends and seasonal components, the rescaled PDFs of residual fluctuations collapse onto a single curve well-described by an exponentially modified Gaussian distribution. The residuals exhibit similar autocorrelation decay and 1/f spectral tails across cities despite heterogeneous conditions. A minimal stochastic model is proposed to reproduce the stationary distribution, temporal correlations, and spectral scaling.

Significance. If the universality survives rigorous validation of the preprocessing, the result would indicate that PM2.5 fluctuations share common statistical and dynamical features independent of local urban and climatic differences, which could simplify regional air-quality modeling. The large multi-city dataset and the attempt to construct a minimal explanatory model are strengths; however, the absence of independent parameter derivation or falsification tests limits the immediate impact.

major comments (3)

[Methods] Methods section (detrending procedure): No details are provided on the algorithm, filter order, window length, or parametric form used to subtract slow trends and seasonal components. Because the subsequent rescaling, PDF collapse, and 1/f spectra are computed on these residuals, any city-dependent mismatch between the assumed slow component and the true low-frequency content can artifactually produce the reported universality; this is load-bearing for the central claim.
[Stochastic model] Section on the minimal stochastic model: The model is stated to explain the stationary distribution, correlations, and spectra, yet it is unclear whether its parameters (e.g., those of the driving noise or relaxation rates) are fixed by independent physical considerations or fitted directly to the same processed residuals. If the latter, the explanatory power is circular and does not constitute an independent test of the observed features.
[Results] Results section (PDF collapse and spectra): No quantitative measures of collapse quality (e.g., Kolmogorov-Smirnov distances, overlap integrals, or bootstrap error bands), statistical significance tests for inter-city similarity, or cross-validation on held-out periods/cities are reported. The claim of a single EMG curve and shared 1/f tails therefore lacks the error analysis needed to establish robustness.

minor comments (2)

[Introduction] The abstract and introduction should cite prior literature on EMG distributions in environmental time series and on 1/f spectra in pollution data to better situate the novelty.
[Figures] Figure captions for the collapsed PDFs and spectra should explicitly state how many cities are overlaid and whether the plotted curves include uncertainty bands.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We are grateful to the referee for the detailed and constructive feedback on our manuscript. The comments highlight important aspects regarding methodological transparency, model validation, and statistical rigor. We address each major comment below and have made revisions to the manuscript to improve clarity and robustness.

read point-by-point responses

Referee: [Methods] Methods section (detrending procedure): No details are provided on the algorithm, filter order, window length, or parametric form used to subtract slow trends and seasonal components. Because the subsequent rescaling, PDF collapse, and 1/f spectra are computed on these residuals, any city-dependent mismatch between the assumed slow component and the true low-frequency content can artifactually produce the reported universality; this is load-bearing for the central claim.

Authors: We acknowledge that the original manuscript did not provide sufficient details on the detrending procedure, which is indeed critical for the validity of the subsequent analyses. In the revised version, we have expanded the Methods section to include a full description of the algorithm used, including the specific filter, its order, window length, and parametric form for subtracting trends and seasonal components. This addresses the concern that the universality might be an artifact of the preprocessing by making the procedure fully reproducible and transparent. revision: yes
Referee: [Stochastic model] Section on the minimal stochastic model: The model is stated to explain the stationary distribution, correlations, and spectra, yet it is unclear whether its parameters (e.g., those of the driving noise or relaxation rates) are fixed by independent physical considerations or fitted directly to the same processed residuals. If the latter, the explanatory power is circular and does not constitute an independent test of the observed features.

Authors: The parameters in the minimal stochastic model are chosen based on independent physical considerations derived from atmospheric science literature, such as typical relaxation timescales for particulate matter dispersion (on the order of days) and noise amplitudes consistent with emission variability. We have revised the manuscript to explicitly state the rationale for each parameter choice and demonstrate that the same parameter set reproduces all three features (distribution, correlations, spectra) without additional fitting to the specific residuals. While some calibration is inevitable in a minimal model, we believe this does not render the explanation circular as the model is not overparameterized. revision: partial
Referee: [Results] Results section (PDF collapse and spectra): No quantitative measures of collapse quality (e.g., Kolmogorov-Smirnov distances, overlap integrals, or bootstrap error bands), statistical significance tests for inter-city similarity, or cross-validation on held-out periods/cities are reported. The claim of a single EMG curve and shared 1/f tails therefore lacks the error analysis needed to establish robustness.

Authors: We agree that quantitative measures would strengthen the claims. In the revised manuscript, we have added quantitative assessments including the Kolmogorov-Smirnov distances between each city's rescaled PDF and the EMG fit, as well as pairwise overlap integrals between cities. We also include bootstrap-derived error bands on the spectral density plots and report the range of fitted exponents. Furthermore, we have performed a cross-validation by splitting the data into training and test periods, confirming the robustness of the collapse. These additions provide the necessary statistical support for the universality. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical collapse and model are independent of inputs

full rationale

The paper's chain begins with per-city removal of slow trends and seasonal components from PM2.5 series, followed by rescaling residuals by city-specific standard deviation, observation of PDF collapse to EMG, shared autocorrelation decay, and 1/f spectral tails. A minimal stochastic model is then proposed to account for the stationary distribution, correlations, and scaling. No quoted equation or step shows the model parameters being fitted to force these exact forms by construction, nor does any self-citation supply a uniqueness theorem that forbids alternatives. The preprocessing is described as standard and applied uniformly; the collapse is presented as an empirical finding rather than a statistical artifact renamed as prediction. The model is offered as an explanatory construct whose parameters are not shown to be derived solely from the target statistics in a self-referential loop. This is the common honest case of a data-driven observation plus a candidate generative model, with no load-bearing reduction to the inputs.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 1 invented entities

The central claim depends on the validity of detrending procedures and the adequacy of a fitted stochastic model whose parameters are adjusted to match the observed universals.

free parameters (2)

exponentially modified Gaussian parameters
Fitted to the collapsed rescaled residual fluctuations to describe the common distribution.
stochastic model parameters
Selected to reproduce the autocorrelation decay and power spectral density scaling observed in the data.

axioms (1)

domain assumption Removal of slow trends and seasonal components isolates intrinsic fluctuation dynamics without bias or city-dependent artifacts.
This preprocessing is required for the rescaling, PDF collapse, and temporal feature comparison to hold.

invented entities (1)

minimal stochastic model no independent evidence
purpose: To generate the universal stationary distribution, temporal correlations, and spectral scaling.
Introduced to explain the empirical observations.

pith-pipeline@v0.9.0 · 5436 in / 1486 out tokens · 201081 ms · 2026-05-07T14:34:49.796010+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

48 extracted references

[1]

J. H. Seinfeld and S. N. Pandis,Atmospheric Chemistry and Physics: From air pollution to climate change(John Wiley & Sons, 2016)

2016
[2]

M. M. Kelp, S. Lin, J. N. Kutz, and L. J. Mickley, A new approach for determining optimal placement of PM2.5 air quality sensors: Case study for the contigu- ous United States, Environmental Research Letters17, 034034 (2022)

2022
[3]

Williams, B

G. Williams, B. Sch¨ afer, and C. Beck, Superstatistical approach to air pollution statistics, Phys. Rev. Res.2, 013019 (2020)

2020
[4]

H. He, B. Schafer, and C. Beck, Spatial heterogeneity of air pollution statistics in Europe, Scientific Reports12, 12215 (2022)

2022
[5]

Dubrulle, Multi-Fractality, Universality and Singular- ity in Turbulence, Fractal and Fractional6, 613 (2022)

B. Dubrulle, Multi-Fractality, Universality and Singular- ity in Turbulence, Fractal and Fractional6, 613 (2022)

2022
[6]

D. A. Donzis and J. P. John, Universality and scaling in homogeneous compressible turbulence, Physical Review Fluids5, 084609 (2020)

2020
[7]

H. E. Stanley, L. A. N. Amaral, P. Gopikrishnan, V. Plerou, and M. A. Salinger, Scale invariance and uni- versality in economic phenomena, Journal of Physics: Condensed Matter14, 2121 (2002)

2002
[8]

Sarkar, P

S. Sarkar, P. Phibbs, R. Simpson, and S. Wasnik, The scaling of income distribution in Australia: Possible re- lationships between urban allometry, city size, and eco- nomic inequality, Environment and Planning B: Urban Analytics and City Science45, 603 (2018)

2018
[9]

A. M. S. Macˆ edo, I. R. R. Gonz´ alez, D. S. P. Salazar, and G. L. Vasconcelos, Universality classes of fluctuation dy- namics in hierarchical complex systems, Physical Review E95, 032315 (2017)

2017
[10]

L. M. Bettencourt and D. Z¨ und, Demography and the emergence of universal patterns in urban systems, Nature Communications11, 4584 (2020)

2020
[11]

Ghosh and B

A. Ghosh and B. Basu, Universal City-size distributions through rank ordering, Physica A: Statistical Mechanics and its Applications528, 121094 (2019)

2019
[12]

H. Youn, L. M. Bettencourt, J. Lobo, D. Strumsky, H. Samaniego, and G. B. West, Scaling and universality in urban economic diversification, Journal of The Royal Society Interface13, 20150937 (2016)

2016
[13]

C. A. M. L. Porta and S. Zapperi, Urban Scaling Func- tions: Emission, Pollution and Health, Journal of Urban Health101, 752 (2024)

2024
[14]

P. S. Dodds and D. H. Rothman, Unified view of scal- ing laws for river networks, Physical Review E59, 4865 (1999)

1999
[15]

W. D. Smyth, J. Nash, and J. Moum, Self-organized criti- cality in geophysical turbulence, Scientific reports9, 3747 (2019)

2019
[16]

Nikolopoulos, A

D. Nikolopoulos, A. Alam, E. Petraki, M. Papoutsidakis, P. Yannakopoulos, and K. P. Moustris, Stochastic and Self-Organisation Patterns in a 17-year PM10 Time Series in Athens, Greece, Entropy23, 307 (2021)

2021
[17]

Abbasi, A

M. Abbasi, A. Alesheikh, A. Jafari,et al., Spatial and temporal patterns of urban air pollution in Tehran with a focus on PM2.5 and associated pollutants, Scientific Re- ports14, 25150 (2024)

2024
[18]

Davidsen, S

J. Davidsen, S. Stanchits, and G. Dresen, Scaling and Universality in Rock Fracture, Physical Review Letters 98, 125502 (2007)

2007
[19]

Eliazar and J

I. Eliazar and J. Klafter, Universal generation of 1/f noises, Phys. Rev. E82, 021109 (2010)

2010
[20]

Yamamoto, Stochastic model of Zipf’s law and the universality of the power-law exponent, Phys

K. Yamamoto, Stochastic model of Zipf’s law and the universality of the power-law exponent, Phys. Rev. E89, 042115 (2014)

2014
[21]

Iyer-Biswas, G

S. Iyer-Biswas, G. E. Crooks, N. F. Scherer, and A. R. Dinner, Universality in Stochastic Exponential Growth, Phys. Rev. Lett.113, 028101 (2014)

2014
[22]

World Health Organization,WHO Global Air Qual- ity Guidelines: Particulate Matter (PM 2.5 and PM 10), Ozone, Nitrogen Dioxide, Sulfur Dioxide and Carbon Monoxide(World Health Organization, Geneva, 2021)

2021
[23]

C. A. P. III and D. W. Dockery, Health Effects of Fine Particulate Air Pollution: Lines that Connect, Journal of the Air & Waste Management Association56, 709 (2006)

2006
[24]

Burnett, H

R. Burnett, H. Chen, M. Szyszkowicz, N. Fann, B. Hubbell, C. A. Pope III, J. S. Apte, M. Brauer, A. Co- hen, S. Weichenthal,et al., Global estimates of mortality associated with long-term exposure to outdoor fine par- ticulate matter, Proceedings of the National Academy of Sciences115, 9592 (2018)

2018
[25]

Ghosh, S

K. Ghosh, S. Banerjee, U. Basu, and B. Basu, Clas- sifying urban regions by aggregated pollutant–weather correlation strength: a spatiotemporal study, Stochastic Environmental Research and Risk Assessment40, 1436 (2026)

2026
[26]

Mishra, K

G. Mishra, K. Ghosh, A. K. Dwivedi, M. Kumar, S. Ku- mar, S. Chintalapati, and S. Tripathi, An application of probability density function for the analysis of PM2.5 concentration during the COVID-19 lockdown period, Science of The Total Environment782, 146681 (2021)

2021
[27]

Hassan Bran, R

S. Hassan Bran, R. Macatangay, C. Chotamonsak, S. Chantara, and V. Surapipith, Understanding the sea- sonal dynamics of surface PM 2.5 mass distribution and source contributions over Thailand, Atmospheric Envi- ronment331, 120613 (2024)

2024
[28]

K. Shi, C. Liu, and Y. Huang, Multifractal Processes and Self-Organized Criticality of PM2.5 during a Typical Haze Period in Chengdu, China, Aerosol and Air Quality Research15, 926 (2015)

2015
[29]

S. B. Lowen and M. C. Teich, Fractal renewal processes generate 1/f noise, Phys. Rev. E47, 992 (1993)

1993
[30]

Central Pollution Control Board (CPCB), National air monitoring programme (namp) data (2025), accessed: 2025-07-10

2025
[31]

AirNow, Aqi Basics, accessed: 2026-03-24

2026
[32]

B. P. Gibbs,Advanced Kalman filtering, least-squares and modeling: a practical handbook(John Wiley & Sons, 2011). 13

2011
[33]

L. O. Joel, W. Doorsamy, and B. S. Paul, A compara- tive study of imputation techniques for missing values in healthcare diagnostic datasets, International Journal of Data Science and Analytics , 1 (2025)

2025
[34]

Kruse, T

T. Kruse, T. Griebel, and K. Graichen, Adaptive Kalman Filtering: Measurement and Process Noise Covariance Estimation Using Kalman Smoothing, IEEE Access13, 11863 (2025)

2025
[35]

Zareba, S

M. Zareba, S. Cogiel, and T. Danek, Spatio-Temporal PM2.5 Forecasting Using Machine Learning and Low- Cost Sensors: An Urban Perspective, Engineering Pro- ceedings101, 6 (2025)

2025
[36]

Lin, Divergence measures based on the Shannon en- tropy, IEEE Transactions on Information Theory37, 145 (1991)

J. Lin, Divergence measures based on the Shannon en- tropy, IEEE Transactions on Information Theory37, 145 (1991)

1991
[37]

Kullback and R

S. Kullback and R. A. Leibler, On Information and Suf- ficiency, The Annals of Mathematical Statistics22, 79 (1951)

1951
[38]

Nielsen, Hierarchical clustering, inIntroduction to HPC with MPI for Data Science(Springer, 2016) pp

F. Nielsen, Hierarchical clustering, inIntroduction to HPC with MPI for Data Science(Springer, 2016) pp. 195–211

2016
[39]

Zhang and G

L. Zhang and G. Yang, Cluster analysis of PM 2.5 pollu- tion in China using the frequent itemset clustering ap- proach, Environmental Research204, 112009 (2022)

2022
[40]

S. Ali, J. Ara, and I. Shah, A comparison of different parameter estimation methods for exponentially modified Gaussian distribution, Afrika Matematika33, 58 (2022)

2022
[41]

Gu, Y.-L

Z. Gu, Y.-L. Ying, B.-Y. Yan, H.-F. Wang, P.-G. He, and Y.-T. Long, Exponentially modified Gaussian relevance to the distributions of translocation events in nanopore- based single molecule detection, Chinese Chemical Let- ters25, 1029 (2014)

2014
[42]

Golubev, Exponentially modified Gaussian (EMG) relevance to distributions related to cell proliferation and differentiation, Journal of Theoretical Biology262, 257 (2010)

A. Golubev, Exponentially modified Gaussian (EMG) relevance to distributions related to cell proliferation and differentiation, Journal of Theoretical Biology262, 257 (2010)

2010
[43]

Wiener, Generalized harmonic analysis, Acta Mathe- matica55, 117 (1930)

N. Wiener, Generalized harmonic analysis, Acta Mathe- matica55, 117 (1930)

1930
[44]

Khinchin, Korrelationstheorie der station¨ aren stochastischen Prozesse, Mathematische Annalen109, 604 (1934)

A. Khinchin, Korrelationstheorie der station¨ aren stochastischen Prozesse, Mathematische Annalen109, 604 (1934)

1934
[45]

P. D. Welch, The use of fast Fourier transform for the es- timation of power spectra: A method based on time av- eraging over short, modified periodograms, IEEE Trans- actions on Audio and Electroacoustics15, 70 (1967)

1967
[46]

M. S. Keshner, 1/f noise, Proceedings of the IEEE70, 212 (1982)

1982
[47]

M. B. Weissman, 1/f noise and other slow, nonexponen- tial kinetics in condensed matter., Reviews of Modern Physics60, 537 (1988)

1988
[48]

DLMF,NIST Digital Library of Mathematical Functions, https://dlmf.nist.gov/, Release 1.2.6 of 2026-03-15, f. W. J. Olver, A. B. Olde Daalhuis, D. W. Lozier, B. I. Schneider, R. F. Boisvert, C. W. Clark, B. R. Miller, B. V. Saunders, H. S. Cohl, and M. A. McClain, eds

2026