Frequent new particle formation events in the Indo-Gangetic Plain occur under reduced condensation sink but are obscured by air mass heterogeneity

Srivastav, Gaurav K.; Wagay, Aasif A.; Lampilahti, Janne; Rathore, Jaswant; Kunchala, Ravi K.; Ganguly, Dilip; Kulmala, Markku; Petäjä, Tuukka; Paasonen, Pauli; Thakur, Roseline C.; Gani, Shahzad

doi:10.5194/egusphere-2026-2212

Preprints

https://doi.org/10.5194/egusphere-2026-2212

Preprints

15 Jun 2026

| 15 Jun 2026

Status: this preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).

Frequent new particle formation events in the Indo-Gangetic Plain occur under reduced condensation sink but are obscured by air mass heterogeneity

Gaurav K. Srivastav, Aasif A. Wagay, Janne Lampilahti, Jaswant Rathore, Ravi K. Kunchala, Dilip Ganguly, Markku Kulmala, Tuukka Petäjä, Pauli Paasonen, Roseline C. Thakur, and Shahzad Gani

Abstract. New particle formation (NPF) is a major source of atmospheric aerosols and cloud condensation nuclei, influencing climate and air quality. In highly polluted regions such as the Indo-Gangetic Plain (IGP), where precursor concentrations and condensation sinks are among the highest globally, NPF remains poorly constrained due to limited observations of particles and ions in sub–10 nm size range.

Here, we investigated the occurrence and microphysical evolution of NPF at a suburban IGP site (CAS-AO, Sonipat, Haryana), during May – December 2023 using ion and particle measurements together with meteorological data. NPF events are frequent during summer and less common in winter. The weak seasonal variability of ion concentrations relative to nucleation-mode particles suggests that neutral pathways dominate NPF in this high-sink environment. The median condensation sinks on event-days (0.024 s⁻¹) are approximately half those on non-event days (0.046 s⁻¹). Particle growth rates (maximum concentration method) increase with size, from 14.7 nm h⁻¹ (3–7 nm) to 19.0 nm h⁻¹ (7–20 nm), indicating size-dependent condensational growth.

In this high-background setting, pollution plumes, and meteorological variability intermittently mask or distort NPF signals, limiting the direct applicability of both visual and automated classification methods and growth rate estimation methods. Our observations highlight that careful application and improvement of data analysis methods, along with precursor gas measurements, are required to better constrain nucleation processes by avoiding methodological sensitivity due to the complexity of aerosol dynamic processes in multi-source, high-condensation-sink environments, such as IGP.

Received: 17 Apr 2026 – Discussion started: 15 Jun 2026

Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Chemistry and Physics.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 2343 KB)

Supplement (950 KB)

Download & links

Gaurav K. Srivastav, Aasif A. Wagay, Janne Lampilahti, Jaswant Rathore, Ravi K. Kunchala, Dilip Ganguly, Markku Kulmala, Tuukka Petäjä, Pauli Paasonen, Roseline C. Thakur, and Shahzad Gani

Status: open (until 27 Jul 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-2212', Anonymous Referee #1, 22 Jun 2026 reply
General assessment
Srivastav present PNSD measurements across almost a year in the IGP, focusing on the occurence of NPF events. These are valuable data. India has some of the worst air quality on the planet, and the question of whether secondary formation contributes meaningfully to particle number there is not answered. In China it turned out to matter a great deal, and a well-characterised sub–10 nm dataset from a suburban IGP site is highly valuable. In this context, I think this dataset is worth publishing.
My main concern is one of framing. As it stands, the manuscript reads as though it's working quite hard to present a clean, finished story, and in doing so it skirts around some real limitations in the dataset (data coverage, SMPS availability, pollution contamination of the growth). My view is that being upfront about those limitations would make this a stronger paper, not a weaker one. I'd encourage the authors to lean into that rather than away from it.
The points below are split into a few major issues that I think need resolving before publication, then specific comments, then a general note on the writing.
Major issues
SMPS data availability
This is my biggest concern. Late in the paper (Sect. 3.2.2) it emerges that SMPS data overlapping with the NAIS were available for only ~11 NPF days, of which 3 were usable. This is also clear if the reader is savvy as it is mentioned in sect. 2.3 (“Data for the above 40 nm diameter range from SMPS were available for only a few days (mention in table T1)”, and viewing T1 only shows three events, but this is far from clear!
The condensation sink, the coagulation sink, and the formation rates all require the full size distribution out to the accumulation mode and in a polluted environment CS and CoagS are dominated by the >40 nm particles that only the SMPS sees. So I would strongly encourage the authors to state, explicitly and early:
For how many days, and which days, is the SMPS actually available?

Is the event-day vs non-event CS comparison (the 0.024 vs 0.046 /s result that the title and abstract are built on) computed over the full record, or only over those few SMPS days?

If the CS conclusion rests on three days, then I'm afraid the central claim of the paper isn't currently supported and needs to be either properly substantiated or removed. If the SMPS coverage is a bit thin, one honest option is to reframe this as a NAIS-led paper, present CS/CoagS/J only for the days where they're defensible (clearly flagged as such), and be explicit that "total particles" actually means "sub 40 nm particles" everywhere the SMPS is absent. None of that is fatal to the paper but the current presentation implies a completeness the data may not have.
Data coverage is never made clear.
Similar to my above point, the measurements run May through December, yet Sect. 3.2.1 implies only 108 data days. While this is a lot of missing time, our instruments are indeed temperamental, so this is quite normal. However, the reader is left guessing where the gaps are and which instruments were running when. I would encourage the authors to open the results with a full time-series / data-availability plot showing NAIS and SMPS coverage across the whole period, alongside the MET data, and anything else appropriate. This single figure would resolve a lot of the ambiguity and would let readers judge the seasonal claims for themselves.
Clarity regarding GRs
Median GRs of ~15 nm/h in the 3–7 nm range (rising to ~19–27 nm/h at larger sizes) are at the very top of anything reported anywhere. If real, that's a very important result with real consequences, as the particle number scales with J, but both the mass of the growing mode and the survival probability are extremely sensitive to GR, so growth this fast could be a major local source of both number and mass.
But the authors themselves acknowledge that pollution plumes intermittently mimic and inflate growth signatures, so I'd want much more reassurance that these GRs aren't contaminated. It would be great if the authors could both
State clearly which classification scheme (Dal Maso or Dada) is used to select the events that feed the GR calculation, how many events that is, and why.

Given the contamination risk, please show some peak fits and some diameter-versus-time plots so the reader can judge whether the growing mode is being cleanly separated from pollution.

Following from the high GRs, I would encourage you to discuss the potential contribution of these events to particle mass, not just number (see Kulmala et al., 2022, https://doi.org/10.1039/D1EA00096A). Estimating mass from the PNSD would be straightforward for such events where SMPS data exist.
Condensation sink equation (line 257).
Please double-check this. The standard formulation is CS = 2πD .... The manuscript has 4πD, which is a factor-of-two error when Dp is a diameter.
Specific comments
Abstract
L13: "in the sub–10 nm size range".

L14: "microphysical evolution" what does microphysical mean here specifically? Either define it or use a plainer phrase.

The frequency of NPF is a headline finding, so give the summer and winter frequencies in the abstract rather than just "frequent in summer, less common in winter". Likewise, since the sub–10 nm gap is a central motivation, state the diameter ranges you actually measured when you say "using ion and particle measurements".

Introduction
L40: a single percentage range can't apply to both total particle number and CCN concentrations as they're different quantities. Please separate them.

L43: "At the nanoscale level" is ambiguous; be specific about the size regime you mean.

L62: I think you mean the "dynamics of particles in the 3–10 nm range", rather than "clusters below 3–10 nm".

L64: if the seasonal contrast isn't yet understood, how is it being asserted as significant? Reword.

L66 onwards: NPF gets formally introduced here, after its impacts have already been discussed. The intro would flow better if NPF were defined earlier. This passage also carries a lot of unnecessary detail, e.g. "NPF emerges as a pivotal process where gas-phase molecules converge to initiate the formation of solid or liquid particles" can go.

L70: "As these nascent particles grow into the Aitken mode, their number concentration often increases further." This isn't a property of NPF, the number in a growing population generally falls (coagulation, dilution). What increases is the observed concentration in a fixed size band as the mode passes through it. That's an observational effect, and the sentence should be reframed accordingly.

Nucleation mechanisms paragraph: I'd point the authors to Kulmala et al., 2023 (https://doi.org/10.1038/s41561-023-01305-0), which is the current reference for how to frame this. Given that you’re observing NPF at 30 degrees Celsius I doubt you’re seeing pure SA-NH3, but I’d bet you’re got a tonne of much stronger bases.

"In the overall aerosol loading, NPF contribution… makes a significant part of that": this is overstated as a general claim. It's true at some sites and not others; please qualify.

Instrumentation (2.2.1, 2.2.2)
Inconsistent treatment of the two instruments: detailed uncertainties are given for the SMPS but not the NAIS. Please balance this (there are some estimates available for NAIS uncertainty, but not as many as for SMPS).

Several things are missing: did the NAIS and SMPS agree in their overlapping size range, and if not, what correction was applied? How was the SMPS inversion performed, and which version of AIM was used? Was retrospect used to invert the NAIS data?

L160: "Multiple instruments are clubbed together" please use more formal phrasing.

L186: pedantic, but the correct quantity is the particle number size distribution (PNSD); please use this throughout.

Figure 2
The colour scales are mismatched: particles span ~2 orders of magnitude while ions span ~5. As drawn, this makes it look as though there are NPF events with no corresponding ions, which I don't think is your intended message. Please harmonise the scaling (or at least pull down that upper range for the ions a bit!)

Panel (a) doesn't directly show particle number concentrations, and if there's no SMPS support then this isn't "total particles" but sub 40 nm particles. Label accordingly. This applies to all discussion.

Figure 3 and associated text
The choice of size windows is a bit odd. 3–5 nm and then 7–12 nm. Why not 3–7 nm, especially given that's the window used for the growth-rate analysis? Please make the size windows consistent across analyses, or justify why they differ.

Figure 4 / classification
Internal inconsistency to fix: the Discussion (around L723) quotes GRs of "~9.7–11.4 nm/h (3–7 nm)" and "~12.9–15.6 nm/h (7–20 nm)", which don't match the medians reported in the Results and abstract (10.6 / 14.7 and 16.8 / 19.0 nm/h). The numbers need to be reconciled.

Figure 6
Needs tidying (de-squishing) before publication; it's hard to read at present.

Since the visual classification is acknowledged to be subjective, it might be worth plotting the mean Class I and Class II events as heatmaps, so readers can see what an average event of each class actually looks like.

Figure 9
I'd strongly recommend estimating particle mass from the PNSD and adding it to this figure (not obligatory!).

Section 3.2 / organisation
L579: we've suddenly returned to a discussion of Dal Maso here, well after it was introduced. The manuscript's organisation needs work overall. The classification discussion is scattered, and a tighter structure would help the reader a lot as you could describe the methods, state which one is best, and move forwards with it.

Figure 10
There looks to be a time offset between the two instruments. The SMPS appears to lag the NAIS by roughly an hour. Please check the timestamps and synchronisation.

The SMPS also looks as though it may be under-reading particle concentrations on these days; worth verifying (the NAIS is known to over-read, so maybe that’s the more appropriate statement).

Figures 11 and 12
These really should be combined into a single figure with a shared, aligned time axis. As it stands they're on separate pages, so it's impossible to read straight down from the aerosol/sink panels to the meteorology.

Table 1
I don't think the J values are valid without the supporting SMPS data: without the full distribution the CoagS is too small and J is therefore underestimated. This ties back to Major Issue 1. Or are these only the days where the SMPS was available?

For what subset of days is the condensation sink actually calculated? If it's only these three days, then the paper's central CS conclusion rests on three points and needs to be reconsidered.

The quantity names aren't in the header row, and inconsistent font sizes are used. Please clean this up.

A general note on the writing
The writing needs a careful check throughout. A few things crop up repeatedly
singular/plural slips ("these cluster" should be "these clusters"),

subject–verb disagreement ("ion concentrations does not show", "are play a more important role"),

missing articles ("in sub–10 nm size range", "from NAIS negative channel"),

There's also some wordiness that could be trimmed hard, e.g. "facilitating the observation of changes in resultant particle populations".
Also in the references there appear to be two separate Aalto et al. (2001) entries; please de-duplicate and check the reference list generally.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-2212-RC1
RC2: 'Comment on egusphere-2026-2212', Anonymous Referee #2, 01 Jul 2026 reply

My comments are given in the accompanying pdf file.

Reply

Citation: https://doi.org/10.5194/egusphere-2026-2212-RC2

Gaurav K. Srivastav, Aasif A. Wagay, Janne Lampilahti, Jaswant Rathore, Ravi K. Kunchala, Dilip Ganguly, Markku Kulmala, Tuukka Petäjä, Pauli Paasonen, Roseline C. Thakur, and Shahzad Gani

Supplement

https://doi.org/10.5194/egusphere-2026-2212-supplement

Gaurav K. Srivastav, Aasif A. Wagay, Janne Lampilahti, Jaswant Rathore, Ravi K. Kunchala, Dilip Ganguly, Markku Kulmala, Tuukka Petäjä, Pauli Paasonen, Roseline C. Thakur, and Shahzad Gani

Metrics will be available soon.

Latest update: 05 Jul 2026

Short summary

Particle formation events affect climate and air quality. Using measurements of particle number concentrations down to very small particles (~3 nm) in polluted northern India, we show that these events occur frequently in summer under reduced background particle load. However, in this environment, emissions and changing air masses can mask or distort these events, making them difficult to detect. Such environments require adaptation of standard analytical methods used in cleaner regions.