the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Jun 2025
Aerosols drive monsoon rainfall spatial modulations over the Indian subcontinent: anthropogenic and dust aerosols impact, mechanism, and control
Abstract. Spatial modulations in monsoon rainfall over the Indian subcontinent, characterized by persistent weakening and strengthening patterns, are yet to be understood in the context of the role of spatial heterogeneity in aerosol species (anthropogenic and dust) radiative perturbations, driving mechanisms, and control. Current inaccuracies in modelling the aerosol species burden in this region have posed challenges to fully addressing these complexities. Here, we successfully simulate for the first time the aerosol-driven impact on monsoon rainfall spatial modulations, adequately accounting for aerosol distributions in a fine-resolved (25×25 km2) regional climate model. The modelled aerosol-induced spatial modulations align consistently with the measured departures in monsoon rainfall. The aerosol-induced weakening of the rainfall (30 %–50 %) over most of the Indian subcontinent, with a maximum deficiency over the eastern coast (−48 mm), is primarily driven by changes in regional wind dynamics induced by anthropogenic aerosol all-sky radiative forcing. The rainfall increase (>50 %) is strengthened by all-sky radiative warming with dust aerosols over most of central/northwestern India and the western coast. Abatement of anthropogenic aerosols can largely mitigate the rainfall deficiency, but by 30 % to 40 % only over the eastern coast; thereby also identifying areas of augmented rainfall excessiveness (e.g., Andhra Pradesh/Gujarat) or their mitigation (e.g., Kerala/northeastern India) driven by anthropogenic aerosol control. These insights are crucial for developing effective water management strategies in the region.
- Preprint
(14433 KB) - Metadata XML
-
Supplement
(1609 KB) - BibTeX
- EndNote
Status: open (extended)
-
CC1: 'Comment on egusphere-2025-2302', R H Kripalani, 22 Jun 2025
reply
The manuscript appears very well written , contains useful scientific information. I don't understand why a discussion is required on this well written manuscript. Recommended for publication
Citation: https://doi.org/10.5194/egusphere-2025-2302-CC1 -
AC1: 'Reply on CC1', Shubha Verma, 23 Jun 2025
reply
Dear Dr Kriplani,
Thank you for your encouraging comments.
Regards
Shubha Verma
(Corresponding author)
Citation: https://doi.org/10.5194/egusphere-2025-2302-AC1
-
AC1: 'Reply on CC1', Shubha Verma, 23 Jun 2025
reply
-
RC1: 'Comment on egusphere-2025-2302', Anonymous Referee #1, 04 Aug 2025
reply
Review of “Aerosols drive monsoon rainfall spatial modulations over the Indian subcontinent: anthropogenic and dust aerosols impact, mechanism, and control” by Sauvik Santra and co-authors.
In this study the authors simulate the response of Indian monsoon rainfall due to regional aerosols. This is achieved by contrasting simulations with and without aerosols. The first analysis uses output from a coarse-resolution global model (IPSL-CM6A-LR) and focuses on the separating the precipitation responses from GHG emissions and aerosols. A second analysis uses a set of simulations performed using a regional configuration of WRF with a detailed representation of present-day aerosols constrained by AOD observations. The aerosols are separated into contributions from dust and anthropogenic sources. Simulations are run with and without aerosols. These simulations are used to try and reproduce/explain recent trends in observed rainfall between 2000 and 2019 over the Indian subcontinent and anomalous rainfall over the same period relative to long-term observations (previous 50 years). Herein lies a potentially fundamental flaw in this study – these observations are representative of relatively recent decades (back to 1970s), yet the simulations are either used to focus on the anthropogenic aerosol effect (so pre-industrial, not 1970) or total aerosol effect (so not applicable to any realistic scenario).
Throughout the manuscript the authors make note of the consistency between the simulated patterns for ‘all aerosols’ and the observed anomaly patterns but this is not comparing like for like. If they just used pre-industrial aerosol vs present-day aerosol, or only focus on the anthropogenic aerosol response, then this may be a useful comparison, but by removing all aerosol and focusing on this response only they are implicitly assuming that all regional aerosol over India in the present-day is due to anthropogenic activity. This is clearly unrealistic – dust and sea-spray emissions are unlikely to have changed significantly due to anthropogenic activity (and others..). The correct way to make this comparison is to use an aerosol scenario that is consistent with overlapping regional AOD/aerosol concentrations from the past 50 years.
There are other issues with this manuscript that I outline below, but for this primary reason I do not recommend publication. There are fundamental flaws in this manuscript that need to be addressed. I do however commend the authors on their ARS model setup – having a realistic aerosol distribution with distinct species is an excellent tool to focus on aerosol impacts to rainfall over the Indian subcontinent. One option for the authors is to use the ‘all aerosol’ simulations to gain an overall perspective on the aerosol-monsoon interactions, but then focus most of the results on the difference between including and excluding anthropogenic aerosols.
Major comments
Primary comment as per above.
Lack of necessary detail on model configuration. For the CMIP6 simulations – were the two simulations free-running or nudged? How were aerosols represented? Are direct and indirect effects represented? For the ARS simulations – what were used as boundary conditions? Were SSTs prescribed? How long was each simulation run for? Was it nudged against reanalysis meteorology? All of these details are considered standard and should be included.
Much more detail is needed to understand how aerosols are coupled to the cloud microphysics scheme in the ARS simulations. Are aerosols activated as CCN and coupled to the cloud droplet number concentration? How is precipitation represented? Is aerosol scavenging represented in the model? Are aerosol emissions interactive in the model or is the aerosol distribution prescribed?
How does sea spray aerosol feature in this study? It is mentioned as being part of the total aerosol scenario but never referred to afterwards.
The manuscript is difficult to read in places and could be substantially more succinct. For example, the following sentence:
“To assess the aerosol-induced spatial modulations in rainfall at a fine-grid scale over the Indian subcontinent evaluating the role of dust and anthropogenic aerosols, we examine the estimates from the designed ARS in a fine-grid resolved aerosol climate model. To represent the aerosol-driven modulations on monsoon-rainfall as realistically as possible, accounting for the aerosol-meteorological interactions, the change in modelled monsoon-rainfall due to aerosols is estimated finding out the difference between modelled rainfall considering the atmosphere with aerosols and that without aerosols or no-aerosols.”
could be alternatively written as “We now focus on the fine-grid scale ARS simulations.” Additionally, some of the section titles are very long and references to sections should be included as numbered items rather than the entire section title.
Minor comments
Line 41. Absorbing aerosols also produce localized heating of the aerosol layer
Line 124. What is Sul-ows?
Line 144. Shallow convection is parameterized – but does a 25km resolution configuration resolve deep convection?
Line 154. R1 and R2 haven’t been defined yet.
Line 160. Example of repeated information – the four scenarios have already been defined.
Line 163. Clear-sky rainfall and wind responses. Which figure is this referring to? If it S3 then the figure caption needs more information – no mention of clear-sky.
Line 193. The percentage departure is not clear. Do you end up with five data points per grid cell? Or 20? Or something else?
Line 193. Given much of this manuscript depends on getting the right spatial pattern of observed anomalies, please can you include some statistical information on the uncertainty/robustness of the values (in Figure 1b).
Line 197. Table S1 is incomplete.
Line 200. This is the fourth time that IMD has been defined – I suggest sticking to one!
Line 203. Ullah et al (2022) https://www.mdpi.com/2072-4292/14/13/3219 present a very different spatial pattern. Is this because of the years being focused on (2000 to 2019 in this study, 1980 to 2020 in Ullah study)? If so does that have implications for this study?
Line 211. OPTSIM already defined.
Line 224. The authors note the consistent spatial pattern in dust+anthro AOD and recent precipitation trends. But what is the recent trend in dust+anthro AOD? This is what is needed to make such a comparison sensible.
Line 231. Is this using the ARS simulations?
Line 246. Other studies have looked at ACI and ARI impact on surface fluxes over India - how do these values compare?
Line 260. Example of a long section title that could be reduced.
Line 265. Why didn’t the authors use the same years as the observations (2000 onwards?)
Line 273. Again – comparing the spatial pattern of the simulated total aerosol response to observed anomalies from recent decades. This is illogical.
Line 282. “Thereby indicating the regional spatial modulations observed in IMD-rainfall (trend and departures) are in general well represented by the spatial changes (weakening and strengthening) in rainfall imposed due to aerosol during present years” This claim cannot be made using these simulations.
Line 288. Suggested section title: ARS simulations.
Line 316. Why don’t the dust and anthropogenic responses linearly combine in the total aerosol response? Why does anthropogenic dominate to such an extent?
Line 316. “The spatial advancement over central India of aerosol-induced increasing monsoon from dust aerosols as obtained from ARS is also supported by the observational studies inferring an increasing trend in heavy to extremely heavy rainfall events in the central Indian region in recent decades”. But has dust actually increased over the recent decades?
Line 326. It looks like Kripalani et al. 2022 found an increase in precipitation alongside a decrease in aerosol – which points towards aerosols supressing precipitation. Isn’t this therefore at odds with your results?
Line 338. The probability distribution for the response over R1 for all aerosols (black solid line) is almost entirely positive, yet the region marked as R1 on Figure 3d is almost entirely negative apart from a small patch near the centre. Are both plots using the same data? If so, I am not sure how this result is possible.
Line 360. The section ‘Comparison of modelled aerosol induced rainfall change with IMD-rainfall: statistical analysis’ does not seem to exist in the SI.
Line 363. “A reasonably good comparison of measured departures with the modelled aerosol-induced spatial modulations indicates that the measured spatial modulations (weakening and strengthening as departures) in the monsoon-rainfall are driven by regional aerosols over the Indian subcontinent.” You cannot make this claim with the simulations you have performed.
Line 393. Should this be Figure 3?
Line 415. Doesn’t the northwestern region show a decrease in water vapour?
Line 428. Why is there such a strong signal in OLR on the east coast where there is not so much dust?
Line 429. Figure 3f?
Line 436. Can you confirm the impact on convection by looking at vertical winds and/or CAPE etc?
Citation: https://doi.org/10.5194/egusphere-2025-2302-RC1 -
AC2: 'Reply on RC1', Shubha Verma, 11 Aug 2025
reply
We sincerely thank the reviewer for their thorough and detailed critique, and we greatly appreciate the recognition of the ARS model setup, especially regarding its realistic representation of aerosol species and distributions. We believe the insights provided will significantly contribute to strengthening the manuscript.
However, we would like to clarify certain key aspects of our study to enhance understanding of the manuscript, particularly concerning the core scientific motivation, methodological approach, and interpretation of results.
Overall, we have not identified any flaws in our study and respectfully request that the reviewer consider our explanations below, which clarify the plan of this research. We will also incorporate a few lines in the manuscript based on these clarifications for added clarity.
Our study got developed evolving strategies to adequately represent and understand the spatial modulations in monsoon rainfall over the Indian subcontinent, characterized by persistent spatially weakening and strengthening patterns, seeking information strategically from observational data (rainfall measurements from the IMD), global aerosol climate model (CMIP6 simulations), and then aerosol scenario simulations designed in a regional aerosol climate model.
We first analyse the spatially gridded (25×25 km2) twenty-year (2000–2019) monsoon-rainfall (June to September averaged) derived from IMD measurements (Indian Meteorological Department, IMD) and identify the regional spatial modulations in monsoon-rainfall over the mainland. This analysis gave us insights into the presence of persistent spatial modulations (regions with weakening and strengthening rainfall trends).
To further enhance the relevant information from available measurements, we also analyse percentage departures of rainfall. The percentage departure of the IMD-rainfall is obtained as the divergence of the IMD-rainfall of a selected year from the defined normal monsoon of the respective year, averaged over the five selected years. The normal monsoon (as defined by the IMD) for a respective year is estimated as the mean of the last 50 years of IMD rainfall (Indian Meteorological Department). We hypothesise that the spatial modulations in the measured departures (positive and negative) contain signals of rainfall changes primarily driven by aerosols. It is seen that the spatial mapping of percentage departures, positive and negative, from measurements spatially overlapped with regions showing increasing and decreasing rainfall trends from measurements.
Eventually, we also find that the modelled aerosol-induced spatial modulations in rainfall, as estimated from ARS for the designed aerosol scenarios, align consistently with the measured departures in monsoon rainfall. Also, the aerosol-induced spatially changing patterns of weakening and strengthening of rainfall, as estimated from the global aerosol climate model, showed spatial concordance with the measured departures.
We have also shown a comparison of the Probability density (PD) of modelled rainfall change for three aerosol scenarios (all-aerosols, anthropogenic and dust) over the identified regions (marked as ‘R1’ and ‘R2’, Figure 1) averaged over 5-y corresponding to the five types of climatic patterns during the monsoon period with the PD of measured rainfall change (Figure 3g) obtained from IMD-rainfall departures averaged over the same period as mentioned above. This was done to show which aerosol scenario shows matching features with the PD from measured departures, and these are discussed in Section 3.4 (page 14 of the manuscript) and also in the supplementary information. A reasonably good comparison of measured departures with the modelled aerosol-induced spatial modulations indicates that the measured spatial modulations (weakening and strengthening as departures) in the monsoon rainfall are driven by regional aerosols over the Indian subcontinent.
Then we evaluate the available long-term simulations (1901 to 2014) of monsoon-rainfall, from the global aerosol climate model (IPSLCM6A-LR) in CMIP6 experiments. Here we framed our analyses using the available simulations so as to obtain information on spatial modulations in monsoon rainfall segregated by forcing due to greenhouse gases (GHG) and aerosols-induced.
To present the change in rainfall due to forcing from GHG and that from aerosols, we used pre-industrial rainfall from model estimates. Hence, we used the relative change (%) in rainfall for the present years (1995–2014) with respect to the pre-industrial (1901-1930) for simulations done with no-aerosols (i.e. atmosphere with GHG-only), and those done with aerosols (aerosols+GHG). From the above analysis, too, we could obtain spatial modulations in rainfall patterns for the present years relative to pre-industrial. However, the change in rainfall for the “no-aerosol” scenario from the global aerosol climate model does not represent or oppose the observed pattern of spatial modulations in the measured monsoon rainfall. In contrast, the spatial modulations in rainfall for the “with-aerosol” scenario (Figure 3b) show similar spatial features to the measured monsoon rainfall. Thereby, revealing the regional spatial modulations (from measured trends and departures) in IMD-rainfall comprising rainfall-deficient and rainfall-excessive areas as identified over the IGP and western India, respectively, are aerosol-driven and cannot be explained by GHG forcing.
Now, knowing that spatial modulations in measured monsoon rainfall are indeed explainable due to forcing by aerosols and cannot be explained by forcing from GHG, we further conduct specific aerosol response simulations (ARS) adequately accounting for regional aerosol distributions in a fine-resolved (25×25 km2) weather research and forecasting (WRF) regional climate model. The above is done to examine spatial modulations (weakening and strengthening) persistent in the measured monsoon rainfall over the Indian subcontinent in the context of spatial heterogeneity in regional aerosol species (anthropogenic and dust), radiative perturbations, driving mechanism and control. These simulations are designed for aerosol scenarios (all-aerosols, dust-aerosols, anthropogenic aerosols) for the present years (see methodology for selected years of simulations), focused on understanding the regional aerosol-monsoon interactions, examining the regional spatial modulations in the measured rainfall. Here we first evaluate aerosol radiative effects (relative to no-aerosols) for aerosol scenarios and for all-sky and clear-sky conditions. It is then followed by evaluating aerosol-induced rainfall changes for three aerosol scenarios, driving mechanisms and control.
Simulating the aerosols-driven role in monsoon rainfall spatial modulations over the Indian subcontinent is necessary to examine the implications of anthropogenic emissions and their mitigation on the required water management measures as presented in this study.
Overall, we present analysis from observations, the global aerosol-climate model, and the regional aerosol climate model, with a specific plan and outcome for each of these. The analysis includes presenting (i) rainfall departures from measurements showing persistent spatial modulations, (ii) segregated information on spatial modulations in rainfall due to GHG forcing (present years relative to pre-industrial) from global aerosol climate model thereby confirming the role of aerosols in driving the spatial modulations in rainfall during present years, (iii) detailed analyses with ARS in regional aerosol climate model estimating aerosol-induced radiative effects, driving mechanism, and implications on control measures.
Our analyses demonstrate that the modelled aerosol-induced spatial modulations consistently align with the measured departures in monsoon rainfall. Our study confirms that these persistent spatial modulations in the measured monsoon rainfall are primarily driven by aerosols. Moreover, the close similarity between the magnitude of the measured rainfall departures and the modelled aerosol-induced spatial modulations suggests that these departures in rainfall intensity (relative to the normal rainfall) are significantly influenced by changes induced by spatially varying aerosol species (dust and anthropogenic).
We sincerely hope that the clarification provided addresses the Reviewer’s concerns. We welcome any further discussion or suggestions and look forward to refining our methodology and scientific rationale in the manuscript for clarity. We are in the process of modifying the manuscript as per the comments received from the Reviewer.
Thank you.
Citation: https://doi.org/10.5194/egusphere-2025-2302-AC2 -
CC2: 'Reply on RC1', Sauvik Santra, 16 Aug 2025
reply
We thank the Reviewer for their comments and assessment of our work. We would like to offer further clarifications on specific comments.
Reviewer: “These simulations are used to try and reproduce/explain recent trends in observed rainfall between 2000 and 2019 over the Indian subcontinent and anomalous rainfall over the same period relative to long-term observations (previous 50 years). Herein lies a potentially fundamental flaw in this study – these observations are representative of relatively recent decades (back to 1970s), yet the simulations are either used to focus on the anthropogenic aerosol effect (so pre-industrial, not 1970) or total aerosol effect (so not applicable to any realistic scenario).”
Response: In our study, we do not use any data or reference from the 1970s. Furthermore, we do not compare the 50-year climatological mean rainfall from IMD directly with either CMIP6 outputs or the ARS simulations, as such a comparison would indeed be scientifically flawed. Instead, our approach follows the definition of “normal” rainfall as per the Indian Meteorological Department (IMD), where the 50-year mean preceding any specific year is considered the climatological normal for that year. Accordingly, for each year in the 2000–2019 period, we compute the deviation of that year’s observed rainfall from its corresponding 50-year normal, also termed as “departures” in the manuscript. These deviations are then compared with the ARS-simulated aerosol-induced changes in rainfall for the same years. Henceforth, we have not identified any flaws in our study.
Reviewer: “they are implicitly assuming that all regional aerosol over India in the present-day is due to anthropogenic activity. This is clearly unrealistic – dust and sea-spray emissions are unlikely to have changed significantly due to anthropogenic activity (and others..).”
Response: The above comment appears to be possibly a misinterpretation. One of the key contributions of this study is the explicit separation of anthropogenic and dust aerosols within the ARS simulations. Our experimental design allows us to isolate and quantify the distinct impacts of these two aerosol types on ISM rainfall. The spatially heterogeneous distribution of anthropogenic and dust aerosols is shown to modulate rainfall in contrasting ways across different regions, which is central to the novelty of this work and is clearly discussed in the results and discussion sections.
Reviewer: “One option for the authors is to use the ‘all aerosol’ simulations to gain an overall perspective on the aerosol-monsoon interactions, but then focus most of the results on the difference between including and excluding anthropogenic aerosols.”
Response: Our manuscript actually adopts an advanced and extended version of the above-mentioned approach. The ARS simulations include separate scenarios for all aerosols, anthropogenic aerosols only, and dust aerosols only, and also a no aerosols scenario, thereby enabling a species-specific attribution of impacts on monsoon rainfall.
Reviewer: “Lack of necessary detail on model configuration. For the CMIP6 simulations, were the two simulations free-running or nudged? How were aerosols represented? Are direct and indirect effects represented? For the ARS simulations – what were used as boundary conditions? Were SSTs prescribed? How long was each simulation run for? Was it nudged against reanalysis meteorology? All of these details are considered standard and should be included.”
Response: The reference to the CMIP6 IPSL-CM6A-LR model has been provided in the manuscript to refer to for details. Nevertheless, as suggested, we have now included more details of the CMIP6 simulation configuration (e.g., nudging, aerosol treatment, forcing) in the updated manuscript. Further details, as suggested for the ARS setup, have also been incorporated in the updated manuscript.
Reviewer: “How does sea spray aerosol feature in this study? It is mentioned as being part of the total aerosol scenario but never referred to afterwards.”
Response: This study focuses on aerosol-induced modulation of ISM rainfall over the Indian mainland. While sea salt aerosols are included in the “all aerosols” scenario to reflect a realistic atmospheric composition. The aim in this paper is to understand the spatial modulations in monsoon rainfall over the Indian subcontinent in the context of the role of spatial heterogeneity in aerosol species (anthropogenic and dust) radiative perturbations, driving mechanism and control.
Reviewer: “The manuscript is difficult to read in places and could be substantially more succinct.”
Response: We are verifying the manuscript for improving the overall readability. We have also incorporated most of the minor suggestions raised by the reviewer in the updated version of the manuscript.
I noticed that “Table S1” of the supplementary material was somewhat incomplete, seems some error during uploading. So I am hereby again uploading the supplementary material.
We believe this will provide further clarity on the implementation of the ARS framework and reinforce the scientific basis of the study.
Thank you again for your time and thoughtful review.
-
AC2: 'Reply on RC1', Shubha Verma, 11 Aug 2025
reply
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 494 | 70 | 17 | 581 | 30 | 12 | 27 |
- HTML: 494
- PDF: 70
- XML: 17
- Total: 581
- Supplement: 30
- BibTeX: 12
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1