the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Oct 2025
Hydrologic implications of aerosol deposition on snow in High Mountain Asia river basins
Abstract. The deposition of light-absorbing particles (LAPs) on snowpack is known to accelerate snowmelt. However, the resulting hydrological impacts, particularly on streamflow, remain under-explored. This study assesses the hydrologic consequences of LAP deposition on snow in High Mountain Asia (HMA) based on model simulations with and without aerosol deposition (“clean snow” scenario) during 2004–2018. We use the Community Land Model with a detailed aerosol-snowpack radiation module and the mizuRoute river model, driven by a 12-km meteorology-aerosol reanalysis dataset – the Model for Atmospheric Transport and Chemistry (MATCHA) – generated by the Weather Research and Forecast model coupled with Chemistry that assimilates satellite aerosol optical depth every three hours. The results show that LAPs advance seasonal snow cover disappearance by two weeks to over a month compared to the clean snow scenario. This shift alters both runoff and evapotranspiration (ET). LAP-deposited snow produces more runoff and ET until it is depleted; afterward, runoff declines due to earlier loss of snowmelt, while elevated ET persists even after LAP-deposited snow disappears as a darker, snow-free surface enhances evaporation from soil. Consequently, the annual runoff is slightly reduced under the LAP-deposited snow condition. Streamflow increases from late winter until snow melts completely but decreases due to earlier snow disappearance in LAP-deposited snow. This pattern is most evident in the headwaters, with the impact diminishing downstream. The semi-arid basins in the western HMA (e.g., Amu-Darya and Indus) show greater sensitivity to LAP deposition than the monsoon-dominated eastern HMA (e.g., Brahmaputra and Ganges). In western HMA regions, where larger perennial snow exists, LAP-enhanced snowmelt persists into summer and fall, leading to greater streamflow during these seasons compared to the clean snow scenario. This study provides important implications for the synergistic control of air pollution and water resource management.
- Preprint
(5115 KB) - Metadata XML
-
Supplement
(3313 KB) - BibTeX
- EndNote
Status: open (until 07 Jan 2026)
- RC1: 'Comment on egusphere-2025-4586', Anonymous Referee #1, 12 Dec 2025 reply
-
RC2: 'Comment on egusphere-2025-4586', Anonymous Referee #2, 22 Dec 2025
reply
This manuscript addresses an important question, namely how deposition of light absorbing particles (LAPs) on snow may alter hydrologic fluxes and streamflow across High Mountain Asia, yet the current framing and evidentiary support do not meet the standard expected for publication in Hydrology and Earth System Sciences. The central claims are derived from an offline CLM5 SNICAR sensitivity experiment in which LAP deposition is set to zero to define a clean snow counterfactual while meteorological forcing is held fixed. This design isolates a single land surface pathway through snow darkening but, by construction, excludes atmospheric aerosol radiative effects, aerosol cloud precipitation interactions, and land atmosphere feedbacks that are integral to the real world hydrologic influence of aerosols. Moreover, the manuscript provides limited quantitative evaluation of the simulated snowpack and streamflow, and it does not present basin integrated effect sizes with interannual variability or water balance closure diagnostics needed to establish the robustness and practical significance of the reported shifts, for example earlier snow disappearance versus order one week hydrograph centroid changes. Given these scope limitations, the unrealistic nature of the zero deposition baseline, and the absence of uncertainty and sensitivity analyses despite known forcing and deposition biases in high elevation regions, the conclusions currently read as qualitatively expected and insufficiently constrained for HESS without substantial reframing, additional diagnostics, and stronger evaluation.
Major comments
-
The attribution is not defensible given the experimental design. The simulations are offline and meteorology is identical in CLM-LAP and CLM-clean, so the study isolates only the land surface snow darkening pathway. Any manuscript level statements that suggest aerosols broadly alter HMA hydrology are not supported because atmospheric aerosol radiative forcing, aerosol cloud precipitation coupling, and land atmosphere feedbacks are excluded by construction. The scope must be narrowed to deposition on snow effects, or the modeling framework must be expanded.
-
The clean snow counterfactual is physically unrealistic and not policy relevant. Setting all LAP deposition fluxes to zero represents an extreme bound rather than a plausible baseline for HMA. A scientifically meaningful counterfactual should be framed as a perturbation around observed present day conditions, for example scaling deposition fluxes by factors f in {0.25, 0.5, 0.75} or applying source sector reductions. Without this, the effect sizes cannot be interpreted as actionable or comparable to mitigation scenarios.
-
Basin scale effect sizes are not quantified in a way that supports the conclusions. The paper relies heavily on maps and qualitative language such as slightly reduced or greater sensitivity. HESS readers will expect basin integrated numbers with interannual variability. At minimum provide, for each basin and for headwaters versus full basin, the following diagnostics computed annually and seasonally: ΔQ = Q_LAP − Q_clean in mm per year and percent, ΔET = ET_LAP − ET_clean in mm per year and percent, and ΔS as change in storage including at least soil water plus groundwater plus snow water equivalent carryover. Verify closure of the incremental water balance ΔP_liquid + ΔM = ΔQ + ΔET + ΔS, where P_liquid is rainfall and M is snow plus ice meltwater production. Without this accounting the interpretation of runoff decreases being driven by earlier depletion versus enhanced ET is not mathematically demonstrated.
-
The manuscript does not show robustness of timing metrics. The text claims snow disappearance advances by two weeks to over a month, yet centroid shifts are at most within a week and peak timing shifts are within a few days. This can occur due to routing integration and rainfall dilution, but the manuscript does not quantify it. Provide a reach wise relationship between local snow disappearance shift and hydrograph centroid shift, for example regression or correlation across reaches, and show how this relationship varies with snow fraction, elevation hypsometry, and rainfall contribution. Otherwise the timing narrative appears internally inconsistent.
-
Streamflow evaluation is insufficient for a study making streamflow conclusions. A brief supplement mention does not meet HESS expectations. At minimum show outlet and selected upstream gauge comparisons of the mean annual cycle and interannual variability, with standard skill metrics such as NSE, KGE, bias, and timing error in centroid day. If gauges are unavailable for some basins, this limitation must be stated explicitly and the discussion must be scaled back accordingly.
-
Snow and energy balance evaluation is missing where it matters most. Because the signal depends on snow persistence and albedo feedback, the manuscript should evaluate simulated snow cover duration and snow water equivalent using independent datasets such as MODIS snow cover and available SWE or snow depth proxies. At minimum, evaluate snow disappearance timing at elevation bands. Without this, the reported two week to one month changes could reflect forcing cold bias or model snow parameterization rather than LAP physics.
-
The forcing dataset has known high elevation biases that directly confound the LAP signal, and uncertainty is not propagated. The manuscript reports strong cold bias at high elevation and underestimation of surface black carbon in mountains. These biases affect both snow duration and LAP loading and therefore the melt response. Provide at least one uncertainty analysis that brackets the response, such as temperature bias correction sensitivity, precipitation partition sensitivity, or deposition scaling to represent underestimation. A single deterministic pair of simulations is not sufficient to claim basin specific sensitivity differences.
-
The routing configuration is too simplified for inference about timing and downstream attenuation. A spatially constant Manning coefficient n = 0.05 over all reaches is a strong assumption. Even if the differential signal between experiments is less sensitive than absolute flow, this must be demonstrated. Provide a sensitivity test for n in a plausible range and report whether ΔQ, centroid shift, and timing of maximum increase and decrease are stable.
-
Glacier processes are not adequately represented for HMA hydrology and may bias seasonal conclusions. CLM5 uses a prescribed glacier land unit and does not simulate evolving glacier geometry or mass balance. In basins such as Indus and Amu Darya, glacier melt can materially influence summer and fall flows. The manuscript should quantify the fraction of runoff attributed to glacier melt in each basin, discuss limitations explicitly, and avoid attributing late season streamflow differences solely to LAP driven snow processes without separating glacier contributions.
-
The ET response is described but not diagnosed mechanistically. The manuscript attributes persistent ET increases after earlier snow disappearance to darker snow free surfaces and soil evaporation, but this is not demonstrated. Decompose ET into soil evaporation, canopy evaporation, and transpiration, and show whether increased ET is energy limited or water limited by presenting changes in available energy and soil moisture. This is essential because the sign and persistence of ET anomalies strongly affect the annual runoff conclusion.
-
Snow cover type classification and thresholds are arbitrary and not tested. The perennial versus seasonal categorization depends on SWE greater than 5 mm and 60 day persistence at 12 km resolution. Report sensitivity to alternative thresholds and demonstrate that the headline conclusions, including the reported 40 percent reduction in perennial snow area, are not an artefact of threshold choice or grid scale smoothing.
-
The novelty is presented as a list of technical upgrades but the manuscript does not demonstrate that these upgrades lead to new hydrologic insight. To justify publication in HESS, the paper needs either a deeper process level contribution, for example elevation band attribution of ΔQ and ΔET tied to LAP loading and radiation changes, or clear improvements relative to prior studies with quantified differences. Otherwise the outcome reduces to an expected earlier melt earlier runoff narrative.
-
The management and policy implications are overstated relative to the experiment. Because the study does not simulate realistic emission reductions and does not include atmospheric feedbacks, statements about air pollution control changing water resources should be limited to a qualitative sensitivity framing. If policy relevance is retained, provide scenario based deposition perturbations and translate impacts into operational metrics such as changes in seasonal low flow quantiles, timing of center of mass, or irrigation season deficits.
Minor comments
-
Title wording is awkward and overly broad. Consider replacing “Hydrologic implications of aerosol deposition on snow” with “Hydrologic implications of light absorbing particle deposition on snow” or “Streamflow impacts of snow darkening by light absorbing particles” to match what is actually simulated.
-
Define all acronyms at first use and use them consistently. LAPs, BC, BrC, SWE, ET, HMA, MATCHA, SNICAR, CLM, PFT. Avoid switching between aerosol, LAP, and BC when the mechanism discussed is specifically snow darkening by LAPs.
-
Use consistent terminology for fluxes. “Snowmelt runoff” is confusing. Use snowmelt for meltwater production and runoff for surface plus subsurface outflow. If you mean the contribution of snowmelt to runoff, state “runoff attributable to snowmelt” and define how it is diagnosed.
-
The phrase “rain plus snowmelt” should be defined once with an explicit equation and then replaced with a shorter term, for example liquid water input or meltwater plus rainfall input. If glacier melt is included in runoff forcing to mizuRoute, clarify whether it is included in this input term.
-
The spin up description is confusing and should be rewritten for clarity. State clearly the cycling period used for equilibrium and the transient period used for analysis, with dates and number of cycles.
-
In Section 2.1, clarify whether perennial snow includes glacier land unit or only snow on the vegetation land unit. At 12 km, this distinction matters for interpretation.
-
Units and symbols need tightening throughout. Replace “C-degree” with “°C”. Use mm per year or mm yr−1 consistently, and m3 s−1 for discharge.
-
Several sentences are long and would benefit from being split to reduce ambiguity, particularly in the Abstract and Conclusions where multiple causal claims are chained.
-
The novelty paragraph in Section 2.3 reads like a checklist. Consider condensing and focusing on what materially changes hydrologic inference, not just model versioning.
-
Figure 1 caption should explicitly define what “1000 m elevation bands” means and how they were computed. Also confirm whether the elevation shown is the CLM 12 km elevation or a higher resolution DEM aggregated.
-
Figure 2 caption has a potential sign confusion. Ensure sign convention and colorbar labeling are unambiguous. Also specify whether day of year is water year or calendar year.
-
Figure 3 needs clearer phrasing and labeling. Add basin mean annotations or an inset table of basin averages. Clarify whether seasonal panels are mm per month averaged over the season or total seasonal sum divided by months.
-
Figure 4 needs axis labels and units for soil moisture, ET, and runoff, and should clarify whether soil moisture is volumetric water content, equivalent depth, or column integrated storage.
-
Section 3.3 centroid metric should be defined mathematically. Provide the equation t_c = (Σ t Q(t)) / (Σ Q(t)) and specify whether Q(t) is daily mean discharge and whether t spans calendar year or water year.
-
Fix typographic inconsistencies and hyphenation. “LAPdeposited” should be “LAP-deposited” consistently. Ensure consistent usage for snow-free, high-elevation, basin-scale, and data assimilation.
-
The Conclusions contain several forward looking statements about air pollution management. These should be softened and clearly labeled as implications of a sensitivity experiment rather than predictions.
Citation: https://doi.org/10.5194/egusphere-2025-4586-RC2 -
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 204 | 101 | 25 | 330 | 39 | 17 | 20 |
- HTML: 204
- PDF: 101
- XML: 25
- Total: 330
- Supplement: 39
- BibTeX: 17
- EndNote: 20
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Find the review file attached!