the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 03 Aug 2026
Multi-year Raman-polarization lidar profiling in Central Asia: Seasonal analysis of dust and non-dust optical properties over the ACTRIS site in Dushanbe, Tajikistan
Abstract. Dushanbe, located in Central Asia, resides in the global dust belt, fundamental for the emission and transport of dust. The Central Asian Dust EXperiment (CADEX, 2015 to 2016) provided foundational insights into the optical properties of atmospheric dust in Central Asia by means of lidar observations. A follow-up dataset of high-quality multi-year (2019 to 2025) lidar observations from a newly established Aerosol Clouds and Trace Gases Research Infrastructure (ACTRIS) site in Dushanbe, Tajikistan have been analyzed with the goal to deepen the knowledge of the extent, duration and interannual characteristics of the annual dust cycle in the Central Asian region. The dataset consists of automatically processed lidar data, including the application of the POlarization Lidar and PHOtometer Network (POLIPHON) method to separate the measured total particle backscatter coefficient into a dust and non-dust fraction. The study found a pronounced cycle of the dust backscatter coefficient over the course of several years with low values in winter, also in the planetary boundary layer, and a slow onset of the dust season in spring with an increased occurrence of lofted dust layers. The main dust season spans from July to September with a sharp end in September/October. During summer, non-dust particles, likely local pollution, additionally get lifted above the PBL. The PBL is normally well-mixed and shows rather constant contribution of local pollution. These results show the importance of long-term lidar observations in the Central Asian region and the value of high-quality automated processing of the retrieved data.
- Preprint
(10869 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 14 Sep 2026)
- RC1: 'Comment on egusphere-2026-3390', Anonymous Referee #1, 13 Aug 2026 reply
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 111 | 43 | 9 | 163 | 6 | 9 |
- HTML: 111
- PDF: 43
- XML: 9
- Total: 163
- BibTeX: 6
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
In this article, the authors present a multi-year Raman polarization lidar dataset from Dushanbe, Tajikistan, a region which is part of the global dust belt and where long-term vertically resolved aerosol observations remain scarce. This study is a continuation of the CADEX campaign realized between 2015 and 2016 providing further insights into the seasonality and inter-annual variability of the lidar-derived particle backscatter coefficient, particle linear depolarization ratio and aerosol mass concentration of dust aerosols over the region. The analysis is based on PollyXT lidar observations processed with the automated PollyNET procedure described by Klamt et al. (2024) for the Raman inversion of optical products and the estimation of dust mass concentrations. Given that aerosol observations in the vertical are scarce in this part of the world, the manuscript presents valuable information extending previous efforts and building towards an aerosol climatology in the area that can be used for satellite and model validation. The manuscript is well organized and supported by relevant literature. Nonetheless, several passages are overly speculative, and some sections contain unnecessary repetition. The manuscript is suited for publication in Atmospheric Chemistry and Physics and can be published after addressing the major comments listed below.
Specific comments:
1. The introduction establishes the relevance of mineral dust to climate, clouds, ecosystems, and human health. It also summarizes previous lidar campaigns conducted within the global dust belt, with particular emphasis on the 2015–2016 CADEX campaign. However, this overview could be shortened and reorganized around the specific knowledge gap that the present study aims to address. At present, the introduction provides a largely descriptive account of previous campaigns without clearly summarizing their main findings, identifying their limitations, or explaining how they motivate the current study. The authors should synthesize the relevant conclusions from earlier work, clearly define the unresolved scientific questions, and focus and explain how the present multi-year dataset advances beyond the existing observations.
2. Airmass source attribution method: Air parcels are considered to be influenced by the surface they pass when they fall below the 2km altitude threshold. The authors then conveniently define 2 regions from 0.7 to 2 and from 2 to 5km which are at the intersection of this altitude threshold and discuss the optical properties in relation to the surface types. Given that this study is using Raman observations which are primarily available during nighttime, I am wondering how representative is the inclusion of 1) daytime back trajectories, 2) the 2 km altitude threshold for dust particles given that the publication by Val Martin et al., (2018) refer to the injection heights of smoke aerosol particles and 3) the actual PBL height during nighttime.
3. The manuscript lacks a quantitative uncertainty analysis. The reported dust and non-dust optical properties, as well as the resulting dust mass concentrations, are not directly measured quantities but depend on several retrieval steps and assumptions. The authors currently characterize the variability among individual profiles and across months or seasons using percentiles or standard deviations. However, this statistical variability should not be confused with the uncertainty of the retrieved quantities themselves. The authors should quantify the retrieval uncertainties and describe how uncertainties associated with the individual processing steps and assumptions are propagated into the dust and non-dust optical properties and, ultimately, the estimated dust mass concentrations (Fig. 8).
4. The reported July-to-September maximum and rapid September/October decline are central findings. At present, however, the study emphasizes what the lidar observes more than why the seasonal cycle occurs. At several places the conclusions are highly speculative. For example, throughout the manuscript, the non-dust component either within PBL or lofted is attributed to local pollution without any further investigation. The manuscript would benefit from a concise meteorological analysis involving, for example, seasonal wind direction and speed, boundary-layer depth, trajectory clusters, fire location information etc. Then, the authors should link their findings to literature and compare them against the past CADEX campaign and other lidar studies in the area.
5. The use of automatically retrieved lidar optical properties is justified given the size of the dataset. However, it is unclear whether the automated retrievals have been validated against manually retrieved and quality-controlled optical properties. Such a comparison is necessary to assess the performance of the automated procedure and quantify potential discrepancies between the two approaches. The authors should validate the automated products using a representative subset of manually analysed profiles covering different seasons, aerosol loads and signal-to-noise conditions. The comparison should include appropriate statistical metrics, as well as a discussion of any systematic differences.
6. Line 49: ‘The aforementioned campaigns were strongly focused on the Saharan region.’. This is not entirely correct. Many of the campaigns mentioned in the text focused on Arabian dust as well. There are also a few more on Asian dust, for example, Cottle et al., (2013) and Hu et al., (2020) and references therein which are not included at all.
7. Line 167: the orange dots are described in the figure legend as valid POLIPHON profiles, whereas the main text appears to refer to them as available usable profiles. Please clarify exactly what the orange dots represent and use consistent terminology in the text and figure caption. It would also be helpful to include the available Raman profiles in the figure, allowing the reader to distinguish among the total number of measurements, profiles suitable for Raman retrievals, and profiles suitable for the POLIPHON analysis. In addition, please standardize the acronyms throughout the manuscript. For example, POLIPHON is introduced in uppercase letters in the text but appears in lowercase letters in the figure legend.
8. Lines 191-204: This paragraph is a bit confusing. For example, it is unclear how profiles with undetected clouds from the automatic procedure were treated. Did you exclude the cloud part manually or removed the entire profile completely?
9. Lines 211-213: This paragraph is difficult to follow and should be rewritten to explain the processing steps and selection criteria more clearly. I understand that this threshold is applied to the particle backscatter coefficient, but the temporal averaging used in the lidar inversion is unclear. Please specify the temporal averaging and clarify whether the threshold value of 0.6 is applied as an upper or lower limit.
10. Lines 215-218: Backscatter-related Angstrom exponents could be utilised below this height threshold and could provide further insights into the aerosol composition.
11. Line 225-226: Provide the reference for the selection of 0.31 and 0.05 values.
12. Please clarify somewhere in the manuscript the effective data time coverage. In Figure 3, there are observations available for 48 months which is equal to 4 years. In Line 71, it mentions 5 years while the overall Jun 2019- Jun 2025 implies 6 years.
13. In Figure 4, the number of profiles available for each season is indicated, but what is the vertical availability of the data? The authors state that the upper limit of each profile is determined by its reference height. Therefore, it is unclear whether the median values at different altitudes are calculated from the same number of profiles. Please indicate the number of valid profiles contributing to the median at each altitude, for example by adding a height-dependent sample-size profile to the figure.
14. The particle linear depolarization ratio is a pivotal parameter for aerosol classification and the POLIPHON method. Have the authors considered presenting this parameter similar to Figure 4?
15. Figure 5. In line with my previous comment, the lidar ratio is another valuable aerosol optical parameter. Given the long-term automated Raman dataset analysed in this study, its inclusion would strengthen the aerosol type interpretation. To this direction, although the study was limited to heights above 700m due to the limitation of the extinction coefficient, this information is not presented in the current study. Instead, a constant lidar ratio is assumed to convert the particle backscatter coefficient to extinction for the dust mass calculations. Could the authors provide further clarification?
16. Line 286: ‘The same seasonal climatology of the aerosol backscatter coefficient depicted in…’ Consider changing the word ‘climatology’ to ‘variability’ since climatology refers to much longer temporal scales.
17. Lines 340-341: Do the authors imply here that fresh smoke is a depolarizing aerosol particle?
18. Figure 8c: 2023 is missing from this plot although there seems to be information according to Figure 3. Please add an error estimation in this figure and the number of profiles available for each year.
Technical corrections:
The manuscript would benefit from careful proofreading to correct grammatical errors and improve the clarity and consistency of the English throughout.
Line 8: The terminology used for the particle backscatter coefficient is inconsistent. The manuscript refers to it as the “backscatter coefficient” in lines 212 and 221 and as the “aerosol backscatter coefficient” in line 286. Please standardize the terminology throughout the manuscript.
Line 12: ‘The PBL is normally well-mixed….’. Do you mean usually well-mixed?
Line 41: Reference for PollyXT is missing.
Line 50: Round-based -> ground-based.
Line 63: ‘PollyNET is part of the European Aerosol Research Lidar Network…’ -> PollyNET is a network of automated multi-wavelength Raman polarization lidars, several of which contribute observations to EARLINET.
Line 119: ‘…which were able to reach dust mass concentrations…’ Do you mean ‘with estimated dust mass concentrations’.
Line 141: Reference is missing for Cimel CE318.
Line 146: Acronym definition for LPDM Flexpart is missing.
Line 156: if -> of
Line 221: Reference for POLIPHON is missing.
Line 663: Remove the extra ‘T’ from the title.
References
Paul Cottle, Detlef Mueller, Dong-Ho Shin, et al. "Studying Taklamakan aerosol properties with lidar (STAPL)", Proc. SPIE 8894, Lidar Technologies, Techniques, and Measurements for Atmospheric Remote Sensing IX, 88940X (22 Oct 2013); https://doi.org/10.1117/12.2029158.
Hu, Q., Wang, H., Goloub, P., Li, Z., Veselovskii, I., Podvin, T., Li, K., and Korenskiy, M.: The characterization of Taklamakan dust properties using a multiwavelength Raman polarization lidar in Kashi, China, Atmos. Chem. Phys., 20, 13817–13834, https://doi.org/10.5194/acp-20-13817-2020, 2020.
Klamt, A., Yin, Z., Floutsi, A. A., Griesche, H., Haarig, M., Radenz, M., Jimenez, C., Gast, B., and Baars, H.: PollyNET: Pollynet Processing Chain, Zenodo, https://doi.org/10.5281/zenodo.13379737, 2024.