the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
19 Jan 2021
19 Jan 2021
Multi-dimensional satellite observations of aerosol properties and aerosol types over three major urban clusters in eastern China
- 1Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China
- 2Key Laboratory of Digital Earth Sciences, The Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
- 3Department of Physics, P.O. Box 64, 00014 University of Helsinki, Helsinki, Finland
- 4Institute for Environmental and Climate Research, Jinan University, Guangzhou, Guangdong 511443, China
- 5Royal Netherlands Meteorological Institute (KNMI), R&D Satellite Observations, 3730AE De Bilt, The Netherlands
- 6Aerospace Information Research Institute, Chinese Academy of Sciences (AirCAS), No.20 Datun Road, Chaoyang District, Beijing 100101, China
- 7Nanjing University of Information Science & Technology (NUIST), School of Atmospheric Physics, No.219, Ningliu Road, Nanjing, Jiangsu, China
- 8University of Mining and Technology (CUMT), School of Environment Science and Spatial Informatics, Xuzhou, Jiangsu 221116, China
- 9Coastal and Ocean Management Institute, Xiamen University, Xiamen 361102, China
- 1Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China
- 2Key Laboratory of Digital Earth Sciences, The Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
- 3Department of Physics, P.O. Box 64, 00014 University of Helsinki, Helsinki, Finland
- 4Institute for Environmental and Climate Research, Jinan University, Guangzhou, Guangdong 511443, China
- 5Royal Netherlands Meteorological Institute (KNMI), R&D Satellite Observations, 3730AE De Bilt, The Netherlands
- 6Aerospace Information Research Institute, Chinese Academy of Sciences (AirCAS), No.20 Datun Road, Chaoyang District, Beijing 100101, China
- 7Nanjing University of Information Science & Technology (NUIST), School of Atmospheric Physics, No.219, Ningliu Road, Nanjing, Jiangsu, China
- 8University of Mining and Technology (CUMT), School of Environment Science and Spatial Informatics, Xuzhou, Jiangsu 221116, China
- 9Coastal and Ocean Management Institute, Xiamen University, Xiamen 361102, China
Abstract. Using nine years (2007–2015) of data from passive (MODIS/Aqua) and active (CALIOP/CALIPSO) satellite measurements over China, we investigate (1) the temporal and spatial variation of aerosol properties over the Beijing-Tianjin-Hebei (BTH) region, the Yangtze River Delta (YRD) and the Pearl River Delta (PRD) and (2) the vertical distribution of aerosol types and extinction coefficients for different aerosol optical depth (AOD) and meteorological conditions. The results show the different spatial patterns and seasonal variations of the AOD over the three regions. Annual time series reveal the occurrence of AOD maxima in 2011 over the YRD and in 2012 over the BTH and PRD; thereafter the AOD decreases steadily. Using the CALIOP vertical feature mask, the contributions of different aerosol types to the AOD were analysed: contributions of dust and polluted dust decrease from north to south, contributions of clean ocean, polluted continental, clean continental and smoke aerosol increase from south to north. In the vertical, the peak frequency of occurrence (FO) for each aerosol type depends on region and season and varies with AOD and meteorological conditions. In general, three distinct layers are observed with the peak FO at the surface (clean continental and clean marine aerosol), at ~1 km (polluted dust and polluted continental aerosol) and at ~3 km (smoke aerosol), whereas dust aerosol may occur all over the altitude range considered in this study (from the surface up to 8 km). In this study nighttime CALIOP profiles were used. The comparison with daytime profiles shows substantial differences in the FO profiles with altitude which suggest effects of boundary layer dynamics and aerosol transport on the vertical distribution of aerosol types.
Yuqin Liu et al.
Status: open (until 16 Mar 2021)
-
RC1: 'Review of Liu et al. - Multi-dimensional satellite observations of aerosol properties and aerosol types over three major urban clusters in eastern China', Anonymous Referee #2, 12 Feb 2021
reply
The paper “Multi-dimensional satellite observations of aerosol properties and aerosol types over three major urban clusters in eastern China” by Liu et al. present and discuss the temporal and spatial variation of aerosol properties over the Beijing-Tianjin-Hebei, Yangtze River Delta and Pearl River Delta regions, and the vertical distribution of the different aerosol subtypes, based on nine-years (2007-2015) of MODIS-Aqua and CALIOP-CALIPSO observations, together with meteorological parameters provided by ERA-Interim reanalysis. The study falls within the scope of ACP. The manuscript is well-written and structured, the presentation clear, the language fluent and the quality of the figures high. The authors have done a thorough job and the results support the conclusions. I recommend publication in ACP, however I recommend the following revisions before it can proceed to be published.
Comments:
- In the framework of the study CALIPSO nighttime observations are used. The authors justify their choice through the illumination conditions and the related lower daytime SNR. However, I would suggest to extend their justification through CALIPSO daytime and nighttime performance, as provided in the literature.
- In the end of Section 2.1, a map encompassing East China, delineating the three BTH, YRD, and PRD study regions, would be practical for the reader.
- I would suggest to the authors to break down Section 2.2 - “Data sources” into three distinct sub-sections: “MODIS/Aqua”, “CALIOP/CALIPSO” and “ERA-Interim”.
- Whenever a web-link is provided, please follow the formalism on adding in a parenthesis the “last visit:” information (e.g. line 137).
- Lines 141-143 “Over China, the differences between the C6 and C6.1 AOD are small, except over certain areas like the Tibetan Plateau, Sichuan Province and the NW of China”. Please include related references.
- Why have the authors selected MODIS AOD 1.5 and CALIPSO AOD 3? Please include references. Moreover, regarding comparison methodology, I would argue the use of similar CALIPSO and MODIS upper AOD limits.
- Please include information of the pre-processing of MODIS/Aqua. Which Quality Assurance procedures and flags are used? Based on Figure 2, the authors have used a re-griding procedure in MODIS L2 AOD, which is not mentioned. Is any final smoothing applied to the data? Moreover, Figure 2 has same “blank” areas, without AOD values. Please discuss these aspects/address in the manuscript.
- “CALIOP is the first space-borne near-nadir dual-wavelength lidar (532 nm and 1064 nm)”. This is not correct (e.g. ICESat). Please revise.
- Lines 155-161: The authors discuss the different levels of processing in CALIPSO algorithms. Please mention which refer to L1B and which to L2, and also that the study is based on L2. Moreover, add the word “Version” beside the “4.1”.
- The authors mention that AOD is the vertical integration of extinction. However, it is not discussed in the manuscript how the mean AOD is calculated. The official and more robust way, as discussed in Amiridis et al. (2013) and Tackett et al. (2018), is calculating the mean quality-assured extinction coefficient profile at the overpass level - based on L2 profiles per overpass, and accordingly using all overpass-mean profiles to calculate the seasonal or annual profile. Not following this approach results in weighting effects, thus in not representative results. Please provide in the manuscript the methodology followed in the processing of CALIPSO profiles at mean-extinction-coefficient and AOD at 532nm, and if the methodology is different, make necessary corrections.
- Lines 168-169: “This is further illustrated …”. Add in the end the Section/Figure that support this sentence.
- Lines 174-180: Profiles of RH are provided in CALIPSO L2 Aerosol and Cloud Profiles, based on MERRA-2 model, which is used in the algorithms of CALIPSO in order to produce the different optical products. Possible use of ERA-Interim may results in some point of model-intercomparison. Have the authors considered the use of RH from CALIPSO datasets?
- Line 182: “cloud-free pixels”. What is a pixel? Please define clarify in the manuscript. Is it the L2 Profile, the grided L3 profile, or the region (e.g. BTH region) to be cloud-free?
- It is not clear why the authors have used the limited period between 2007 and 2015. Since this period has already been discussed in Proestakis et al. (2017), it would be interesting and of added-value to include more years in the analysis. Expanding the timeseries would be of added value for timeseries analysis, especially due to the special orbital characteristic and overpass frequency of CALIPSO. Moreover, extending the observational period would address the question whether emissions have increased after 2015 or whether they are still declining due to regulations applied.
- In Section 3.1, please provide in the manuscript the trends, the related statistical significance, and discuss the outcomes.
- In Figures 1b and 1c please include vertical error-bars/uncertainty-bars. Moreover, what is missing is the information on the number of overpasses/profiles used for the calculation. Discuss in the manuscript the outcomes in combination with the number of overpasses/profiles, with respect to the representativeness of the results.
- In the figures, please add the sensor name. For instance in Figure 1, modify the caption to “CALIPSO annually (a), …”.
- Although the language is fluent and the manuscript smooth, at some points the language could be more formal (e.g. line 253: “The monsoon brings heavy rains which effectively washout aerosols”).
- Throughout the manuscript, in the framework of the discussion of the different types of aerosols, the related sources, and atmospheric mass origin, phrases such as “likely”, “maybe” and more are frequently used. To this end, of high value would be to add aerosol backtrajectories cluster analysis, to strengthen the discussion, and avoid hypothesis used, especially in this extend.
- In Figure 2 add annual figures and corresponding discussion.
- Figure 3: use the CALIPSO official colors if possible, to the aerosol subtypes.
- It would be of added value to the reader of the manuscript and to the manuscript itself to add a brief discussion and description on the CALIPSO aerosol classification algorithm (Kim et al. 2018) and to possible errors in the classification (Burton et al. 2013), since the aerosol subtyping is a cornerstone to this study.
- Lines 324-325: Include AOD limits of the different “moderately polluted”, “polluted” and “heavily polluted” conditions. A histogram of the AOD values delineating the different categories would be nice also.
- Figure 4: Please add variability-bars, and maybe “number of cases-used” to the right-axes, to provided to the reader a degree of representativeness. Moreover please explain to the manuscript the feature of extinction coefficient increasing close to the surface (0 km).
- FO is not explained clearly. Is it the “number of a specific aerosol subtype to the total number of aerosol”, the “number of a specific aerosol subtype to the total number of aerosol including Clear-Air”, or something else? For instance in the FO figures, the aerosols subtypes is between 0 and 1. Have the authors converted the FO to percentage? If not the FO is very unexpectedly/unphysically low.
- In Figure 5 please add the “Annual figures”.
- Lines 483-484: The use of Layer A, B, C, is not very convenient for the reader. Please consider the use of alternative ways of formalism.
- What I am missing in the manuscript is a connection between observations and physics. Observations are discussed, but the study does not go deeper. For example the authors could discuss that polluted dust and smoke are hydrophilic aerosols, in presence of high RH the act as effective CCN aerosols, releasing Latent Heat and contributing to instability, while dust aerosols act as effective IN aerosols, having significant effects in higher altitude, … Please consider improving the manuscript in relation to physical interpretations of the outcomes.
- In the end, conclusion section, what is missing is a section of the way the observations can be used and their added value. Some examples for the authors could be effect on human health, transport, deposition, and more, to extend and discuss them.
-
RC2: 'Comments on “Multi-dimensional satellite observations of aerosol properties and aerosol types over three major urban clusters in eastern China” by Yuqin Liu et al.', Anonymous Referee #1, 20 Feb 2021
reply
General comments
The authors use MODIS/Aqua and CALIOP/CALIPSO observations to retrieve the temporal, horizontal and vertical variation of aerosol properties (AOD/extinction and aerosol type) over the regions of Beijing-Tianjin-Hebei (BTH), Yangtze River Delta (YRD) and Pearl River Delta (PRD) in the period from 2007 to 2015. The paper is comprehensive, well written and suitable for ACP. However, there are several methodological issues that must be tackled before the manuscript can be accepted. There are deficits in the understanding of the vertical information provided by the CALIPSO retrievals, and the paper suffers from a tendency to overinterpret the results. My major concerns are:
1) It is stated (in Sec. 2.3) that MODIS AOD greater than 1.5 were discarded. However, mean values in areas with high pollution (see Fig. 2) seem to be close to 1.5, and even white pixels appear in the red areas where the mean AOD is larger than 1. These findings clearly hint to a bias that is introduced by discarding any AOD > 1.5. The entire study provides only long-term mean values, without any statistical investigation regarding the range of values, standard deviation, median, percentiles etc. to underline the significance of the findings and to detect potential biases. Therefore, conclusions, e.g., about trends in AOD are not trustworthy.
2) The study aims at the synergetic use of aerosol products from passive and active sensors. However, the results are presented just next to each other, without fully exploiting the synergy. A comparison of AOD values retrieved from imager and lidar is missing, and it remains unclear whether the results are consistent (see also previous comment on possible biases).
3) Background information on the original satellite retrievals and their limitations is missing. Overinterpretation of results and even circular reasoning are the consequence. The interpretation of the appearance of aerosol types in the vertical column above the three regions requires the understanding of the decision tree for assigning an aerosol type to aerosol layers detected in the CALIOP signals (see Kim et al., 2018, Fig. 1). The aerosol subtype selection depends, e.g., on the surface type and the aerosol layer height. In the paper, atmospheric “findings” are discussed that actually originate from the CALIPSO retrieval input and threshold parameters. One should always keep in mind that the CALIPSO aerosol typing is a pre-condition for the L2 algorithms (to select a proper lidar ratio for the extinction retrieval) and is relying only on L1 data and auxiliary information.
4) There is a contradiction in the discussion of CALIOP daytime vs. night-time data. First, it is stated that the study is restricted to night-time observations because of the lower SNR and corresponding biases at daytime (L165 ff.). Later, day- and night-time results are directly compared, and the differences are solely related to atmospheric processes, without considering any biases in the retrievals. A corresponding error estimation is not provided.
5) It is unclear why the investigation is restricted to the period 2007-2015. Several more years of data are available until today.
Specific comments
L38, “Aerosol effects depend on particle size distribution”: Very generic. Which effects concretely?
L40, “... important role in Earth’s climate change”: Very generic. How exactly do aerosol particles change the climate (and not just influence the energy balance)?
L41, “…serve as cloud condensation nuclei”: Why only condensation nuclei? What about ice nuclei?
L35 vs. L44 etc.: “Aerosol” is defined in singular (L35), but then often used in plural. What exactly do you mean with “aerosols” (aerosol particles or aerosol types or something else)?
L48, “…aerosol indirect effects are strongly influenced by aerosol types”: What about the direct effects? How do direct and indirect effects depend on the aerosol type?
L168, “This is further illustrated…”: What is illustrated and where?
L181 ff.: The description of data processing is too brief. What does “cloud-free pixels” exactly mean? To which instrument and which cloud-detection scheme does it refer? Does it only hold for MODIS? What is about CALIOP? How is aerosol-cloud discrimination considered (e.g., are aerosol layers in cloudy profiles included or are only fully cloud-free profiles considered)? How is the AOD from CALIOP calculated? What do the quality control flags mean, i.e., which lidar data are discarded? Are the AOD retrievals from imager and lidar equivalent (e.g., in terms of coverage) and how do the values compare?
L200 ff.: Does the description refer to MODIS or CALIOP AOD?
L227 ff.: “…larger boundary layer heights … allow for mixing over a deeper layer resulting in elevated AOD”: Why? Usually, mixing leads to a dilution of aerosol in the BL while AOD remains constant.
L231 ff.: Trajectory analysis and source attribution are needed instead of speculations about the sources of high AOD. Aerosol removal processes like washout and deposition must be considered in the discussion as well.
Fig. 1: What is the data source, MODIS or CALIOP? More statistical information is required (e.g., box plots). At least the standard deviation must be provided and explained for all panels.
Fig. 2: What is the data source? What do white pixels mean? How can mean values close to 1.5 be explained, if all AOD values larger than 1.5 are discarded?
L285 ff.: It is stated that the relative contribution of each aerosol type to the aerosol burden is calculated. However, for such an investigation it would be necessary to weight the aerosol type occurrence with the layer mean extinction. Using only the occurrence does not say anything about the contribution to the aerosol burden.
Sec. 3.2.2 and 3.3: See major comment no. 3. It is important to understand the CALIPSO typing scheme for the interpretation of the findings.
L317, “The vertical profiles of the aerosol extinction coefficients describe the variation of the attenuation of the laser light…”: Vice versa. The aerosol extinction coefficient profile is the atmospheric property that causes variations in the attenuation of laser light, which can thus be used to describe the extinction.
L321, “…modulated by the boundary layer”: What does it mean?
L323, “…soothes the features and results in rather smooth profiles”: What does it mean?
L323 ff.: Which AOD is used and what are the concrete intervals? Please provide numbers!
L333 ff.: This discussion is strange (input = output).
Fig. 4: Figure caption is wrong.
Fig. 4: Please provide the variance and the AOD values for each profile.
L355: The definition of “layers” in this context is a bit misleading, since the retrieval is originally starting from distinct layers detected by CALIOP. It would be better to speak of “height ranges” here. These height ranges should also be indicated in the figures, in order to guide the reader in the discussion. It should be discussed if and how the CALIPSO typing scheme artificially introduces the boundaries of these height ranges.
L381 ff.: Please explain, under consideration of the CALIPSO typing scheme, where marine and dusty marine profiles come from. Also explain the occurrence of smoke in view of the typing scheme.
Fig. 5: Figure caption is wrong.
L437, “dusty marine aerosol”: From Fig. 6, it seems that it is clean marine aerosol. Again, please consider the surface-dependent typing in the interpretation.
Fig. 6: Figure caption is wrong.
Sec. 3.4.1: What does the RH at 950 hPa have to do with the vertical distribution of aerosols up to 8 km height? This discussion is very misleading. RH analysis can only be done when the RH for each detected layer is considered. A trajectory analysis would be much more appropriate to support the discussion of air mass origin.
Fig. 7: While the RH discussion is already strange, the provision of 2 decimal places is even more useless.
Sec. 3.4.2: Horizontal transport (according to trajectory analysis) and limitations of the typing scheme need to be considered in the discussion.
L533 ff.: Again, consider how smoke is assigned to a layer in the typing scheme and do not overinterpret the results. This discussion is mainly based on circular reasoning.
Sec. 3.5: As mentioned in major comment no. 4, it is unclear how the detection limitations at daytime influence the obtained differences. Detection thresholds, noise, and the influence of the quality control flags must be investigated, before conclusions can be drawn from the findings.
Fig. 10: Figure caption is wrong.
L575 ff.: Discussion on the influence of the diurnal variation of emissions is missing.
Fig. 11: Figure caption is wrong.
Conclusions: Reconsider your conclusions in view of all the comments above.
L654, “both flying on the Aqua satellite”: Really?!
Acknowledgements, “We also thank the reviewers of this paper for their valuable comments which helped improve the manuscript.”: Well, I do hope that this statement will become true.
Technical corrections
L146, “…we use of the MODIS”: delete “of”
L187: typo in index 532nm,unc
L265: North China Plain (not Plan)
L349: certain (not certainly)
Yuqin Liu et al.
Yuqin Liu et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 116 | 38 | 7 | 161 | 1 | 2 |
- HTML: 116
- PDF: 38
- XML: 7
- Total: 161
- BibTeX: 1
- EndNote: 2
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1