the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
The role of refractive indices in measuring mineral dust with high-spectral resolution infrared satellite sounders: Application to the Gobi Desert
Abstract. Mineral dust significantly influences the Earth's climate system by affecting the Earth's radiative balance through the absorption and scattering of solar and terrestrial radiation. Understanding the physico-chemical properties of dust in the longwave region of the electromagnetic spectrum is crucial for a more accurate estimation of the radiative budget. The complex refractive index (CRI) of dust in the infrared (IR) is a key parameter for mineral dust characterization from satellite remote sensing. Particularly, high-spectral-resolution instruments have shown the ability to measure these aerosol properties, e.g., the Infrared Atmospheric Sounding Instrument (IASI). This work reviews six prior laboratory Complex Refractive Index (CRI) datasets, which focus on advancements in laboratory measurement techniques aimed at characterizing dust properties via IASI measurements during a dust storm event over the Gobi Desert in May 2017. We evaluate the sensitivity of IASI to different CRI datasets using the ARAHMIS radiative transfer algorithm and explore their impact on retrieving size distribution parameters by mapping their spatial distribution. The results indicate that the new laboratory CRI datasets decrease the total error by 30 % and that the choice of CRI significantly impacts the accuracy of dust detection and characterization from satellite observations. Notably, datasets that incorporate aerosol generation techniques with higher spectral resolution and samples from the case study region show enhanced compatibility with IASI observations. The outcomes of this research emphasize two key points: the crucial connection between dust's chemical composition and its optical properties, and the need to consider the specific composition of the CRI dataset for improved retrieval of the microphysical parameters. Moreover, this study highlights the critical role of continuous enhancements in CRI measurement approaches, as well as the potential of high-spectral-resolution infrared sounders for aerosol atmospheric investigation and understanding their radiative impacts.
- Preprint
(2369 KB) - Metadata XML
-
Supplement
(354 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2024-888', Anonymous Referee #1, 22 Apr 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-888/egusphere-2024-888-RC1-supplement.pdf
-
AC1: 'Reply on RC1', Perla Alalam, 19 Jul 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-888/egusphere-2024-888-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Perla Alalam, 19 Jul 2024
-
RC2: 'Comment on egusphere-2024-888', Anonymous Referee #2, 31 May 2024
Review of paper "The role of refractive indices in measuring mineral dust with high-spectral resolution infrared satellite sounders: Application to the Gobi Desert" by Perla Alalam et al.
The paper is dedicated at assessing the influence of different determinations of mineral dust complex refractive indices (CRI) in the IR spectral range on the aerosol retrieval from IASI observations. Six different determinations of CRI are used to assess the capability of retrieving some dust properties (diameter and mixing ratio), and are applied to the retrieval for an intense dust storm occurred in 2017 over the Gobi area.The study addresses an important topic, because of the role that dust plays in the climate system through different mechanisms, as well as on the surface, atmospheric, and top of the atmosphere radiation budget.
Some aspects of the paper require additional investigation and a clearer presentation. In particular:
- some of the used methods are not well described and require more details or at least a brief summary.
In particular, there are no details or references to the principal component analysis method briefly described in lines 111-114, also applied elsewhere (e.g., l. 189-190).l. 94-99: the "V-shape" dust criterion is mentioned here, but a short description of the method and what is its sensitivity is needed (there is reference to a phd thesis, but a more specific reference would be preferable). Does a difference between two brightness temperatures necessarily imply a v-shaped spectrum? Are there specific assumptions on the spectral variability of surface properties needed?
- no verification data are presented, except for a generic comparison with measurements at two sites of the SONET network. Did the authors verify if other data are available, for example in the Skynet network (https://www.skynet-isdc.org/obs_sites.php)?
- the retrieval of the dust properties is made under several assumptions (single mode, if I understood well fixed standard deviation of the size distribution, uniform refractive index, fixed vertical distribution). Thus, although the correspondence between modeled and measured IR spectra is a relevant constraining method, the derived parameters might better be defined as radiativelly equivalent aerosol model, than specific aerosol properties. The possible influence of different distribution standard deviations and vertical dstributions (which has been shown to play a significant role on the satellite radiances, see e.g. Clarisse et al., 2019) should be addressed and investigated.
- Section 4 needs a better explanation and a clearer treatment. Also the notation is not clear. It is not clear why the AOD at 1020 nm is used here and how it is combined with determinations in the IR. Also here, all the calculations are made under the assumption that the only driving parameters are the median diameter and the concentration. How different factors affect the information content?
- Results shown in figure 6 might need a deeper discussion. Is there a reason for the large differences in the results obtained for DB17, DSC22 and VMA? All of these should somewhat take into account the specific mineralogy of the source. Two sets of CRI are for dry samples; might humidity play a role? Without validation data it is difficult to derive conclusions.
- The abstract needs to be improved, and some sentences need clarifications.
Minor comments are reported below.The same event in May 2017 was analysed in the paper by Alalam et al. (2022). This study seems an expanded sensitivity study of the same retrieval, leading to similar conclusions. It may be useful to highlight overlapping parts, differences, and new results.
Citation to references (e.g., parentheses) are not correct throughout the text.
line 8: emission should also be mentioned in addition to absorption and scattering
line 11: reference to IASI seems to be misplaced
l. 12: The sentence "this work reviews six prior ..." is misleading.
l. 16: here and elsewhere, the authors should clearly define what are "old" and what "new" CRI datasets. Please, indicate on what quantity you find a
decrease of total error by 30%. Reconstructed spectra should be mentioned.l. 27: please add the reference to IPCC.
l. 39: simulate spectral fits? Please, clarify.
l. 45: lack of size distribution knowledge in addtition to CRI is a relvant factor.
l. 39-59: it is not clear if the authors refer to determinations and calculations in the visible or IR or both. There is a wide literature on dust characterization
based on visible spectra. If they include visible spectra the discussion should be expanded.l. 69: is there e reference on the CESAM system? What is meant for "relevant atmospheric conditions"?
l. 75: see also comment to l. 16: please, define what are the new CRI measurements.
l. 83-84: accuracy of mineral dust retrievals: of which properties?
l. 93: "most dispersion and visibility to IASI observations". Please, clarify. Is it due to clouds? To the dust amount, vertical distribution, surface conditions?
l. 133: same as above: please specify what is intended for "atmospheric relevant conditions".
l. 151: it seems that using soil from the correct area (i.e., a more similar mineralogy) may have a strong impact as well.
l. 168: I would suggest using "median diameter" or "geometric mean diameter" instead of geometric size diameter.
l. 183: is the surface temperature retrieval affected by the occurrence of dust?
l. 194-195: if I understand well, the dust vertical distribution is assumed to be a vertically homogeneous layer between 1.5 and 2.5 km height. As discussed above, how is the assumption on the dust vertical distribution affecting the results? The authors state that a reduction of visibility was associated with this event. This does not seem fully compatible with a dust layer above 1.5 km altitude.
l. 246-247: is the effect of spectral stability included in SNR? It is expected to contribute.
l. 251: SNR set to 500: this requires a justification.
l. 260: please, explain better why the accuracy on albedo (in which spectral range?) is related with the uncertainty on surface emissivity.
l. 283: please, clarify what is the range of values needed to avoid saturation and to avoid loosing sensitivity.
l. 294: please, for clarity, use percent (used elsewhere) instead of fractional values for Sx.
l. 294-295: "The relative behavior of the CRIs ...": the sentence is unclear.
l. 313-314: "... is by tenth ...": sentence unclear.
l. 324-325: The sentence needs clarification. In my opinion, you can not use RMS on spectra as a measure of uncertainty on CRI. What would be the RMS corresponding to the measured spectra noise level (SNR=500)? This might give a reference value.
l. 325-326. "We selected spectra with large spectral features, whose simulation presents significant challenges and thereby provides a rigorous test for the CRIs" please, clarify. The sentence seems contradictory (you need large spectral features for detection).
l. 328-330: looking at fig. 5, it seems that for all cases and all CRI datasets there are spectral regions with model/measurement differences which probably exceed SNR=500 on measurements. May this provide useful suggestions on lacking knowledge on dust properties?
l. 330-331: although this might be the case, you might have compensating effects among different variables.
l. 339: is there an optimal range of absorption/extinction for detection? Is the largest extinction due to too much absorption (ie., imaginary part of CRI)?
l. 352: what was local time?
l. 354: "The mean value indicates the range of each parameter": please, clarify.
Citation: https://doi.org/10.5194/egusphere-2024-888-RC2 -
AC1: 'Reply on RC1', Perla Alalam, 19 Jul 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-888/egusphere-2024-888-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Perla Alalam, 19 Jul 2024
Status: closed
-
RC1: 'Comment on egusphere-2024-888', Anonymous Referee #1, 22 Apr 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-888/egusphere-2024-888-RC1-supplement.pdf
-
AC1: 'Reply on RC1', Perla Alalam, 19 Jul 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-888/egusphere-2024-888-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Perla Alalam, 19 Jul 2024
-
RC2: 'Comment on egusphere-2024-888', Anonymous Referee #2, 31 May 2024
Review of paper "The role of refractive indices in measuring mineral dust with high-spectral resolution infrared satellite sounders: Application to the Gobi Desert" by Perla Alalam et al.
The paper is dedicated at assessing the influence of different determinations of mineral dust complex refractive indices (CRI) in the IR spectral range on the aerosol retrieval from IASI observations. Six different determinations of CRI are used to assess the capability of retrieving some dust properties (diameter and mixing ratio), and are applied to the retrieval for an intense dust storm occurred in 2017 over the Gobi area.The study addresses an important topic, because of the role that dust plays in the climate system through different mechanisms, as well as on the surface, atmospheric, and top of the atmosphere radiation budget.
Some aspects of the paper require additional investigation and a clearer presentation. In particular:
- some of the used methods are not well described and require more details or at least a brief summary.
In particular, there are no details or references to the principal component analysis method briefly described in lines 111-114, also applied elsewhere (e.g., l. 189-190).l. 94-99: the "V-shape" dust criterion is mentioned here, but a short description of the method and what is its sensitivity is needed (there is reference to a phd thesis, but a more specific reference would be preferable). Does a difference between two brightness temperatures necessarily imply a v-shaped spectrum? Are there specific assumptions on the spectral variability of surface properties needed?
- no verification data are presented, except for a generic comparison with measurements at two sites of the SONET network. Did the authors verify if other data are available, for example in the Skynet network (https://www.skynet-isdc.org/obs_sites.php)?
- the retrieval of the dust properties is made under several assumptions (single mode, if I understood well fixed standard deviation of the size distribution, uniform refractive index, fixed vertical distribution). Thus, although the correspondence between modeled and measured IR spectra is a relevant constraining method, the derived parameters might better be defined as radiativelly equivalent aerosol model, than specific aerosol properties. The possible influence of different distribution standard deviations and vertical dstributions (which has been shown to play a significant role on the satellite radiances, see e.g. Clarisse et al., 2019) should be addressed and investigated.
- Section 4 needs a better explanation and a clearer treatment. Also the notation is not clear. It is not clear why the AOD at 1020 nm is used here and how it is combined with determinations in the IR. Also here, all the calculations are made under the assumption that the only driving parameters are the median diameter and the concentration. How different factors affect the information content?
- Results shown in figure 6 might need a deeper discussion. Is there a reason for the large differences in the results obtained for DB17, DSC22 and VMA? All of these should somewhat take into account the specific mineralogy of the source. Two sets of CRI are for dry samples; might humidity play a role? Without validation data it is difficult to derive conclusions.
- The abstract needs to be improved, and some sentences need clarifications.
Minor comments are reported below.The same event in May 2017 was analysed in the paper by Alalam et al. (2022). This study seems an expanded sensitivity study of the same retrieval, leading to similar conclusions. It may be useful to highlight overlapping parts, differences, and new results.
Citation to references (e.g., parentheses) are not correct throughout the text.
line 8: emission should also be mentioned in addition to absorption and scattering
line 11: reference to IASI seems to be misplaced
l. 12: The sentence "this work reviews six prior ..." is misleading.
l. 16: here and elsewhere, the authors should clearly define what are "old" and what "new" CRI datasets. Please, indicate on what quantity you find a
decrease of total error by 30%. Reconstructed spectra should be mentioned.l. 27: please add the reference to IPCC.
l. 39: simulate spectral fits? Please, clarify.
l. 45: lack of size distribution knowledge in addtition to CRI is a relvant factor.
l. 39-59: it is not clear if the authors refer to determinations and calculations in the visible or IR or both. There is a wide literature on dust characterization
based on visible spectra. If they include visible spectra the discussion should be expanded.l. 69: is there e reference on the CESAM system? What is meant for "relevant atmospheric conditions"?
l. 75: see also comment to l. 16: please, define what are the new CRI measurements.
l. 83-84: accuracy of mineral dust retrievals: of which properties?
l. 93: "most dispersion and visibility to IASI observations". Please, clarify. Is it due to clouds? To the dust amount, vertical distribution, surface conditions?
l. 133: same as above: please specify what is intended for "atmospheric relevant conditions".
l. 151: it seems that using soil from the correct area (i.e., a more similar mineralogy) may have a strong impact as well.
l. 168: I would suggest using "median diameter" or "geometric mean diameter" instead of geometric size diameter.
l. 183: is the surface temperature retrieval affected by the occurrence of dust?
l. 194-195: if I understand well, the dust vertical distribution is assumed to be a vertically homogeneous layer between 1.5 and 2.5 km height. As discussed above, how is the assumption on the dust vertical distribution affecting the results? The authors state that a reduction of visibility was associated with this event. This does not seem fully compatible with a dust layer above 1.5 km altitude.
l. 246-247: is the effect of spectral stability included in SNR? It is expected to contribute.
l. 251: SNR set to 500: this requires a justification.
l. 260: please, explain better why the accuracy on albedo (in which spectral range?) is related with the uncertainty on surface emissivity.
l. 283: please, clarify what is the range of values needed to avoid saturation and to avoid loosing sensitivity.
l. 294: please, for clarity, use percent (used elsewhere) instead of fractional values for Sx.
l. 294-295: "The relative behavior of the CRIs ...": the sentence is unclear.
l. 313-314: "... is by tenth ...": sentence unclear.
l. 324-325: The sentence needs clarification. In my opinion, you can not use RMS on spectra as a measure of uncertainty on CRI. What would be the RMS corresponding to the measured spectra noise level (SNR=500)? This might give a reference value.
l. 325-326. "We selected spectra with large spectral features, whose simulation presents significant challenges and thereby provides a rigorous test for the CRIs" please, clarify. The sentence seems contradictory (you need large spectral features for detection).
l. 328-330: looking at fig. 5, it seems that for all cases and all CRI datasets there are spectral regions with model/measurement differences which probably exceed SNR=500 on measurements. May this provide useful suggestions on lacking knowledge on dust properties?
l. 330-331: although this might be the case, you might have compensating effects among different variables.
l. 339: is there an optimal range of absorption/extinction for detection? Is the largest extinction due to too much absorption (ie., imaginary part of CRI)?
l. 352: what was local time?
l. 354: "The mean value indicates the range of each parameter": please, clarify.
Citation: https://doi.org/10.5194/egusphere-2024-888-RC2 -
AC1: 'Reply on RC1', Perla Alalam, 19 Jul 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-888/egusphere-2024-888-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Perla Alalam, 19 Jul 2024
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
327 | 83 | 31 | 441 | 40 | 16 | 16 |
- HTML: 327
- PDF: 83
- XML: 31
- Total: 441
- Supplement: 40
- BibTeX: 16
- EndNote: 16
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1