Impact of 3D Cloud Structures on the Atmospheric Trace Gas Products from UV-VIS Sounders – Part II: impact on NO2 retrieval and mitigation strategies
- 1Royal Belgian Institute for Space Aeronomy (BIRA-IASB), Brussels, Belgium
- 2Ludwig-Maximilians-University (LMU), Meteorological Institute, Munich, Germany
- 3Norwegian Institute for Air Research (NILU), Kjeller, Norway
- 4ESA-ESTEC, Noordwijk, the Netherlands
- 1Royal Belgian Institute for Space Aeronomy (BIRA-IASB), Brussels, Belgium
- 2Ludwig-Maximilians-University (LMU), Meteorological Institute, Munich, Germany
- 3Norwegian Institute for Air Research (NILU), Kjeller, Norway
- 4ESA-ESTEC, Noordwijk, the Netherlands
Abstract. Operational retrievals of tropospheric trace gases from space-borne spectrometers are based on one-dimensional radiative transfer models. To minimize cloud effects, trace gas retrievals generally implement Lambertian cloud models based on radiometric cloud fraction estimates and photon path length corrections. The latter relies on measurements of the oxygen collision pair (O2-O2) absorption at 477 nm or on the oxygen A-band around 760 nm. In reality however, the impact of clouds is much more complex, involving unresolved sub-pixel clouds, scattering of clouds in neighboring pixels and cloud shadow effects, such that unresolved three-dimensional effects due to clouds may introduce significant biases in trace gas retrievals. In order to quantify this impact, we study NO2 as a trace gas example, and apply standard retrieval methods including approximate cloud corrections to synthetic data generated by the state-of-the-art three-dimensional Monte Carlo radiative transfer model MYSTIC. A sensitivity study is performed for simulations including a box-cloud, and the dependency on various parameters is investigated. The most significant bias is found for cloud shadow effects under polluted conditions. Biases depend strongly on cloud shadow fraction, NO2 profile, cloud optical thickness, solar zenith angle, and surface albedo. Several approaches to correct NO2 retrievals under cloud shadow conditions are explored. We find that air mass factors calculated using fitted surface albedo or corrected using the O2-O2 slant column density can partly mitigate cloud shadow effects. However, these approaches are limited to cloud-free pixels affected by surrounding clouds. A parameterization approach is presented based on relationships derived from the sensitivity study. This allows identifying measurements for which the standard NO2 retrieval produces a significant bias, and therefore provides a way to improve the current data flagging approach.
Huan Yu et al.
Status: final response (author comments only)
-
RC1: 'Comment on amt-2021-338', Anonymous Referee #1, 14 Jan 2022
Review of “Impact of 3D Cloud Structures on the Atmospheric Trace Gas Products from UV-VIS Sounders - Part II: impact on NO2 retrieval and mitigation strategies” by Yu et al.
In this manuscript, the authors report on a study quantifying the effect of 3d cloud effects on satellite retrievals of tropospheric NO2. Such effects are currently not taken into consideration but are expected to play a role at the spatial resolutions of current and future instruments. The study identifies and quantifies different effects on simple synthetic test data concluding that the cloud shadow effect is dominant. This error is then analysed systematically and different methods are developed to correct it using albedo fitting, O2-O2 scaling or a parametrisation approach. All of them result in a reduction of the errors at least in the clear sky part of the measurements, both in the simple cases studies and in a complex cloud scenario. Finally, two case studies are presented on real TROPOMI data indicating that the developed approaches might improve NO2 retrievals close to cloud edges.
The manuscript is very interesting, clearly structured and written and contains a wealth of good ideas and relevant results. I, therefore, recommend publication with only minor revisions as listed below.
Detailed comments:
It would be good to have a short discussion of what the expected effect of aerosols is on the discussed 3d effects which here are discussed in a Rayleigh atmosphere.
Page 2, last line: This sentence is a bit unclear as spatial heterogeneity will also be relevant in clear sky scenes and several effects are addressed at the same time here. Please separate into two (or more) sentences.
Page 3, line 14 / 15: It would be nice to have a very brief indication also of what Várnai et al. found in their work.
Page 9, line 21: I think it would be good to iterate here that only one aspect of possible errors introduced by cloud correction is covered. Perfect knowledge of all parameters is assumed and in particular, the NO2 profile is assumed to be the same inside and outside of the cloud.
Figure 2: I think that this display is somewhat misleading – I was tempted to see points close to the 1:1 line as “good” points while in reality, they are just points for which both cloud retrievals perform similarly. The main point of the discussion here is how large errors are and I think histograms of relative errors would be more appropriate.
Figure 10: It would be nice to have the same x-axis in both plots to allow direct comparison
Section 4.1.1 It would be interesting to add a short discussion of what you think about the surface albedo fitting implemented in the current TROPOMI lv2 product where the surface albedo is determined from radiance in case it is lower than the climatological value for a scene.
Cases where the retrieved albedo is 0 appear to be problematic – can you discuss this a bit more? Is that because the atmosphere is illuminated less than it would without cloud which reduces the backscattered intensity but does not change the layer AMF in the same way as a small albedo?
The application to TROPOMI data is based on the assumption that NO2 retrievals should yield the same column in cloudy and clear regions as well as in the cloud shadow. However, considering the reduced actinic flux in the cloud shadow (and the increased values inside the cloud), shouldn't we actually see differences?
-
AC1: 'Reply on RC1', Huan Yu, 13 Mar 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2021-338/amt-2021-338-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Huan Yu, 13 Mar 2022
-
RC2: 'Comment on amt-2021-338', Anonymous Referee #2, 03 Feb 2022
The study "Impact of 3D Cloud Structures on the Atmospheric Trace Gas Products from UV-VIS Sounders - Part II: impact on NO2 retrieval and mitigation strategies"
investigates the impact of 3D cloud structures - which are ignored so far - on the NO2 retrieval from satellite measurements.
The paper is generally well written and matches the scope of AMT.However, I see a fundamental flaw in the design of this study, as it is doing the second step before the first one:
The quantities directly affected by 3d cloud effects would be the retrieved cloud fraction and cloud height.
These quantities are generally used for calculating NO2 AMFs, and, as far as I understand, this should not be changed according to the authors.
But then it is essential to first check how far the cloud retrievals are affected by 3D effects before analysing the effects on trace gases.For instance, a cloud shadow causes lower reflectance. This might actually be dealt with in the existing algorithms if negative cloud fractions would be allowed. This way it might be actually quite simple to account for cloud shadow effects without introducing new concepts/quantities like CSF.
Also other 3d effects (clouds in neighboring pixels) will affect the cloud fraction and cloud height retrieved based on IPA. It would be interesting to see to which extent these "wrong" CF/CH parameters do the NO2 AMF correction intrinsically (such as aerosol effects being partly accounted for by the cloud algorithms yielding higher CF and lower CH than "reality").I would thus like the authors to add an analysis of 3D effects on the cloud products first. The further mitigation strategy might be different if 3D effects could already be accounted for by e.g. negative cloud fractions. In any case, the mitigation strategies cannot be discussed without knowledge on the effect of 3D cloud structures on the standard cloud products themselves.
Minor comments:
Page 1, Line 2: "generally implement Lambertian cloud models": This is not true, see for instance OCRA/ROCINN.
Page 1, Line 3: "photon path length corrections": to my understanding, the cloud algorithms interprete the measured O2 or O4 absorption in terms of a cloud height. This should be stated here.
Page 2, line 6: "amount of the trace gas along the average path": this sounds like the average path could be calculated and then linked to the amount of trace gas. It is rather the average absorption along light paths.
Page 2, line 19: "A simplified Lambertian cloud model is generally used": This is not true, see for instance OCRA/ROCINN.
As the remaining text might change considerably, I do not provide further detailed comments for now.-
AC2: 'Reply on RC2', Huan Yu, 13 Mar 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2021-338/amt-2021-338-AC2-supplement.pdf
-
EC1: 'Reply on AC2', Diego Loyola, 15 Mar 2022
Hola Huan,
Reading the reply to Referee #2, I had the impression that only the classical Lambertian cloud model (CRB) retrieval will be discussed in “AMF using extended cloud retrievals”. I strongly recommend to include in this section also a discussion of AMF calculations using the more realistic cloud-as-layer (CAL) model as it is done in Liu et al., 2021:
https://doi.org/10.5194/amt-14-7297-2021Chao
Diego-
AC3: 'Reply on EC1', Huan Yu, 10 Apr 2022
NO2 retrieval corrected by the cloud retrieval based on CAL model is added in the “AMF using extended cloud retrievals” part. Instead of OCRA/ROCINN cloud algorithm, a simple cloud retrieval approach is presented, which assumes the cloudy scenes are 100% covered by a uniform layer of water cloud with a 1-km geometrical thickness. The cloud single scattering albedo sets as 1 and the asymmetry parameter is 0.85, these values are consistent with those used in the cloud and NO2 retrieval (Liu et al., 2020, 2021). This approach retrieves cloud top pressure and optical thickness based on the measured reflectance at 460 nm and O2-O2 SCD or three 1-nm (758–759 nm, 760–761 nm and 765–766 nm) averaged radiances around O2-A band.
-
AC3: 'Reply on EC1', Huan Yu, 10 Apr 2022
-
EC1: 'Reply on AC2', Diego Loyola, 15 Mar 2022
-
AC2: 'Reply on RC2', Huan Yu, 13 Mar 2022
Huan Yu et al.
Huan Yu et al.
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
389 | 120 | 24 | 533 | 11 | 8 |
- HTML: 389
- PDF: 120
- XML: 24
- Total: 533
- BibTeX: 11
- EndNote: 8
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1