Investigations into the development of a satellite-based aerosol climate data record using ATSR-2, AATSR and AVHRR data over north-eastern China from 1987 to 2012

Che, Yahui; Guang, Jie; de Leeuw, Gerrit; Xue, Yong; Sun, Ling; Che, Huizheng

doi:https://doi.org/10.5194/amt-12-4091-2019

Articles | Volume 12, issue 7

https://doi.org/10.5194/amt-12-4091-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/amt-12-4091-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 12, issue 7

Research article

|

26 Jul 2019

Research article |

| 26 Jul 2019

Investigations into the development of a satellite-based aerosol climate data record using ATSR-2, AATSR and AVHRR data over north-eastern China from 1987 to 2012

Yahui Che, Jie Guang, Gerrit de Leeuw, Yong Xue, Ling Sun, and Huizheng Che

Download

Final revised paper (published on 26 Jul 2019)
Preprint (discussion started on 22 Jan 2019)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of ATSR/AVHRR AOD combination paper by Che et al', Andrew Sayer, 20 Feb 2019
- AC1: 'Response to reviewer 1', Jie Guang, 15 May 2019
- AC4: 'Revised Manuscript', Jie Guang, 15 May 2019
RC2: 'Discussion on "Investigations into the Development of a Satellite-Based Aerosol Climate Data Record using ATSR-2, AATSR and AVHRR data"', Anonymous Referee #2, 28 Feb 2019
- AC2: 'Response to reviewer 1', Jie Guang, 15 May 2019
- AC4: 'Revised Manuscript', Jie Guang, 15 May 2019
RC3: 'Review of amt-2019-26', Anonymous Referee #3, 16 Mar 2019
- AC3: 'Response to reviewer3', Jie Guang, 15 May 2019
- AC4: 'Revised Manuscript', Jie Guang, 15 May 2019

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Jie Guang on behalf of the Authors (15 May 2019)

ED: Referee Nomination & Report Request started (28 May 2019) by Thomas Eck

RR by Andrew Sayer (03 Jun 2019)

Suggestions for revision or reasons for rejection

I reviewed a previous version of this paper and posted the review under my own name (Andrew Sayer). My recommendation for that version was major revisions to have the paper better reflect its goal of bridging the ATSRs and SLSTR aerosol records. Briefly, the AATSR record ends in 2012 and the successor SLSTR launched in 2016, so there’s a 4-year gap, which the authors chose to try and fill for aerosol data over northeastern China using the old AVHRR sensor series (1980s onwards). MODIS satellite data are also used as a semi-reference here although it’s not clear that MODIS should be trusted more than ATSR.

In this revision the authors have addressed the technical concerns I raised, but in response to the conceptual issue (the paper was about bridging ATSR and SLSTR, but the reasons for using AVHRR initially were not strong, and no SLSTR data were shown) the analysis has been reframed in terms of using AVHRR as a way to continue the ATSR record back in time to the 1980s. So it’s the same analysis as before, but a different message. The main stated reason for changing focus is not having SLSTR data processed at present (which makes the original focus strange). I suppose that is ok, but I will admit I feel the original intent (bridging gaps going forwards) is the more scientifically interesting. I also raised an issue with the AVHRR algorithm surface model used but that was more or less ignored (acknowledged that it exists but declared as something for future work). I’m not really sure how satisfactory that is given that this is now a paper looking back at long-term consistency, when it’s these earlier AVHRR records which are the most problematic (see previous review). In this sense the study does seem premature since there are neither SLSTR data to provide a bridge going forwards, nor the robust AVHRR algorithm necessary to extend the data record backwards with confidence. Still there has been a lot of work here and the topic of long-term AOD variations is important, so it would be a shame to have nothing come out of it. So I am a bit conflicted in my recommendation and would leave that to the Editor and other reviewers.

There are a few remaining (mostly small technical) issues:

Section 2.2.1 – the text reads like the Xue and Deep Blue AVHRR data sets are the only land ones (e.g. page 5 lines 17). But there’s also some (in various states of availability) by groups including Mei et al, Gao et al, Hauser et al, etc.

Page 9 line 24: I would not described 0.06 as a “small” bias, and suggest removing that word.

Figure 2 (and later): Should the middle column say “ATSR_ADV” instead of “AVHRR_ADV”? It would also be better to have captions without underscores, e.g. write “AVHRR ADL AOD, JJA” instead of “AVHRR_ADL_AOD_JJA”. The algorithm abbreviations could also be dropped, since only one algorithm is applied to each sensor here.

Figure 3: The caption refers to least squares fit parameters. This text should be removed because these are not shown in the figure (to be clear: they should not be added to the figure; the caption should just be shortened).

Table 2: likewise “for a LSQ fit to all data points” can be deleted.

Discussion: the time series shown near Beijing all look in reasonable agreement, which is encouraging. But the later figures in the paper with larger maps (figure 9) show large systematic offsets between the data sets. I still feel that the discussion glosses over this, and just says that merging is “not straightforward”, which is something we knew beforehand. So I still don’t feel that it pushes forward with actionable recommendations that much. For example we still don’t know from this whether sampling or algorithm are more important or if they’re both important. The Sogacheva analyses cited did some three-way matchups (ATSR, MODIS, AERONET) to try to isolate algorithm from sampling effects, perhaps something like this could be done in this study too. If it’s sampling limits which dominate then perhaps we should try to glue on a daily basis rather than monthly. This could have been tested in this study but wasn’t. The analysis is framed now as extending ATSR backwards with AVHRR but there’s still no combination of the records done or attempt to create an extended data set, just a comparison of time series. Perhaps this is enough of a step for now, but as I indicated earlier, I’m not certain what to recommend here so would suggest the Editor draw on additional reviewers if they did not already.

Hide

RR by Anonymous Referee #3 (20 Jun 2019)

ED: Publish subject to technical corrections (21 Jun 2019) by Thomas Eck

AR by Jie Guang on behalf of the Authors (27 Jun 2019) Author's response Manuscript

Short summary

The use of AOD data retrieved from ATSR-2, AATSR and AVHRR to produce a very long time series is investigated. The study is made over a small area in northern China with a large variation of AOD values. Sun photometer data from AERONET and CARSNET and radiance-derived AOD are used as reference. The results show that all data sets compare well. However, AVHRR underestimates high AOD (mainly occurring in summer) but performs better than (A)ATSR in winter.