Aridity synthesis for eight selected key regions of the global climate system during the last 60&thinsp;000 years

Fuhrmann, Florian; Diensberg, Benedikt; Gong, Xun; Lohmann, Gerrit; Sirocko, Frank

doi:https://doi.org/10.5194/cp-16-2221-2020

Articles | Volume 16, issue 6

https://doi.org/10.5194/cp-16-2221-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue:

Paleoclimate data synthesis and analysis of associated uncertainty...

https://doi.org/10.5194/cp-16-2221-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 16, issue 6

Research article

|

16 Nov 2020

Research article |

| 16 Nov 2020

Aridity synthesis for eight selected key regions of the global climate system during the last 60 000 years

Florian Fuhrmann, Benedikt Diensberg, Xun Gong, Gerrit Lohmann, and Frank Sirocko

Download

Final revised paper (published on 16 Nov 2020)
Supplement to the final revised paper
Preprint (discussion started on 27 Sep 2019)
Supplement to the preprint

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Comments', Anonymous Referee #1, 24 Oct 2019
- AC1: 'Point-by-point response to RC#1', Florian Fuhrmann, 07 Nov 2019
RC2: 'review#2', Anonymous Referee #2, 15 Nov 2019
- AC2: 'Point-by-point response to RC#2', Florian Fuhrmann, 18 Dec 2019
RC3: 'Review', Anonymous Referee #3, 22 Nov 2019
- AC3: 'Point-by-point response to RC#3', Florian Fuhrmann, 18 Dec 2019
RC4: 'review', Anonymous Referee #4, 30 Nov 2019
- AC4: 'Point-by-point response to RC#4', Florian Fuhrmann, 18 Dec 2019

Peer-review completion

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

ED: Reconsider after major revisions (20 Dec 2019) by Christian Ohlwein

AR by Florian Fuhrmann on behalf of the Authors (10 Jan 2020) Author's response Manuscript

ED: Referee Nomination & Report Request started (03 Feb 2020) by Christian Ohlwein

RR by Anonymous Referee #4 (10 Feb 2020)

Suggestions for revision or reasons for rejection

Dear Authors,
many thanks for your revisions.

In the replies to all four reviewers you mention the requirements of the PalMod Project. Apparently all reviewers consider these requirements difficult, and suggest other and scientifically more useful data selection. This should make you think how useful these constraints are, and if you can really derive robust aridity for all regions with these constraints. These constraints are clearly impacting on the robustness of your results (negatively, according to four reviewers), and I do think that this needs to be openly stated in the manuscript.

I do see the synthesis as an important contribution, and I do see that you carefully thought about how to combine information from different geoarchives.
I do see that you are trying to improve your manuscript, but when you are unable to make your data basis wider and follow reviewers’ suggestions, please discuss how the PalMod requirements impact on the robustness of your result – this is yet missing in the discussing part. Generally, I have the impression that you neglect reviewers’ well-meant suggestions in some cases without need.

Before publication I expect that the discussion needs to be more self-critical and needs to include statements for all regions how the PalMod requirements impact on/bias your overall result and the aridity synthesis. Without this, the paper should not be published in Climate of the past in my opinion.

Further, in the last reviews you were asked you to explain your methodology in a way that is 100% reproducible – from data to result. This still is not at all the case. As science should be reproducible, this is not acceptable as it is at the moment. I really demand from you to provide methods in detail and name where required software is available – possibly in Supplements. For speleothems you consider AGE uncertainty, for the other proxies you ‘estimate’ PROXY uncertainty – but you state that you use data on their original age models. This seems unlogical to mix. Please elaborate how you derive proxy uncertainty in a reproducible way.

Two reviewers mention that some of your your data selections are difficult or not useful for a local compilation. Please take these concerns serious and do NOT use the Palod requirements as reason to compare not-comparable datasets. These don’t force you to compare different things. Please don’t use the data from New Zealand and Australia together, I seriously think that mixing these is scientifically misleading. Possibly this reduces the number of regions – but that is in my opinion better than producing misleading results. The Lynch Crater is influenced by monsoons, NZ by westerlies. Although these may have similar temperature and precipitation, that is not comparable.

‘China’ is a political statement, not a geographic one – what about ‘east Asia’?

Hide

RR by Anonymous Referee #1 (13 Feb 2020)

Suggestions for revision or reasons for rejection

The authors indeed made some efforts to improve the manuscript from the first version. However, there are several aspects from my comments that in fact still remain unanswered, in particular concerning the methodology. Therefore I still recommend major revisions to the manuscript. This includes passages with typos and poor use of the English language.

SITE SELECTION

You now explain more clearly the PalMod requirements you stuck to for site selection, in particular the requirement for the datasets to be available in publicly accessible databases, and that they cover the last 60000 years. I recommend explaining also what the criteria were in terms of “high sampling resolution” and “highly cited”.

ARIDITY INDEX

You now try to better explain how the aridity index is calculated. You say it’s the sum of the three scores from the individual (speleothem, pollen, dust) proxy records, each one first scaled to the 0-100% interval, and then discretized as follows: 0/1 for speleothems, and 0/1/2 for pollen and dust (0-33%, 33-66%, 66-100%), resulting in an overall aridity index score between 0 and 5 (e.g. Table 1). Yet, in Figures 2, 3, etc. you have aridity index time series with non-integer values. This is clearly inconsistent with the definition you gave of the aridity index. Aren’t you by any chance linearly rescaling the 0-100% dust and pollen time series between 0 and 2?

I also do not understand how you merge the 3 individual time series, each one representing the “aridity score”. The three time series have different temporal resolution. Please explain how this is dealt with.

In connection to the point just above, how do you treat “no data” versus “0 or hiatus”? For instance, in the ELSA dust stack (Fig. 2e) you see no values for the last ~3000 years. Is this because there is no dust or no dated material?

The Central-European aridity index plotted in Fig. 2f is different (truncated) than in Fig. 3a.

You claim to be using one and only one record, per type of proxy, per region (with the exception of Santa Barbara basin), but in fact I still see two speleothem records in Figure 2. Please try to be consistent.

In connection to the point just above, you have also not explained how you deal with multiple records of the same type within each region. For instance, for Central Europe I still see both Bunker Cave and Spannagel Cave (Fig. 2a,b). Which one is used for calculating the Central Europe speleothem aridity score? If you used both, please justify why and explain how you merged the two records, i.e. based on AND / OR logical operator.

UNCERTAINTIES

This section is still too generic. What is an error for pollen and dust? (You use all kinds of dust proxies by the way, and I am not sure how you deal with the different cases). Mentioning that you learn from the experience of the ELSA stack is just not enough. Okay the description of the Montecarlo procedure, but where do you get the perturbations from in the first place? This needs to be spelled out clearly, and the actual values/ranges of these perturbations properly justified.

The uncertainties in Table 2 are expressed as percentages. With respect to what? Age or value of the proxy variable (i.e. x or y axis)? To the actual value of the aridity index in the 0-1 or 0-2 range for the individual proxies?

How do you combine these individual uncertainties into the overall aridity index uncertainty (the grey shaded area)?

RESULTS

In this section (and the corresponding supplementary sections for the different regions) you could indeed introduce in the discussion the relevant records that are not used for the calculation of the aridity index, for instance the records of Nussloch and Dunaszecsko you cited, and compare qualitatively with the records you actually used for the calculation of the aridity index.

Hide

RR by Anonymous Referee #3 (26 Feb 2020)

Suggestions for revision or reasons for rejection

Review of revised paper by Fuhrmann et al. “Aridity synthesis for 10 selected key regions of the global climate system during the last 60 000 years”

The manuscript by Fuhrmann et al. has been substantially improved. However, there are a few remaining issues that I strongly recommend to revise:

-P1/L26: “This paper emerges from the PalMod project which develops a long GCM time series of past global temperatures...”. PalMod is doing much more than just a long GCM time series of past temperatures!

-P1/L27: “One prerequisite of the project was to work only with publically available datasets from the most cited papers”. I doubt that “most cited” was a PalMod prerequisite. Personally, I also think that this is a very questionable criterion. Generally, the number of citations is not a faithful quality indicator.

-P2/L8: Note that the “...the boundaries of the MIS have been developed by Imbrie et al. (1984) and Martinson et al. (1987)” are outdated. The stacks provided by Lisiecki and Raymo (2005, Paleoceanography, 20, doi:10.1029/2004PA001071) or Lisiecki and Stern (2016, Paleoceanography, 10.1002/2016pa003002) are up to date.

-P2/L13: What is “highly cited”?

-P3/Fig 1: It seems that the location of the Susah Cave is wrong in Fig. 1 (c.f. Rogerson et al., Clim. Past, 15, 1757–1769, https://doi.org/10.5194/cp-15-1757-2019.). Also, it seems that Susah Cave is actually in the Mediterranean group? In Fig. 5, Susah Cave is listed under NW-Africa

-P4/L19: “The global climate evolution is well documented within Greenland and Antarctica ice cores”. I still disagree. For example, tropical precipitation is not represented in ice cores.

-P5/L3: “In sediments from the Cariaco Basin, the Al/Ti ratio gives the proportion between terrigenous river sediments with higher Al/Ti ratios and Saharan dust with respective lower Al/Ti ratios.” Please back this statement up by a citation or robust evidence. Yes, there is some Saharan dust arriving in the Caribbean, but this is negligible compared to river input, and I doubt that the dust signal is visible in the Cariaco Basin.

-P7/L25: “The Atlantic sea surface temperature pattern (caused by the Atlantic meridional overturning circulation - AMOC) strongly influence the whole European continent today”. Please revise this sentence. The Atlantic sea surface temperature pattern is not only caused by the AMOC.

-P7/L28/29; “... of central ??..”, “...with varved ??”

-P10/Fig. 3, P13/Fig. 5: As already argued in my first review I still find it very unreasonable to merge the Bahamas cave with the records from the Cariaco Basin and potentially (see comment on Fig 5 above) the Susah Cave (located at 33N/22E close to the Mediterranean) with NW African records. The prerequisites from PalMod cannot be relevant for my assessment of this paper.

Hide

ED: Reconsider after major revisions (09 Mar 2020) by Christian Ohlwein

AR by Florian Fuhrmann on behalf of the Authors (20 Apr 2020) Author's response Manuscript

ED: Referee Nomination & Report Request started (11 May 2020) by Christian Ohlwein

RR by Anonymous Referee #1 (28 May 2020)

RR by Anonymous Referee #4 (01 Jun 2020)

Suggestions for revision or reasons for rejection

Dear Authors,
Still, I regard the idea of synthesising information on aridity useful, and probably your results are valid by now. However, I still have issues following your methods and uncertainty assessment. Please take the comments serious, science needs to be reproducible.

In contrast to your statement, I do not see a thorough discussion on how the Palmod requirements impact on your results. You discuss this in the beginning of chapter 4, but I am missing a statements if you generally expect the same results, or different results? You make a statement exclusively about Europe and one other reference archive. A more general statement would be very welcome.
Still I miss a reproducible description on how the aridity index was derived. You explain the method for the individual data treatment well, but you add additional information (page 11, lines 13ff). I would like to see a table (and a Figure in Supplements) how these adjust your aridity index, and the manuscript requires information on which papers you use for this purpose. At the moment you give only examples, which is clearly insufficient. It needs to be clear how your aridity index was created. Otherwise publication is not recommended. Please take this serious (this is the 3rd review round where I keep on asking for the same: an understandable and reproducible explanation of the aridity index and synthesis).

The first part of the introduction is an opinion, not a fact, please clarify this.
p.1 . line 26: what is ‘long’?
p. 2, l.12: why ‘highly cited’? That is not necessary a criterion of quality. You may discuss this in chapter 4 (as the usefulness of other PalMod requirements).
P2, l26f: I disagree, please read the papers on rivers and loess by Ian Smalley for example. Please note the large dust storms in the US in the last century in arable land with precipitation >>200mm/a.
p. 3, l.2: please number figures according to their occurrence in the text in a logical manner.
As Fig. 4 is the synthesis from Figs. 5 & 6. Logically, it should come after these.
p.3, last line: evidence for what?
p.4, last lines: please add a sentence why K/C is a useful dust proxy
p. 5, l. 25f: at least for China there is more than one dataset available, would it be useful to test these for consistency? You may discuss this.
Chapter 2.2. in general: Thank you for including this, I do see an estimation of uncertainty as useful and pragmatic approach, but it needs better explanation. How do you define ‘error’? In Table 2 you seem to use ‘error’ and ‘uncertainty’ for the same thing. Is this useful? Uncertainty may not be an error. Further, which uncertainties are you referring to in Table 2? 1 Sigma, 2 Sigma? Of what distribution? Why are these different for different regions? Are these referring to proxy uncertainty or age uncertainty or both? This needs to be clarified in a manner that non-statisticians understand what you do, and why you do so.
p.5 line 11: please specify “These estimates are based on the experience of the ELSA pollen record”. This is NOT an acceptable statement, you want to present a scientifically robust aridity index, and not an index based on experience. Therefore, please explain what experiences were made, what is learned, and why you use uncertainties which you use. This is really necessary.
p.5 lines 13f: please refer to data, not to axes.
P5 line 20: please be more specific: which probability? Which distribution?
p. 11, l.9: please specify ‘Holocene’, so far Im unsure I can see this.
P. 12: l. 14: can you specify the influence of Heinrich events?
P. 17, l. 10f: the statement on loess is not generally true. See e.g. Yang & Ding 2014, the opposite seems the case in most of Asia.
I like your paragraph 4.1, adding a figure may be a useful addition?  your choice
Caption to Fig.7: anomalies relative to what?
Conclusions generally: These may be written in a nicer form, and may be extended a bit. You may refer to figures here and add why you come to these conclusions briefly.
Conclusion: (1) – seems not the case for the Santa Barbara Basin according to your synthesis, please correct
Conclusion: (2) you find similarity and infer cause – please phrase more self-critical and scientifically
Conclusion: (3) – really ‘easily’? This paper finds a way how to do so because it is NOT straightforward
Conclusion: (5): second part is true at the moment, but better data may improve on this. What is ‘large scale’?

Hide

ED: Reconsider after major revisions (27 Jun 2020) by Christian Ohlwein

AR by Florian Fuhrmann on behalf of the Authors (09 Jul 2020) Author's response Manuscript

ED: Referee Nomination & Report Request started (13 Aug 2020) by Christian Ohlwein

RR by Anonymous Referee #4 (02 Sep 2020)

ED: Publish subject to minor revisions (review by editor) (25 Sep 2020) by Christian Ohlwein

AR by Florian Fuhrmann on behalf of the Authors (30 Sep 2020) Manuscript

ED: Publish as is (05 Oct 2020) by Christian Ohlwein

AR by Florian Fuhrmann on behalf of the Authors (08 Oct 2020) Manuscript

Short summary

Proxy data of sediment cores, speleothem, pollen and isotope data were used to reconstruct past aridity of eight regions of the world over the last 60 000 years. These regions show humid conditions during the early MIS3 (60 to 45 ka). Also the early Holocene (14 to 6 ka) was humid throughout the regions. In contrast, MIS2 and the LGM were arid in Northern Nemisphere records. On- and offsets of aridity/humidity differ between the regions. All this is in good agreement with recent model results.