the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Dec 2023
Climate variability, heat distribution and polar amplification in the unipolar ‘doubthouse’ of the Oligocene
Abstract. The Oligocene (33.9–23.03 Ma) was characterised by generally warm climates, with flattened meridional temperature gradients while Antarctica retained a significant cryosphere. This makes the Oligocene an imperfect analogue to long-term future climate states with unipolar icehouse conditions. Although local and regional climate and environmental reconstructions of Oligocene conditions are available, the community lacks synthesis of regional reconstructions. In order to provide a comprehensive overview of marine and terrestrial climate and environmental conditions in the Oligocene, as well as a reconstruction of trends through time, we here review marine and terrestrial proxy records and numerical climate model simulations of the Oligocene. Results display weaker temperature gradients during the Oligocene compared to modern times, with generally warm poles and colder-than-modern temperatures around the equator. Sea surface temperatures (SSTs) show similar trends to the land temperatures, with warm temperatures around the mid and high latitudes (~60–90°) of especially the Southern Hemisphere. Vegetation-based precipitation reconstructions of the Oligocene suggest drier conditions compared to modern times, in particular around the equator. When compared to proxy-based data, climate modelling approaches overestimate Oligocene precipitation in most areas, in particular the tropics. Temperature around the mid to high latitudes is generally underestimated in models compared to proxy data and tend to overestimate the warming in the tropics. In line with previous conclusions models underestimate polar amplification and the equator-to-pole heat distribution that prevailed during the Oligocene. Despite prevalent glaciation on Antarctica, the Oligocene “icehouse” experienced warm global average temperatures while still maintaining a unipolar icehouse state.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(5237 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5237 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
icehouseclimate. We extend an existing marine climate proxy data compilation and present a new compilation and analysis of terrestrial plant assemblages to assess long-term climate trends and variability. Our data–climate model comparison reinforces the notion that models underestimate polar amplification of Oligocene climates, and we identify potential future research directions.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-2738', Anonymous Referee #1, 15 Jan 2024
Review of “Climate variability, heat distribution and polar amplification in the unipolar ‘doubthouse’ of the Oligocene”
General comments:
The purpose of the paper is confusing. It wasn’t clear from the start that it is a literature review and interpretation. The paper also tries to do too much by reviewing the chronostratigraphy, the geography, model-data comparison, fauna and flora, introducing new data using NRL, etc., etc. This could be mitigated by stating earlier (title, abstract, intro) and more clearly the purpose and structure of the paper and by restructuring sections of the paper (see below). The discussion/conclusion sections are a little disappointing. The discussion section doesn’t offer any insights into the mysteries or speculation about the remaining Oligocene mysteries (e.g. model-data mismatch, CO2-climate mismatch). The conclusion section was obviously a last-minute addition and needs to be redone. Critiques aside, I appreciated the thorough overview and learned some new things about the Oligocene. This could be a nice addition with some additional work.
Title:
The title is misleading. This is a review paper and the title should reflect that.
0 Abstract:
Page 2, Line 14-15 – I don’t understand this sentence. Wouldn’t the Oligocene be a good analogue for a future climate state with unipolar glaciation?
Page 2, Line 28 – Delete “while still maintaining a unipolar icehouse state.”
1 Introduction:Most of the paper is well written. The Introduction is the exception and requires careful editing.
Page 3, Line 4 – “equilibrate” is used incorrectly here.
Page 3, Line 43-44 - Awkward sentence. Please fix.
Page 3, Line 51-52 - Awkward phrasing. Please fix.
Page 4, Line 80 - I’m not sure why whales are mentioned here.
Page 5, Figure 1 caption - Maybe the colors didn’t transfer correctly, but some of the colors and shapes mentioned in the caption don’t match the colors in the legend of the figure (for (b)). Misspelling of “Transition” as well.
Page 6, Line 94 – The authors mention variability in continental ice volume paced by eccentricity and obliquity. The authors should follow that with an explanation of why/if that’s significant and how it might be seen in the proxy records (given that they are low resolution).
Page 6, Line 99 – “Here..” Should be start of a new paragraph.
Page 6, Line 100 – “after a chronostratigraphic section” Awkward phrasing.
Page 6, Lines 99-104 - At the end of the introduction, the authors finally explain that the purpose of this paper is to review the current state of knowledge regarding the Oligocene climate and thus provide constraints on boundary conditions, compile proxy data, and identify points of interest. This point should be made much sooner so that readers understand the purpose of the paper.
Page 6, Lines 99-104 - There is no mention of numerical climate simulations as in the abstract. Model-data comparisons are made later in the paper, and should be mentioned here. There is also no mention of constraining ocean circulation or ice dynamics though a large part of the introduction discusses these. The authors need to make clearer how climate simulations are used in this paper with some quick details on them, if anything on the ocean or ice state of the Oligocene is explored further, and potentially more details on the proxy records used and for what reason? There is not much on the purpose of the paper and it’s all at the very end of the introduction - they could expand on what the paper will discuss and could mention the main purpose it serves sooner in the introduction.2 Oligocene chronostratigraphy and event nomenclature:
This section should be removed completely from the paper or moved to Supplemental Data.
Page 7, Line 112 – “GGSP” should be “GSSP”, right?
Page 7, Line 116 - The OMB is stated as 23.04 Ma here and 23.03 Ma in the rest of the sections
Page 8, Line 134 – Spell out ‘GPTS’ as ‘geomagnetic polarity time scale (GPTS)’
Page 9, Line 177 - Why does the last zone end at 23.07 Ma when the Oligocene ends at 23.03 or 23.04 Ma?
Page 10, Line 196, 198, 201 - Why is there a question mark after ‘Axoprunum’?3 Boundary conditions for Oligocene climate:
Page 11, Line 236 – “boundary conditions” is a modeling term and inappropriate here. Replace with “geography” or “continental position”
Page 12, Figure 3 caption – Add an elevation scale bar. Also ‘NRL’ is supposed to be ‘NLR’.
Page 13, Line 266 – “strongly affected regional ocean circulation…” Here the authors are repeating a hypothesis that is by no means certain and should equivocate appropriately “may have strongly affected”. They do this throughout the paper, stating hypotheses, sometimes model results, as fact or certainty. Please clarify appropriately.
Page 15, Line 348 – The authors mention an increase in ocean methane hydrates, peat, and wetlands as a result of colder temperatures and ice expansion. They could add a sentence that explains how the expansion of those environments are a consequence of colder temperatures and if there is evidence of these expanding during the Oligocene or if it is just a probable guess given the cooling climate.
Page 15, Line 358-380 – How does these CO2 levels/trends compare with the new CO2 from Honisch et al., Science?
Page 16, Line 377 - “16esults” typo4 Climate proxy data:
Page 16, Line 385 - Why did the authors choose fossil plant remains to analyze and why only fossil plant remains (not paleosol carbonates, etc)?
Page 17, Line 405-407 – The authors mention an absence of high latitude data (no fossil plant remains suitable for their method). How could this absence affect their results, or how much uncertainty does this add when estimating a meridional temperature gradient for the Oligocene?
Page 17, Line 415 - One data point in the NH mid-latitudes is significantly different from the other mid-latitude data points - where is this NH data point? What sort of biome is it from?
Page 18, Figure 4 – Use the same y-axis for each plot. As it lies now, the winter temperatures appear the same as the MATs until you see the slight difference in the y-axes. Also, in the figure caption, it states the (b) plot PI values are in brown but they appear gray just like in plot (a).
Page 19, Line 460 - Random parenthesis in statement. Also, the authors could add one more sentence on how pH affects the fractionation of oxygen isotopes in this environment (why an overestimation of 1.5C has to be taken into consideration?).
Page 20, Figure 5 - This is a nice figure, though the shapes are so small I can’t really place many of the sites to a line. They could instead do just three different shapes for each type of SST proxy. There are also some SST sites in Figure 3 that aren’t in Figure 5 (i.e. Site 913 and others), if they aren’t included in Figure 5 then are they still used / needed in Figure 3? Why weren’t some used in Figure 5?
Page 21, Figure 6 - This is a great figure, very concise with a lot of information. It’s interesting too that SST seems ~constant from about 40S to 40N for the Oligocene and sort of late Eocene as well, but modern SSTs have a narrower curve peaking from about 20S to 20N. They could mention something on this.
Page 22, Figure 7 - Why reconstruct driest month precipitation and not wettest month precipitation? Along those lines, why reconstruct winter temperatures and not summer temperatures (Figure 4)? Could just include one sentence stating the benefits of understanding winter temperatures and driest month precipitation.
Page 22, Line 496 – The authors generally state that the Oligocene MAP values are similar to modern day, but the mid-latitude values (for either hemisphere) seem higher than modern values (Figure 7)?
Pages 23 and 24 – The authors don’t mention similarities or differences between the models here at all (HadCM3L vs CESM). According to Figure 8, it seems there are a lot of similarities between how the models reconstruct surface temperature compared to the proxy records. It’s interesting that you see such similarities between CESM and HadCM3L (though on that - parts C/F/I of Figure 8 could use a darker color for CESM mean temperature difference since the light pink is a bit difficult to see).
Page 24, Line 533 – The authors say the differences plotted are ‘calculated as a pointwise difference between the proxy mean value and the model annual mean’. Two thoughts on this: a) how was the model annual mean calculated (was the annual mean for the corresponding grid cell used or the general region / a group of grid cells) and b) are there any seasonal biases within the proxy records that could be causing greater differences since only annual means are used for the simulations? If not, they could just state somewhere that there are no known seasonal biases for these records.
Pages 25 and 26 - Similar to what I said above, similarities and differences between CESM and HadCM3L could be mentioned, I think the light pink color for CESM in Figure 9 parts C/F/I is too difficult to see, and I’m a bit confused on the ‘pointwise difference’ method regarding which grid cells were used for the model values. I also think red should correspond with drier (according to data) conditions (negative precipitation difference) and blue should correspond with wetter (according to data) conditions (positive precipitation difference). It’s a bit confusing that we are looking at precipitation as a ‘data minus model’ value where the data usually represents wetter conditions than the model simulates and yet the data points on the plot are in red colors which traditionally represent drier conditions but this is really just a preference.
Page 27, Line 583 – See inappropriate capitalizations. “Is of great Importance”
Page 27, Section 4.4 – This section on Ice Sheets is out of place. Unlike the rest of Section 4, it has no original data or interpretation. It’s simply a literature review and should be included in Section 3.5 Flora and faunal changes:
Similar to last comment, why is this section not with the literature review in section 3?
6 Discussion:
Page 31 – The authors mention there’s a discrepancy between SH and NH MAP (in Oligocene and not in modern) in the result section (Figure 7). A discussion of why would be welcome here.
Page 31, line 686-687 – The authors conclude that the Oligocene was anomalously warm relative to CO2 levels. They should discuss potential explanations for this apparent mismatch.
7 Conclusions:
It reads as though the authors ran out of energy before the Conclusions. There’s no introductory statement and the bulleted list is very sparse. Besides the abstract, the conclusions are the most frequently read section of a paper. The authors don’t state that the manuscript is a review or distinguish between their contributions and what’s review.
Page 31, Line 704 - The second conclusion is specific to the proxy records, not so much what the simulations show, so maybe mention that, though I guess they have a conclusion point specific to the simulations below.
Page 31, Line 706 - I generally agree with the third conclusion, except Oligocene MAP values around the mid-latitudes appear higher than modern MAP mid-latitude values according to Figure 7.
Page 31, Line 711 - In line with the last conclusion point, the authors could mention that that may be related to the lack/variability/uncertainty in Oligocene CO2 records. CO2 is not well constrained so perhaps the Oligocene is, in part, not the “icehouse” it seemed to be before because CO2 was, in reality, a bit higher than some older records estimated.Citation: https://doi.org/10.5194/egusphere-2023-2738-RC1 - AC1: 'Reply on RC1', Dominique Jenny, 16 Feb 2024
-
RC2: 'Comment on egusphere-2023-2738', Anonymous Referee #2, 17 Jan 2024
Jenny et al present a review of the Oligocene, an 11-million-year long time interval. The review is concerned with:
(1) a general introduction,
(2) Chronostratigraphy and event nomenclature,
(2.1) Planktonic foraminifera zones,
(2.2) Calcareous nannofossil zones,
(2.3) Radiolarian biostratigraphic zones,
(2.4) Dinoflagellate cyst zones,
(3) Boundary conditions for Oligocene climate,
(3.1) Geographical boundary Conditions,
(3.2) Ocean circulation,
(3.4) Carbon cycle,
(4) Climate Proxy data,
(4.1) Temperature,
(4.1.1) Continental Mean Annual Temperature,
(4.1.2) Sea Surface Temperature,
(4.2) Precipitation,
(4.3) Data-model comparison,
(4.3.1) Temperature proxy to model comparison,
(4.3.2) Precipitation proxy to model comparison,
(4.4) Ice sheets,
(5) Flora and Faunal changes,
(6) Discussion,
(Note, 6.1 is missing).
(6.2) Temperature trends and variability,
(6.3) Precipitation, and
(7) Conclusions.
This review constitutes a very ambitious, admirable, and important project, that necessarily needs to make broad-brush summaries and interpretations about the main climatic and oceanographic factors that determined the Oligocene palaeoclimate system. In principle, I am sympathetic to this effort. The paper is well written.
However, I believe that for this paper to become an authoritative reference paper much more work is needed. I am no expert on all topics reviewed here. But on topics on which I do consider myself reasonably knowledgeable I find that they are often marked by too many inaccuracies, some outdated discussion points, and several key omissions. By proxy, I have therefore little faith that the parts of the review concerned with topics on which I am no expert, accurately capture the current state of knowledge. I am left with the impression that this paper ‘could have been so much better’.
To me the paper suffers from not having included enough Oligocene workers, many of whom have dedicated their professional lives to understanding topics ranging from age models, marine microfossils (calcareous nannofossils, foraminifera, radiolaria, etc.), macrofossils (animals and plants), ice sheet modelling, palaeogeography, etc. Compare, for example, this Oligocene review to Steinthorsdottir et al., 2021, who consulted 3 to 4 times as many experts for their Miocene review. (Here my criticism is mainly directed at coauthors Bijl, Huber, and Sluijs; all experienced and senior workers, some of whom have been involved in the Steinthorsdottir effort. To me, and as far as I can judge, it seems that with greater input from these authors, also in study design, the manuscript could have been lifted to a higher level already. Jenny’s efforts in leading this study, as a PhD student, is admirable to say the least.)
I believe that revising this manuscript will result in a much better paper. To me the authors have two options: Option 1) shorten the manuscript considerably by removing much of the review text that is not pertinent to the temperature/precipitation compilation and model comparison, i.e. what I consider the main strength of this paper, and expertise of the authors, or Option 2) include more expert knowledge, work with the text and develop the discussion to lift this paper to a higher level. I recommend the second option, and I believe it could make, after substantial revisions, for an authoritative Oligocene reference paper. Ultimately, it is the author’s choice.
I will provide some below and wish the authors much success with revising this document and planning the next steps ahead. I am sorry that I am not more positive in this instance.
Major comments
- Structure of the paper: I believe that inclusion of subheadings concerned with magnetostratigraphy, astrochronology, oxygen and carbon isotope stratigraphy would strengthen the paper. Perhaps compare to Steinthorsdottir et al., 2021, to see how they went about their Miocene review. If this comparison is unjustified, please say so in the rebuttal.
- Doubthouse/Unipolar Doubthouse: Most workers use “Doubthouse” to describe the latest Eocene (between the MECO and EOT), where the climate system was cooling and “doubting” to jump from a “Greenhouse” state to a “Unipolar Icehouse” state. I much prefer to reserve the phrase “Doubthouse” for the Late Eocene, and the usage of “Unipolar Icehouse” to describe the state of the Oligocene to Pliocene time interval. As an aside, I also prefer “Unipolar Icehouse” over the phrase “Coolhouse”. The usage of “Unipolar Doubthouse” in your title is a mixture of “Unipolar Icehouse” and “Doubthouse” and is confusing to me. I am not entirely sure what is meant by this phrase. Is the system “doubting” to be unipolarly glaciated? Or “doubting” to go back to a “Greenhouse” state, or to a “Bipolar Icehouse” state?
- Figure 1: This figure has many inaccuracies. Almost all boundaries or events (apart from those around the EOT, which seem to be OK) are indicted in the wrong place. Furthermore, the MOG is not a previously defined event, the MOGI is. The MOGI as drawn does not cover the MOGI as defined in the literature. The OMT does not cover the largest glaciation across the OMB. The Mi-1 points to an d18O low, not a d18O high, and moreover it is best to abandon Mi and Oi nomenclature completely (more about this below).
- Line 108/109, the Oligocene time scale: For an Oligocene review paper, it is important to get the age scale right, and to describe it explicitly (not just mentioning it in a figure caption). The absolute ages of the polarity time scale in the GTS2012 (i.e., the GPTS2012) are of greater quality than those incorporated in the GTS2020/GPTS2020, especially for the latest Oligocene where the GTS2020 incorporated erroneous ages. Admittedly, large efforts have been put toward improving biostratigraphic markers and zonations for the GTS2020, but also their absolute ages are unfortunately too often based on an inferior GPTS2020 scale (compared to the GPTS2012). I recommend using GPTS2012 to any Oligocene (age model) worker, and an Oligocene review paper needs to address the errors in the G(P)TS2020, if the authors decide to stick with this age scale as their preferred option.
- Line 353 to 380: This CO2 discussion needs more work. What are the implications for radiative forcing? For example, how to explain the Late Oligocene Warming whilst CO2 is decreasing? This is a major topic in Oligocene research, and it is not mentioned or discussed. For this manuscript to become an authoritative review, such a major climatic enigma needs to be addressed, because it will help with formulating new and important research questions.
- Discussion and Conclusions, both seem quite superficial to me. Can we take a step back, and distil major research questions for the Oligocene?
Minor comments
- Line 13: Perhaps start the abstract with why the Oligocene is of any interest to anybody? It may help draw some more readers in.
- Line 14: Not entirely sure why this would lead from the first line of the abstract. To me the gradients could be a good analogue. Ice sheet geometry, palaeogeography, etc, seem to me the limiting factors in the Oligocene being a palaeo analogy for long term future climate states.
- Line 23: proxy data instead of proxy based data?
- Line 25: “In line with previous…” Does this refer to Oligocene modelling efforts? Or modelling of warmer climates in the general. I thought that this was a feature of Eocene modelling output but was unaware of this being a problem in general, also for the Oligocene unipolar icehouse. Perhaps clarify?
- Line 23 to 27: should we learn a lesson from these observations? Is it worth making that lesson explicit in the abstract (e.g., much more effort needs to be invested in improving Oligocene climate simulations?)
- 1: Position Mi-1 is wrong. In general, it is best to refrain from using Oi and Mi numbering. Oi and Mi numbers were originally defined (in very low-resolution records from the 1980s and early 1990s) as oxygen isotope zones, like biozones, that lasted several hundreds of thousands to millions of years (see original Miller papers and Wright papers). Over the years, these zone numbers were used to refer to the oxygen isotope “highs” (often called “events”) at the base of the original oxygen isotope zones, yet this is not how they were originally defined. The arrows in Fig 1, suggest that the authors also use the Oi and Mi numbers to indicate events. If this usage is adopted, which I do not recommend, then please refer to high oxygen isotope values. E.g., the position of the Mi-1 is not correct in this figure (you refer to an oxygen isotope low). Furthermore, even when the Mi-1 (short for “Miocene oxygen isotope zone 1”) is indicated correctly in Fig. 1, it would still fall in the latest Oligocene (yet another reason to abandon Oi and Mi numbers). Lastly, nobody has even been able to explain to me what makes the Oi1b, Oi2a, and Oi2b stand out as particularly interesting, or exceptional “events”, probably because these numbers were devised as zones originally.
- 1: MOGI instead of MOG? Also, this band is not drawn according to the original definition. You may choose to redefine the main intervals of interest, but then they need to be defined in the text.
- Line 90: Oligoene (missing c)
- Line 90: MOG, where G stands for “glacial”. In MOGI, this stands for Glacial Interval, and refers to a longer period of mean elevated d18O values (punctuated by shorter lasting d18O decreases of about 1 per mil amplitude, interpreted to be Antarctic “interglacials” and/or deep-water warming events). The use of glacial in MOG is a bit strange and makes me think of something more similar to the last Pleistocene glacial cycle.
- Line 99: See Major Comment #1
- Line 109: I strongly recommend using G(P)TS2012 over G(P)TS2020. See Major Comment #3.
- Line 115: Beddow et al., 2018 did not work on the OM Boundary type section (Lemme Carrosio) but derived an astronomically calibrated age for the base of C6Cn.2n from Pacific sediments (perhaps make that clearer, because the current sentence can be interpreted as if Beddow et al did work on that section). Also, the 100 kyr error is not mentioned by Beddow et al. and seems too large.
- Line 110 to 125: This section would be much strengthened with the addition of a description of astrochronozones and potentially astro-unit-stratotypes. GSSPs are useful to an extent, but have their limitations (see e.g., Hilgen et al., 2020, Newsl. On stratigraphy). Potentially add a magnetostratigraphic description as well
- Lines 126 to 145: Overall a good summary, however, as stated above, I would shy away from reiterating Oi and Mi numbers. EOT, MOGI, and OMT are more useful terms in my opinion. Potentially in addition with the Late Oligocene Warming (LOW). MOGI definition used here, instead of MOG in figure caption. Please note that the MOGI is defined as “a generally cold but highly unstable mid-Oligocene time interval (∼0 My to 26.3 My ago), which we refer to as the Mid-Oligocene Glacial Interval (MOGI)”, not as “a ~1 Ma year phase of profound cooling/glacial expansion”. Usage of Ma is erroneous in this context too (Ma refers to an age not a duration). In general, this section needs some work/detailed checking, as I have not checked all statements, and I may have overlooked other potential errors.
- Section 2.1. The GTS2020 is abandoned here?
- Section 2.2. Similar comment. Agnini et al., 2014, cannot have used GTS2020. Please mention on which age scale these bioevents are given.
- Section 2.3. Same here.
- Section 2.4. and here. For Chapter 2, in general, best to refer to an accepted general age scale. I recommend G(P)TS2012. Where biostratigraphic workers have deviated from these scales, it is important to mention so, and to estimate how large uncertainties/discrepancies are between scales.
- Line 329 to 352: This discussion is only concerned with the EOT interval. What about the 400 kyr, and 2.4 Myr cycle that dominate the Oligocene benthic d13C records?
- Line 353 to 370. What about the drop in CO2 at around 24 Ma, during a 1.2 Myr Obliquity cycle “node”. May this drop have preconditioned the system for the (large amplitude of the) OMT.
- Line 377: 16esults. Typo?
- Figure 5: OMT, MOGI, EOGM: please check ages.
- Line 647 to 657: Very qualitative language. Warm, warmer, etc. Can we not put some numbers to how warm the Oligocene was?
- Line 658 to 672: this section is mixing things up. It starts with a description of high resolution benthic d18O records and compilations. Then merges into a discussion of low-resolution SST data, with a reference to Fig. 5 as the only give-away that the topic has changed (admittedly still temperature, but now surface instead of deep ocean). It then comes back to comparing low-res SST (without mentioning that that is what is being compared) to benthic d18O.
- Line 660: “most likely affected by a shift of geographical”. As far as I am aware there is no doubt about it. Not sure if it is still meaningful to dig out this splicing issue in Zachos 2001. We are two good compilations further (Zachos 2008 and Westerhold 2020). Also confusing this with a true signal in both the Atlantic and Pacific benthic d18O of the Late Oligocene Warming is unhelpful. Is it not smarter to have a section in the paper that discusses benthic d18O. To me, the striking absence of a LOW signal in SST is something more interesting, but that is not discussed.
- Line 667: “obliquity”. Not true. Both the 1218 and 1264 (both incorporated in Westerhold et al, 2020) are dominated by eccentricity in their spectral power. Low latitude forcing appears to be dominating global climate (in agreement with theory I would argue). How, power from precession is transferred to eccentricity is a topic of discussion, not included in this review.
- Line 668 -669: “Additionally, stratigraphic constraints on these records are insufficient to assess if this variability is consistent between sites.” This is not true either. In the Westerhold et al., 2020 files, Site 1218 and 1264 are correct on the ~110 kyr time scale of tuning. Testing of orbital variability between sites is thus possible, at least on eccentricity time scales.
- Line 669: are we still talking about benthic d18O?
Citation: https://doi.org/10.5194/egusphere-2023-2738-RC2 - AC2: 'Reply on RC2', Dominique Jenny, 16 Feb 2024
-
EC1: 'Comment on egusphere-2023-2738', Gerilyn (Lynn) Soreghan, 20 Jan 2024
This manuscript has potential, but both reviewers suggest significant revisions in order to bring it to an acceptable state. They provide substantial, constructive, and detailed suggestions. Please respond to each review comment; I expect to be encouraging a revised version assuming the authors address the reviewers' issues.
Citation: https://doi.org/10.5194/egusphere-2023-2738-EC1
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-2738', Anonymous Referee #1, 15 Jan 2024
Review of “Climate variability, heat distribution and polar amplification in the unipolar ‘doubthouse’ of the Oligocene”
General comments:
The purpose of the paper is confusing. It wasn’t clear from the start that it is a literature review and interpretation. The paper also tries to do too much by reviewing the chronostratigraphy, the geography, model-data comparison, fauna and flora, introducing new data using NRL, etc., etc. This could be mitigated by stating earlier (title, abstract, intro) and more clearly the purpose and structure of the paper and by restructuring sections of the paper (see below). The discussion/conclusion sections are a little disappointing. The discussion section doesn’t offer any insights into the mysteries or speculation about the remaining Oligocene mysteries (e.g. model-data mismatch, CO2-climate mismatch). The conclusion section was obviously a last-minute addition and needs to be redone. Critiques aside, I appreciated the thorough overview and learned some new things about the Oligocene. This could be a nice addition with some additional work.
Title:
The title is misleading. This is a review paper and the title should reflect that.
0 Abstract:
Page 2, Line 14-15 – I don’t understand this sentence. Wouldn’t the Oligocene be a good analogue for a future climate state with unipolar glaciation?
Page 2, Line 28 – Delete “while still maintaining a unipolar icehouse state.”
1 Introduction:Most of the paper is well written. The Introduction is the exception and requires careful editing.
Page 3, Line 4 – “equilibrate” is used incorrectly here.
Page 3, Line 43-44 - Awkward sentence. Please fix.
Page 3, Line 51-52 - Awkward phrasing. Please fix.
Page 4, Line 80 - I’m not sure why whales are mentioned here.
Page 5, Figure 1 caption - Maybe the colors didn’t transfer correctly, but some of the colors and shapes mentioned in the caption don’t match the colors in the legend of the figure (for (b)). Misspelling of “Transition” as well.
Page 6, Line 94 – The authors mention variability in continental ice volume paced by eccentricity and obliquity. The authors should follow that with an explanation of why/if that’s significant and how it might be seen in the proxy records (given that they are low resolution).
Page 6, Line 99 – “Here..” Should be start of a new paragraph.
Page 6, Line 100 – “after a chronostratigraphic section” Awkward phrasing.
Page 6, Lines 99-104 - At the end of the introduction, the authors finally explain that the purpose of this paper is to review the current state of knowledge regarding the Oligocene climate and thus provide constraints on boundary conditions, compile proxy data, and identify points of interest. This point should be made much sooner so that readers understand the purpose of the paper.
Page 6, Lines 99-104 - There is no mention of numerical climate simulations as in the abstract. Model-data comparisons are made later in the paper, and should be mentioned here. There is also no mention of constraining ocean circulation or ice dynamics though a large part of the introduction discusses these. The authors need to make clearer how climate simulations are used in this paper with some quick details on them, if anything on the ocean or ice state of the Oligocene is explored further, and potentially more details on the proxy records used and for what reason? There is not much on the purpose of the paper and it’s all at the very end of the introduction - they could expand on what the paper will discuss and could mention the main purpose it serves sooner in the introduction.2 Oligocene chronostratigraphy and event nomenclature:
This section should be removed completely from the paper or moved to Supplemental Data.
Page 7, Line 112 – “GGSP” should be “GSSP”, right?
Page 7, Line 116 - The OMB is stated as 23.04 Ma here and 23.03 Ma in the rest of the sections
Page 8, Line 134 – Spell out ‘GPTS’ as ‘geomagnetic polarity time scale (GPTS)’
Page 9, Line 177 - Why does the last zone end at 23.07 Ma when the Oligocene ends at 23.03 or 23.04 Ma?
Page 10, Line 196, 198, 201 - Why is there a question mark after ‘Axoprunum’?3 Boundary conditions for Oligocene climate:
Page 11, Line 236 – “boundary conditions” is a modeling term and inappropriate here. Replace with “geography” or “continental position”
Page 12, Figure 3 caption – Add an elevation scale bar. Also ‘NRL’ is supposed to be ‘NLR’.
Page 13, Line 266 – “strongly affected regional ocean circulation…” Here the authors are repeating a hypothesis that is by no means certain and should equivocate appropriately “may have strongly affected”. They do this throughout the paper, stating hypotheses, sometimes model results, as fact or certainty. Please clarify appropriately.
Page 15, Line 348 – The authors mention an increase in ocean methane hydrates, peat, and wetlands as a result of colder temperatures and ice expansion. They could add a sentence that explains how the expansion of those environments are a consequence of colder temperatures and if there is evidence of these expanding during the Oligocene or if it is just a probable guess given the cooling climate.
Page 15, Line 358-380 – How does these CO2 levels/trends compare with the new CO2 from Honisch et al., Science?
Page 16, Line 377 - “16esults” typo4 Climate proxy data:
Page 16, Line 385 - Why did the authors choose fossil plant remains to analyze and why only fossil plant remains (not paleosol carbonates, etc)?
Page 17, Line 405-407 – The authors mention an absence of high latitude data (no fossil plant remains suitable for their method). How could this absence affect their results, or how much uncertainty does this add when estimating a meridional temperature gradient for the Oligocene?
Page 17, Line 415 - One data point in the NH mid-latitudes is significantly different from the other mid-latitude data points - where is this NH data point? What sort of biome is it from?
Page 18, Figure 4 – Use the same y-axis for each plot. As it lies now, the winter temperatures appear the same as the MATs until you see the slight difference in the y-axes. Also, in the figure caption, it states the (b) plot PI values are in brown but they appear gray just like in plot (a).
Page 19, Line 460 - Random parenthesis in statement. Also, the authors could add one more sentence on how pH affects the fractionation of oxygen isotopes in this environment (why an overestimation of 1.5C has to be taken into consideration?).
Page 20, Figure 5 - This is a nice figure, though the shapes are so small I can’t really place many of the sites to a line. They could instead do just three different shapes for each type of SST proxy. There are also some SST sites in Figure 3 that aren’t in Figure 5 (i.e. Site 913 and others), if they aren’t included in Figure 5 then are they still used / needed in Figure 3? Why weren’t some used in Figure 5?
Page 21, Figure 6 - This is a great figure, very concise with a lot of information. It’s interesting too that SST seems ~constant from about 40S to 40N for the Oligocene and sort of late Eocene as well, but modern SSTs have a narrower curve peaking from about 20S to 20N. They could mention something on this.
Page 22, Figure 7 - Why reconstruct driest month precipitation and not wettest month precipitation? Along those lines, why reconstruct winter temperatures and not summer temperatures (Figure 4)? Could just include one sentence stating the benefits of understanding winter temperatures and driest month precipitation.
Page 22, Line 496 – The authors generally state that the Oligocene MAP values are similar to modern day, but the mid-latitude values (for either hemisphere) seem higher than modern values (Figure 7)?
Pages 23 and 24 – The authors don’t mention similarities or differences between the models here at all (HadCM3L vs CESM). According to Figure 8, it seems there are a lot of similarities between how the models reconstruct surface temperature compared to the proxy records. It’s interesting that you see such similarities between CESM and HadCM3L (though on that - parts C/F/I of Figure 8 could use a darker color for CESM mean temperature difference since the light pink is a bit difficult to see).
Page 24, Line 533 – The authors say the differences plotted are ‘calculated as a pointwise difference between the proxy mean value and the model annual mean’. Two thoughts on this: a) how was the model annual mean calculated (was the annual mean for the corresponding grid cell used or the general region / a group of grid cells) and b) are there any seasonal biases within the proxy records that could be causing greater differences since only annual means are used for the simulations? If not, they could just state somewhere that there are no known seasonal biases for these records.
Pages 25 and 26 - Similar to what I said above, similarities and differences between CESM and HadCM3L could be mentioned, I think the light pink color for CESM in Figure 9 parts C/F/I is too difficult to see, and I’m a bit confused on the ‘pointwise difference’ method regarding which grid cells were used for the model values. I also think red should correspond with drier (according to data) conditions (negative precipitation difference) and blue should correspond with wetter (according to data) conditions (positive precipitation difference). It’s a bit confusing that we are looking at precipitation as a ‘data minus model’ value where the data usually represents wetter conditions than the model simulates and yet the data points on the plot are in red colors which traditionally represent drier conditions but this is really just a preference.
Page 27, Line 583 – See inappropriate capitalizations. “Is of great Importance”
Page 27, Section 4.4 – This section on Ice Sheets is out of place. Unlike the rest of Section 4, it has no original data or interpretation. It’s simply a literature review and should be included in Section 3.5 Flora and faunal changes:
Similar to last comment, why is this section not with the literature review in section 3?
6 Discussion:
Page 31 – The authors mention there’s a discrepancy between SH and NH MAP (in Oligocene and not in modern) in the result section (Figure 7). A discussion of why would be welcome here.
Page 31, line 686-687 – The authors conclude that the Oligocene was anomalously warm relative to CO2 levels. They should discuss potential explanations for this apparent mismatch.
7 Conclusions:
It reads as though the authors ran out of energy before the Conclusions. There’s no introductory statement and the bulleted list is very sparse. Besides the abstract, the conclusions are the most frequently read section of a paper. The authors don’t state that the manuscript is a review or distinguish between their contributions and what’s review.
Page 31, Line 704 - The second conclusion is specific to the proxy records, not so much what the simulations show, so maybe mention that, though I guess they have a conclusion point specific to the simulations below.
Page 31, Line 706 - I generally agree with the third conclusion, except Oligocene MAP values around the mid-latitudes appear higher than modern MAP mid-latitude values according to Figure 7.
Page 31, Line 711 - In line with the last conclusion point, the authors could mention that that may be related to the lack/variability/uncertainty in Oligocene CO2 records. CO2 is not well constrained so perhaps the Oligocene is, in part, not the “icehouse” it seemed to be before because CO2 was, in reality, a bit higher than some older records estimated.Citation: https://doi.org/10.5194/egusphere-2023-2738-RC1 - AC1: 'Reply on RC1', Dominique Jenny, 16 Feb 2024
-
RC2: 'Comment on egusphere-2023-2738', Anonymous Referee #2, 17 Jan 2024
Jenny et al present a review of the Oligocene, an 11-million-year long time interval. The review is concerned with:
(1) a general introduction,
(2) Chronostratigraphy and event nomenclature,
(2.1) Planktonic foraminifera zones,
(2.2) Calcareous nannofossil zones,
(2.3) Radiolarian biostratigraphic zones,
(2.4) Dinoflagellate cyst zones,
(3) Boundary conditions for Oligocene climate,
(3.1) Geographical boundary Conditions,
(3.2) Ocean circulation,
(3.4) Carbon cycle,
(4) Climate Proxy data,
(4.1) Temperature,
(4.1.1) Continental Mean Annual Temperature,
(4.1.2) Sea Surface Temperature,
(4.2) Precipitation,
(4.3) Data-model comparison,
(4.3.1) Temperature proxy to model comparison,
(4.3.2) Precipitation proxy to model comparison,
(4.4) Ice sheets,
(5) Flora and Faunal changes,
(6) Discussion,
(Note, 6.1 is missing).
(6.2) Temperature trends and variability,
(6.3) Precipitation, and
(7) Conclusions.
This review constitutes a very ambitious, admirable, and important project, that necessarily needs to make broad-brush summaries and interpretations about the main climatic and oceanographic factors that determined the Oligocene palaeoclimate system. In principle, I am sympathetic to this effort. The paper is well written.
However, I believe that for this paper to become an authoritative reference paper much more work is needed. I am no expert on all topics reviewed here. But on topics on which I do consider myself reasonably knowledgeable I find that they are often marked by too many inaccuracies, some outdated discussion points, and several key omissions. By proxy, I have therefore little faith that the parts of the review concerned with topics on which I am no expert, accurately capture the current state of knowledge. I am left with the impression that this paper ‘could have been so much better’.
To me the paper suffers from not having included enough Oligocene workers, many of whom have dedicated their professional lives to understanding topics ranging from age models, marine microfossils (calcareous nannofossils, foraminifera, radiolaria, etc.), macrofossils (animals and plants), ice sheet modelling, palaeogeography, etc. Compare, for example, this Oligocene review to Steinthorsdottir et al., 2021, who consulted 3 to 4 times as many experts for their Miocene review. (Here my criticism is mainly directed at coauthors Bijl, Huber, and Sluijs; all experienced and senior workers, some of whom have been involved in the Steinthorsdottir effort. To me, and as far as I can judge, it seems that with greater input from these authors, also in study design, the manuscript could have been lifted to a higher level already. Jenny’s efforts in leading this study, as a PhD student, is admirable to say the least.)
I believe that revising this manuscript will result in a much better paper. To me the authors have two options: Option 1) shorten the manuscript considerably by removing much of the review text that is not pertinent to the temperature/precipitation compilation and model comparison, i.e. what I consider the main strength of this paper, and expertise of the authors, or Option 2) include more expert knowledge, work with the text and develop the discussion to lift this paper to a higher level. I recommend the second option, and I believe it could make, after substantial revisions, for an authoritative Oligocene reference paper. Ultimately, it is the author’s choice.
I will provide some below and wish the authors much success with revising this document and planning the next steps ahead. I am sorry that I am not more positive in this instance.
Major comments
- Structure of the paper: I believe that inclusion of subheadings concerned with magnetostratigraphy, astrochronology, oxygen and carbon isotope stratigraphy would strengthen the paper. Perhaps compare to Steinthorsdottir et al., 2021, to see how they went about their Miocene review. If this comparison is unjustified, please say so in the rebuttal.
- Doubthouse/Unipolar Doubthouse: Most workers use “Doubthouse” to describe the latest Eocene (between the MECO and EOT), where the climate system was cooling and “doubting” to jump from a “Greenhouse” state to a “Unipolar Icehouse” state. I much prefer to reserve the phrase “Doubthouse” for the Late Eocene, and the usage of “Unipolar Icehouse” to describe the state of the Oligocene to Pliocene time interval. As an aside, I also prefer “Unipolar Icehouse” over the phrase “Coolhouse”. The usage of “Unipolar Doubthouse” in your title is a mixture of “Unipolar Icehouse” and “Doubthouse” and is confusing to me. I am not entirely sure what is meant by this phrase. Is the system “doubting” to be unipolarly glaciated? Or “doubting” to go back to a “Greenhouse” state, or to a “Bipolar Icehouse” state?
- Figure 1: This figure has many inaccuracies. Almost all boundaries or events (apart from those around the EOT, which seem to be OK) are indicted in the wrong place. Furthermore, the MOG is not a previously defined event, the MOGI is. The MOGI as drawn does not cover the MOGI as defined in the literature. The OMT does not cover the largest glaciation across the OMB. The Mi-1 points to an d18O low, not a d18O high, and moreover it is best to abandon Mi and Oi nomenclature completely (more about this below).
- Line 108/109, the Oligocene time scale: For an Oligocene review paper, it is important to get the age scale right, and to describe it explicitly (not just mentioning it in a figure caption). The absolute ages of the polarity time scale in the GTS2012 (i.e., the GPTS2012) are of greater quality than those incorporated in the GTS2020/GPTS2020, especially for the latest Oligocene where the GTS2020 incorporated erroneous ages. Admittedly, large efforts have been put toward improving biostratigraphic markers and zonations for the GTS2020, but also their absolute ages are unfortunately too often based on an inferior GPTS2020 scale (compared to the GPTS2012). I recommend using GPTS2012 to any Oligocene (age model) worker, and an Oligocene review paper needs to address the errors in the G(P)TS2020, if the authors decide to stick with this age scale as their preferred option.
- Line 353 to 380: This CO2 discussion needs more work. What are the implications for radiative forcing? For example, how to explain the Late Oligocene Warming whilst CO2 is decreasing? This is a major topic in Oligocene research, and it is not mentioned or discussed. For this manuscript to become an authoritative review, such a major climatic enigma needs to be addressed, because it will help with formulating new and important research questions.
- Discussion and Conclusions, both seem quite superficial to me. Can we take a step back, and distil major research questions for the Oligocene?
Minor comments
- Line 13: Perhaps start the abstract with why the Oligocene is of any interest to anybody? It may help draw some more readers in.
- Line 14: Not entirely sure why this would lead from the first line of the abstract. To me the gradients could be a good analogue. Ice sheet geometry, palaeogeography, etc, seem to me the limiting factors in the Oligocene being a palaeo analogy for long term future climate states.
- Line 23: proxy data instead of proxy based data?
- Line 25: “In line with previous…” Does this refer to Oligocene modelling efforts? Or modelling of warmer climates in the general. I thought that this was a feature of Eocene modelling output but was unaware of this being a problem in general, also for the Oligocene unipolar icehouse. Perhaps clarify?
- Line 23 to 27: should we learn a lesson from these observations? Is it worth making that lesson explicit in the abstract (e.g., much more effort needs to be invested in improving Oligocene climate simulations?)
- 1: Position Mi-1 is wrong. In general, it is best to refrain from using Oi and Mi numbering. Oi and Mi numbers were originally defined (in very low-resolution records from the 1980s and early 1990s) as oxygen isotope zones, like biozones, that lasted several hundreds of thousands to millions of years (see original Miller papers and Wright papers). Over the years, these zone numbers were used to refer to the oxygen isotope “highs” (often called “events”) at the base of the original oxygen isotope zones, yet this is not how they were originally defined. The arrows in Fig 1, suggest that the authors also use the Oi and Mi numbers to indicate events. If this usage is adopted, which I do not recommend, then please refer to high oxygen isotope values. E.g., the position of the Mi-1 is not correct in this figure (you refer to an oxygen isotope low). Furthermore, even when the Mi-1 (short for “Miocene oxygen isotope zone 1”) is indicated correctly in Fig. 1, it would still fall in the latest Oligocene (yet another reason to abandon Oi and Mi numbers). Lastly, nobody has even been able to explain to me what makes the Oi1b, Oi2a, and Oi2b stand out as particularly interesting, or exceptional “events”, probably because these numbers were devised as zones originally.
- 1: MOGI instead of MOG? Also, this band is not drawn according to the original definition. You may choose to redefine the main intervals of interest, but then they need to be defined in the text.
- Line 90: Oligoene (missing c)
- Line 90: MOG, where G stands for “glacial”. In MOGI, this stands for Glacial Interval, and refers to a longer period of mean elevated d18O values (punctuated by shorter lasting d18O decreases of about 1 per mil amplitude, interpreted to be Antarctic “interglacials” and/or deep-water warming events). The use of glacial in MOG is a bit strange and makes me think of something more similar to the last Pleistocene glacial cycle.
- Line 99: See Major Comment #1
- Line 109: I strongly recommend using G(P)TS2012 over G(P)TS2020. See Major Comment #3.
- Line 115: Beddow et al., 2018 did not work on the OM Boundary type section (Lemme Carrosio) but derived an astronomically calibrated age for the base of C6Cn.2n from Pacific sediments (perhaps make that clearer, because the current sentence can be interpreted as if Beddow et al did work on that section). Also, the 100 kyr error is not mentioned by Beddow et al. and seems too large.
- Line 110 to 125: This section would be much strengthened with the addition of a description of astrochronozones and potentially astro-unit-stratotypes. GSSPs are useful to an extent, but have their limitations (see e.g., Hilgen et al., 2020, Newsl. On stratigraphy). Potentially add a magnetostratigraphic description as well
- Lines 126 to 145: Overall a good summary, however, as stated above, I would shy away from reiterating Oi and Mi numbers. EOT, MOGI, and OMT are more useful terms in my opinion. Potentially in addition with the Late Oligocene Warming (LOW). MOGI definition used here, instead of MOG in figure caption. Please note that the MOGI is defined as “a generally cold but highly unstable mid-Oligocene time interval (∼0 My to 26.3 My ago), which we refer to as the Mid-Oligocene Glacial Interval (MOGI)”, not as “a ~1 Ma year phase of profound cooling/glacial expansion”. Usage of Ma is erroneous in this context too (Ma refers to an age not a duration). In general, this section needs some work/detailed checking, as I have not checked all statements, and I may have overlooked other potential errors.
- Section 2.1. The GTS2020 is abandoned here?
- Section 2.2. Similar comment. Agnini et al., 2014, cannot have used GTS2020. Please mention on which age scale these bioevents are given.
- Section 2.3. Same here.
- Section 2.4. and here. For Chapter 2, in general, best to refer to an accepted general age scale. I recommend G(P)TS2012. Where biostratigraphic workers have deviated from these scales, it is important to mention so, and to estimate how large uncertainties/discrepancies are between scales.
- Line 329 to 352: This discussion is only concerned with the EOT interval. What about the 400 kyr, and 2.4 Myr cycle that dominate the Oligocene benthic d13C records?
- Line 353 to 370. What about the drop in CO2 at around 24 Ma, during a 1.2 Myr Obliquity cycle “node”. May this drop have preconditioned the system for the (large amplitude of the) OMT.
- Line 377: 16esults. Typo?
- Figure 5: OMT, MOGI, EOGM: please check ages.
- Line 647 to 657: Very qualitative language. Warm, warmer, etc. Can we not put some numbers to how warm the Oligocene was?
- Line 658 to 672: this section is mixing things up. It starts with a description of high resolution benthic d18O records and compilations. Then merges into a discussion of low-resolution SST data, with a reference to Fig. 5 as the only give-away that the topic has changed (admittedly still temperature, but now surface instead of deep ocean). It then comes back to comparing low-res SST (without mentioning that that is what is being compared) to benthic d18O.
- Line 660: “most likely affected by a shift of geographical”. As far as I am aware there is no doubt about it. Not sure if it is still meaningful to dig out this splicing issue in Zachos 2001. We are two good compilations further (Zachos 2008 and Westerhold 2020). Also confusing this with a true signal in both the Atlantic and Pacific benthic d18O of the Late Oligocene Warming is unhelpful. Is it not smarter to have a section in the paper that discusses benthic d18O. To me, the striking absence of a LOW signal in SST is something more interesting, but that is not discussed.
- Line 667: “obliquity”. Not true. Both the 1218 and 1264 (both incorporated in Westerhold et al, 2020) are dominated by eccentricity in their spectral power. Low latitude forcing appears to be dominating global climate (in agreement with theory I would argue). How, power from precession is transferred to eccentricity is a topic of discussion, not included in this review.
- Line 668 -669: “Additionally, stratigraphic constraints on these records are insufficient to assess if this variability is consistent between sites.” This is not true either. In the Westerhold et al., 2020 files, Site 1218 and 1264 are correct on the ~110 kyr time scale of tuning. Testing of orbital variability between sites is thus possible, at least on eccentricity time scales.
- Line 669: are we still talking about benthic d18O?
Citation: https://doi.org/10.5194/egusphere-2023-2738-RC2 - AC2: 'Reply on RC2', Dominique Jenny, 16 Feb 2024
-
EC1: 'Comment on egusphere-2023-2738', Gerilyn (Lynn) Soreghan, 20 Jan 2024
This manuscript has potential, but both reviewers suggest significant revisions in order to bring it to an acceptable state. They provide substantial, constructive, and detailed suggestions. Please respond to each review comment; I expect to be encouraging a revised version assuming the authors address the reviewers' issues.
Citation: https://doi.org/10.5194/egusphere-2023-2738-EC1
Peer review completion
Journal article(s) based on this preprint
icehouseclimate. We extend an existing marine climate proxy data compilation and present a new compilation and analysis of terrestrial plant assemblages to assess long-term climate trends and variability. Our data–climate model comparison reinforces the notion that models underestimate polar amplification of Oligocene climates, and we identify potential future research directions.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 448 | 180 | 41 | 669 | 27 | 23 |
- HTML: 448
- PDF: 180
- XML: 41
- Total: 669
- BibTeX: 27
- EndNote: 23
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
Dominique K. L. L. Jenny
Tammo Reichgelt
Charlotte L. O'Brien
Xiaoqing Liu
Peter K. Bijl
Matthew Huber
Appy Sluijs
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5237 KB) - Metadata XML