the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Feb 2026
A compilation of sea surface temperature anomalies from the Southwest Atlantic during the Common Era: challenges and opportunities to support future research
Abstract. Here, we present a compilation of sea surface temperature (SST) anomalies from the Southwest Atlantic during the Common Era. We aim to highlight the challenges and opportunities in advancing our understanding of regional ocean variability and allow to understand the reasons behind the notable scarcity of such records and explore how we can improve mechanisms to prevent misinterpretation and ensure that these efforts effectively support future research. About 24 records indicated that SST anomalies fluctuated throughout the entire Common Era, with values between (-5.84 to 0.62). The limited availability of high-resolution SST records constrains our ability to fully grasp regional climate dynamics and their broader implications for the global climate system. To address this, it is essential to prioritize acquiring records that minimize coastal influences, thereby providing clearer insights into large-scale oceanic and climatic patterns. Enhancing the representation of currently under-sampled regions is crucial for constructing a more comprehensive picture of Earth’s climate history. However, addressing these gaps involves more than data collection alone. It requires a concerted effort to produce and disseminate SST reconstructions spanning the Common Era, while also raising awareness of their value to the scientific community. Ultimately, such initiatives will enhance our ability to anticipate and respond to future climate change, equipping policymakers and communities with the knowledge necessary to build resilience and adapt to an increasingly dynamic climate system.
- Preprint
(1115 KB) - Metadata XML
-
Supplement
(314 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on essd-2026-122', Anonymous Referee #1, 26 Mar 2026
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2026-122/essd-2026-122-RC1-supplement.pdfCitation: https://doi.org/
10.5194/essd-2026-122-RC1 -
AC1: 'Reply on RC1', Renata Hanae Nagai, 29 May 2026
RC1: Major Comments
Overall comment. I understand that the authors aim to raise awareness of the importance and quality of Common Era data in the Southwest Atlantic, and I recognize this as a major gap in paleoclimate studies. However, stronger arguments and a more in-depth discussion are needed to convince the reader—particularly those less familiar with Southern Hemisphere datasets. This is especially relevant for the Introduction, which is currently too brief. Moreover, given that this journal focuses on datasets, a detailed description of the compiled data should be central to the manuscript. The authors should clarify why this compilation is suitable for publication in this journal. While the global importance of this region is evident from PAGES 2k records, the manuscript should better articulate the regional significance—specifically for South Atlantic and South American studies. In my view, several aspects of the manuscript require further explanation and clarification to ensure a proper understanding of both the dataset and its contribution to the identified scientific gap.
Author’s response: We thank the reviewer for this thoughtful and constructive comment. We agree that the previous version of the manuscript did not sufficiently emphasize the dataset's broader scientific relevance or provide sufficient detail on the compilation's structure and significance. In response, we have substantially revised the manuscript to address these points. Specifically, we expanded the Introduction section and now provide a more comprehensive view of current gaps in Common Era paleoclimate records from the SW Atlantic, as well as the broader relevance of this region. We also strengthened the discussion of the compilation's scientific importance and its contribution to ongoing regional and global paleoclimate synthesis efforts. In addition, we improved the description of the compiled dataset throughout the manuscript, including clearer explanations of the data sources, temporal coverage, proxy characteristics, regional distribution, and reconstruction approaches.
The Abstract presents the main objective and briefly describes the data used; however, it does not clearly describe the methods applied, the main findings, or how the analysis leads to the stated conclusions. Temperature values are presented without units, and a definition of the “Common Era” is missing. In addition, the sentence beginning with “We aim to…” is overly long, making it difficult to identify the study’s main objectives. I recommend rewriting this sentence. More generally, it is not clear whether the stated objectives are achieved. The Abstract reads as fragmented and lacks clear connections between sentences, and I therefore suggest a more thorough revision.
Author’s response: We agree that the previous version did not clearly describe the methodological approach, the principal findings, or how the analyses supported the conclusions presented in the manuscript. In response, we substantially revised the Abstract. Specifically, we added a clearer description of the methods used to calculate SST anomalies and construct the regional composite, including the definition of the reference period and comparison with instrumental datasets. We also defined the Common Era (CE) upon first mention and corrected the omission of temperature units throughout the text. In addition, we rewrote the sentence beginning with “We aim to…” to make the objectives more concise and easier to follow. The revised Abstract now more clearly presents the study's main findings, including the range of SST anomalies observed during the Common Era, the regional cooling trend identified after ~1850 CE, and the interpretation of the role of regional oceanographic processes. We also revised the overall organization and transitions between sentences to improve readability and better demonstrate how the stated objectives are achieved.
The authors successfully place the study within a global context by referring to PAGES 2k results (Figure 1), clearly highlighting the scarcity of data in the Southern Hemisphere and within the study region. However, the Introduction is very brief and lacks a more detailed discussion of previous regional studies and the broader implications of these data gaps for paleoclimate research. This section would benefit from further development, including discussion of: (i) the underlying reasons for the data gap; (ii) whether this limitation is specific to the Common Era or also affects other time periods; (iii) recent regional studies (as most cited references are global and pre 2020).
Author’s response: We thank the reviewer for this important comment and for recognizing the relevance of placing the study within the broader PAGES 2k framework. We agree that the original Introduction was too brief and did not sufficiently discuss the regional context, previous studies, and the broader implications of the identified data gaps for paleoclimate research in the Southwest Atlantic and Southern Hemisphere. In response, we substantially expanded and revised the Introduction to address all of these points. Specifically, we added a more detailed description of the underlying reasons for the scarcity of Common Era SST records in the SW Atlantic. We also clarified that these spatial and temporal data limitations are mainly restricted to the Common Era. We believe that this has strengthened the manuscript's regional perspective and more clearly demonstrates the scientific importance of compiling and synthesizing Common Era SST records from this underrepresented region.
Data Section Supplementary Table 1 provides detailed information on the records; however, key general information should also be included in the main text. For example, the authors should report the age range of the records (oldest and most recent) and clearly describe the criteria used for record selection, especially since some records do not meet PAGES 2k criteria. This could be achieved by moving Figure 5 to appear earlier in the manuscript (perhaps as Fig. 1?), as it provides essential information on the records, such as age range and proxy types. The information from this Figure should also appear summarized in the main text. When stating that the records are “influenced by diverse oceanographic conditions” (L. 53), the authors should specify what these conditions are. Additionally, the methodology used to compute anomalies is not sufficiently clear (e.g., whether they were averaged over the entire domain), and the construction of Figure 2 requires further explanation. I recommend providing more methodological details in the main text.
Author’s’ response: We thank the reviewer for these detailed and constructive suggestions. We agree that several important aspects of the dataset description and methodology required further clarification in the main text. In response, we substantially revised the manuscript to address all of these points. Therefore, we added additional information summarizing the temporal coverage of the compiled records, including the oldest and most recent intervals represented in the dataset. We also clarified the criteria used for record selection, emphasizing that all available SST records from the Southwest Atlantic spanning the Common Era were included in the compilation. Although the records do not fully satisfy the PAGES 2k criteria, we now explicitly discuss this limitation in the manuscript and clarify that these records nevertheless provide valuable insights into regional climate variability and its relationship to broader-scale climate patterns. Following the reviewer’s suggestion, we moved the former Figure 5 to an earlier position in the manuscript (now Figure 2), as it provides essential contextual information regarding proxy types, temporal coverage, and record distribution. We also expanded the figure caption and incorporated a summary of the key information from this figure directly into the main text to improve readability and accessibility. In addition, we clarified the statement about the influence of “diverse oceanographic conditions” by explicitly identifying the main regional oceanographic features affecting the study area, including the Brazil Current, the La Plata River Plume, and coastal upwelling systems, which are now discussed in greater detail throughout the manuscript. We also substantially revised the methodological description of the SST anomaly calculations and composite construction. The revised text now clearly explains that anomalies were calculated individually for each record relative to the adopted reference period, and that no spatial averaging was applied during this stage. We further clarified how the individual records were combined to form the composite SST anomaly series and improved the explanation of Figure 3 (formerly Figure 2) and related analyses. In addition, we refined terminology throughout the manuscript, including replacing “stack curve” with “composite SST anomaly series,” and clarified the comparison with previous SST reconstructions.
RC1: Minor Comments
- Please define “Common Era (CE)” in terms of ages (i.e., CE begins in year 1 AD and extends to the present day), as this term, although well established, is not defined in the manuscript. The expression “the last two thousand years” (Line 24) is too vague. In addition, in Line 80 the Roman Warm Period is described as starting in year 0 CE, but year zero does not exist. I recommend revising these aspects accordingly
Author’s response: We have revised the manuscript to explicitly define the Common Era (CE) as the period from 1 CE to the present day, both in the Abstract and in the main text. We have also replaced “the last two thousand years” with the explicit temporal range (1–2018 CE) of the dataset used in this study. We corrected the definition of the Roman Warm Period (~1-400 CE), ensuring consistency with standard chronological conventions, and aligned its description (~1–400 CE) according to Mann et al. (2009) and Neukom et al. (2019).
- In Lines 13–14, the sentence presenting the SST anomaly values is not well integrated with the surrounding text. I suggest either removing it or rephrasing it to better place it in context.
Author’s response: We have rewritten the abstract to improve readers’ understanding.
- Figure 1 clearly shows the study domain; however, in panel (b) the colors and marker sizes make it difficult to distinguish the record locations. I suggest increasing the marker size and improving the color contrast
Author’s response: We have revised Figure 1b by increasing the marker size to enhance the visibility and distinction of record locations.
- It is not clear why subsection 2.1 is included if there are no additional subsections. I suggest revising the section structure.
Author’s response: This subtitle was removed.
- Please revise the sentence in Lines 51-52 for clarity
Author’s response: The sentence has been revised to improve clarity and readability. In addition, abbreviations have been standardized to ensure that each term is defined before its first use.
- Lines 65–67: The terms “pre-industrial cold and warm periods” are not clearly defined. Please clarify their meaning and revise the sentence accordingly
Author’s response: This has been clarified in the revised manuscript. The terminology has been replaced by explicitly defined Common Era climatic intervals following Mann et al. (2009) and Neukom et al. (2019). The sentence has been revised to improve clarity (Lines).
- Figure 2 caption: The definition of the time intervals should also be included in the main text.
Author’s response: This has been addressed by including the definition of the time intervals in the main text (Lines). The figure number has also been updated (now Figure 3) following the reorganization of figures in the manuscript.
- Figure 3. It is not clear how the data used in this figure were obtained. In the main text, the authors refer to “tropical global SST anomalies presented by Abram et al. (2016)” (Line 68), whereas the figure refers to global, Southwest Atlantic, and Southern Hemisphere SST anomalies. Please clarify this information in both the figure and the text.
Author’s response: We thank the reviewer for pointing out this lack of clarity regarding the data sources used in Figure 4 (former Figure 3). We agree that the original description in the main text and figure caption was ambiguous and could lead to confusion regarding the origin and nature of the different SST anomaly series shown. We have revised both the main text and the figure caption to clearly specify the datasets used. We now clarify that the global and Southern Hemisphere SST anomaly series shown in the figure are derived from Kennedy et al. (2019), while the Southwest Atlantic SST composite presented in this study is newly generated from the compiled proxy records. We also explicitly distinguish between these external datasets and our regional reconstruction to avoid any ambiguity. In addition, the figure caption has been expanded to clearly describe each time series and its origin, and the figure has been renumbered (now Figure 4) to ensure consistency throughout the manuscript.
Citation: https://doi.org/10.5194/essd-2026-122-AC1
-
AC1: 'Reply on RC1', Renata Hanae Nagai, 29 May 2026
-
RC2: 'Comment on essd-2026-122', Anonymous Referee #2, 19 Apr 2026
Review of dos Reis de França et al., A compilation of sea surface temperature anomalies from the Southwest Atlantic during the Common Era: challenges and opportunities to support future research, submitted to ESSD.
Summary of review. With apologies for the lateness of this review: I have substantive requests for revision in all sections of this manuscript.
Specific comments.0. Abstract
l. 11-13. Goal is to "to highlight the challenges and opportunities in advancing our understanding of regional ocean variability and allow to understand the reasons behind the notable scarcity of such records and explore how we can improve mechanisms to prevent misinterpretation and ensure that these efforts effectively support future research".
--> Why is this region of interest and importance?
l. 14-15. Results: "About 24 records indicated that SST anomalies fluctuated throughout the entire Common Era, with values between (-5.84 to 0.62)."
--> Further analysis needs to be performed (see comments below), and then this statement of what has been learned from the dataset can be made more specific.
1. Introduction
l. 38-43: Make a stronger case for the need for southwest Atlantic records.
(1) For available paleoclimate observations, the authors can point out that this scarcity is not resolved in updated curated Common Era paleoclimatic databases (e.g. Evans et al 2026, ESSD, and references to component databases therein).
(2) Most importantly, the authors need to make the case here: what we might learn from southwestern Atlantic paleoceanographic records, perhaps drawing from prior paleoclimatic studies, and from modern physical, chemical and biological oceanographic studies. What are the specific outstanding questions concerning regional ocean and climate change? What is the specific role of the southwestern Atlantic in (l.42) "driving and modulating global ocean and atmospheric circulation patterns"?
(3) What is meant by (l. 43) "...ensuring optimal efficiency" in efforts to address the scarcity of observations in this region?
l. 46-50: Missing from the introduction, but needed: an introduction to the ways in which SST was reconstructed from the various proxy observations made in the gathered sediment cores: MAT, UK'37, Mg/Ca; the uncertainties in the reconstruction of SST from these proxies and in this particular region, especially considering the relatively shallow water depths and regional oceanography from which the sediment cores were retrieved and the potential for the dataset to reflect terrestrial influences such as runoff, erosion, and anthropogenic accelerations of effects; and atmospheric influences on the coastal ocean setting. This will be useful for assessing and expanding results and discussion from analysis of the dataset. Some of the regional oceanographic background is at l. 102-110.
2. Data
l. 49-59: Provide original data source URLs in the main text. Describe the specific criteria, and give the rationale for inclusion of data in the database presented here: proxy observation, region, time coverage, time resolution, chronological control, other ... ? These would be important for replicating the work and for adding to the database in the future.
l. 60-71: exploratory visualization of SST anomalies. In this description of how the data were studied, I think the use of broad time windows and regional integration should be justified from a scientific standpoint. Then it should also be justified by the chronological control (see notes on revisions to Supplementary Material). Because the averaging is over time intervals between 150 and 550 years long, it might be better to use a historical climatology averaged over a larger spatial area and a longer time period as reference period. Instead of "summer" (l. 68), define the months within the year over which the average is performed. The rationale for a seasonal average should be described and sources cited, in the Introduction (see notes on l. 46-50).
l. 74: Fig 2:
(1) Please revise to indicate the differential lengths of the time periods studied: make boxplots the width of the named time periods, and give time in years CE on the horizontal axis in addition to the names. This will help readers unfamiliar with these names to understand the results. It also may provide a discussion point: since most of these names, if not all, originated with studies in the North Atlantic, to what extent does the exploratory visualization assume an Atlantic-wide common temporal expression of SST anomalies? This is not certain to be the case for nearshore/shallow water depth records from the southwest Atlantic (really - Brazil-Argentinian Margin might be a better description of the region sampled).
(2) Fig 2: Because the boxplots indicate nonnormal distributions within time intervals, I suggest you plot medians for each record within each period and boxplot.
l. 67-68: Explain the rationale for comparing the results to tropical global SST anomalies from Abram et al (2016). Background for this rationale should be in the Introduction (see notes on l. 38-43).
l. 69-71: These two sentences seem to be elements of Discussion; please move them there.
3. Results:
l. 73: I suggest creating a section 3, Results, which contains Figs 2-4 and the text describing the results of the analysis of the database.
l. 80-96 and Fig 2: Because the SST reconstructions are from shallow water depths, and from Fig 2 results, I suggest making more robust inference on boxplot medians rather than ranges. This might be also justified by introduction to the uncertainties in creating the SST reconstructions, and estimates of the SST reconstruction uncertainties (l. 46-50).
l. 115 and Fig 4: I think this figure logically comes next, as it addresses the question of whether the Brazil-Argentinian Margin dataset introduced here shows similar Common Era variability as does the Ocean2k compilation of global SST. I think that data should be presented similarly to the boxplots as shown in Fig 2, for reasons that are detailed in the McGregor et al (2015) paper which is cited and sourced for the global mean reconstructed SST comparison. Specifically, all the gathered data for that 2015 paper were on different timescales, resolutions, and nominal chronologies, so were perhaps best considered as binned composites, similar to what is done in Fig 1. This also has the benefit of allowing the authors to make inference on medians and changes in medians between one time period and the next. It looks like the same resolvable features are in both composites: relatively flat in the first millennium, with a cooling in the 1300-1800 period.l. 121-122: "These findings demonstrate that the SW Atlantic lacks adequate records to represent Common Era SST anomaly changes...": This discussion argument seems to be contradicted by the results in Fig 3.
l. 121-122: I understand that the authors argue that it is a change in the relative representation of warm vs cold current influenced regions in the regional SST reconstruction that creates the apparent agreement noted in the prior comment. This needs more support from data analysis. For instance, looking at Fig 1 for core locations, and then Fig 2 and Fig 5, it seems like:
(1) l. 110-112: The authors write: "Notably, the sedimentary
records north of 25°S span 778–2010 CE and are influenced by the Cabo Frio coastal upwelling system, which may amplify the regional SST cooling signal from the LIA to the present (Fig. 5)." The authors seem to be referring to record NAP62-2; other northern records are entirely within the LIA. Yet it seems from Fig 1 that at least 3 other records span this period and are within the current (no pun intended) influence of the warm Brazil Current (Fig 1) during the LIA.(2) Within the Little Ice Age period (Fig 2) there are 10 records available spanning a wide range of color codings. If those color codings are as indicated in Fig 5, then a wide range of sites along the coast, and in the influence of the warm Brazil Current, are sampled.
(3) A comparison between record LaPAS-KF02 spanning nearly the entire time period, and NAP62-2, might indicate whether this is the case. Of course, if the data are composited as anomalies relative to modern means, then the effects of differences in mean SST near the upwelling region and across the transect will be removed from the composite already.
(4) I am unable to further assess the authors' arguments by, for instance, making a plot of mean reconstructed SST vs site latitude, or a composite of selected subsets of the data by latitude, or comparing across subsets by proxy observation, because the authors have only provided a single value for each time point in the compilation in their Figshare repository (l. 150-152), rather than the SST reconstructions by core and providing their associated metadata. Please provide the complete database and all metadata in the repository, such that further exploration of the dataset is possible.
(5) If the authors wish to argue that there has been a shift in SST along part of this transect, which might be related to the global or Atlantic basin SST reconstructions over the Common Era, as composited by McGregor et al (2015): this might be speculated using physical oceanographic arguments and changes in winds in the historical era, but evidence has not been presented to support that idea.
Unfortunately, given these comments (1-4), I find that the authors have not provided convincing support for their argument at l. 112-14, and their conclusion at l. 120-126.
l. 197 and Fig 3: I think this figure lacks rationale and I would differ in my interpretation of the results.
(1) l. 194-196: "Overall, unlike the global trend of rising SSTs observed in several studies, especially highlighting the marked post-industrial warming (e.g., Neukom et al. 2019), the SW Atlantic shows an opposite pattern with a cooling trend from 1850 to 2020 (Fig. 2)." With only a single boxplot in Fig 2 for post-1850 data, this cannot be judged. The results in Fig 3 (blue line) seem to indicate that all values are cold relative to the 1961-1990 mean, and perhaps this is not surprising, as the 1961-1990 period postdates the vast majority of the data in the series. But if anomalies were calculated relative to the 1961-1990 mean, and these remain 2-3 degrees colder than that reference period through the temporal overlap period (Fig 3) - something seems wrong. please check if there might be a mistake in the figure?
(2) Fig 3, time series: how was this constructed? Are there uncertainty estimates on the x and y axes? See also comments on introduction to the reconstructions and apply that information to the revision of this figure. The boxplot composite formulation used in Figs 2 and 4 might usefully enable representation of age and observational uncertainty here.
(3) It's unclear to me why a Southern Hemisphere and global average SST record would be comparable to the SW Atlantic anomaly series, but the legend gives me too little information to understand what is here. Are the SH and GL averages from historical gridded estimates? I don't find Kennedy et al (2019) in the reference list. I don't know how the averages were made. Earlier there was reference to "summer" values; is summer also used for creating these averages?
(4) The scale of variation for the regional composite is a factor of 4x greater. If the imprint of forced climate change locally is roughly the same as on global SST (regress the global SST field on the global SST time series to find out), then is it possible that the unforced variability, for coastal locations in this region, especially for shallow water depths, is more atmospherically and therefore stochastically controlled? In which case you might expect to see what is in the blue line: no change over the historical period and a high level of interannual variability which is overlain by variability associated with changing sampling resolution for the reconstructions.
4. Summary and conclusions
l. 128-137: I can agree with the first 2 sentences. The third is not supported by evidence presently provided in the manuscript, and might be contradicted by evidence that is presented. The fourth could be true but is not supported by data analysis, and is not currently accessible by data and metadata provided with the manuscript, but could be.
l. 138-148: A discussion of the extent to which high resolution offshore sediment records could be obtained in the region would help support conclusions here. The authors could cite recent publications of datasets in the NOAA/NCEI repository (https://www.ncei.noaa.gov/access/paleo-search/), and also publications in ESSD, for instance:
https://doi.org/10.5194/essd-12-2261-2020
https://doi.org/10.5194/essd-12-2579-2020
https://doi.org/10.5194/essd-15-2081-2023
https://doi.org/10.5194/essd-16-1933-2024
https://doi.org/10.5194/essd-2025-812/
https://doi.org/10.5194/essd-18-1185-2026
https://doi.org/10.5194/essd-18-1921-20265. SM:
Table 1: It would be helpful to have columns for 'bottom date' and 'top date' to allow the reader to see which age models might be best constrained by chronological control, which ones include chronological extrapolation (if any); additional information as mentioned previously would be valuable for dataset users.
Supplementary Figure 1: these data are overwhelming from very shallow water depths <150m. Therefore very close to present shoreline. Is there discussion of what this might mean for any interpretation of the results? Yes, there is! But see prior comments for further analysis and interpretation.
6. References:
Please include links to original publicly accessible datasets, if available, and if not, provide the full dataset and all its metadata in the Figshare repository.
Citation: https://doi.org/10.5194/essd-2026-122-RC2 -
AC2: 'Reply on RC2', Renata Hanae Nagai, 29 May 2026
RC2: Major comments
Review of dos Reis de França et al., A compilation of sea surface temperature anomalies from the Southwest Atlantic during the Common Era: challenges and opportunities to support future research, submitted to ESSD. Summary of review. With apologies for the lateness of this review: I have substantive requests for revision in all sections of this manuscript.
- Abstract
- 11-13. Goal is to "to highlight the challenges and opportunities in advancing our understanding of regional ocean variability and allow to understand the reasons behind the notable scarcity of such records and explore how we can improve mechanisms to prevent misinterpretation and ensure that these efforts effectively support future research".
- --> Why is this region of interest and importance?
Author's response: We thank the reviewer for this important observation. We agree that the original text did not sufficiently explain why the SW Atlantic is a region of scientific interest and importance. In response, we revised the Introduction and Abstract to better emphasize the region's climatic relevance. We now clarify that the role of this basin is to interhemispheric heat transport and ocean-atmosphere interactions, and is strongly influenced by major oceanographic systems such as the Brazil Current, the La Plata River Plume, and coastal upwelling processes. These features directly affect regional climate variability and influence broader South Atlantic and South American climate dynamics. We also emphasize that the region remains substantially underrepresented in global paleoclimate compilations, particularly for the Common Era, limiting our understanding of Southern Hemisphere climate variability and reducing the spatial representativeness of global SST reconstructions.
- 14-15. Results: "About 24 records indicated that SST anomalies fluctuated throughout the entire Common Era, with values between (-5.84 to 0.62)." Further analysis needs to be performed (see comments below), and then this statement of what has been learned from the dataset can be made more specific.
Author's response: We revised the anomaly calculation and updated the methodology accordingly. As a result, the SST composite and associated interpretations were revised, and additional methodological clarifications were added to the manuscript (see revised Data section and detailed responses below).
- Introduction
- 38-43: Make a stronger case for the need for southwest Atlantic records.
(1) For available paleoclimate observations, the authors can point out that this scarcity is not resolved in updated curated Common Era paleoclimatic databases (e.g. Evans et al 2026, ESSD, and references to component databases therein).
(2) Most importantly, the authors need to make the case here: what we might learn from southwestern Atlantic paleoceanographic records, perhaps drawing from prior paleoclimatic studies, and from modern physical, chemical and biological oceanographic studies. What are the specific outstanding questions concerning regional ocean and climate change? What is the specific role of the southwestern Atlantic in (l.42) "driving and modulating global ocean and atmospheric circulation patterns"?
(3) What is meant by (l. 43) "...ensuring optimal efficiency" in efforts to address the scarcity of observations in this region?
Author's response: We have substantially revised the Introduction to strengthen the motivation for focusing on the SW Atlantic. We now explicitly highlight that the scarcity of Common Era paleoclimate records in this region persists even in updated, carefully curated global compilations, underscoring a major spatial gap in current datasets. We also expanded the Introduction section to better articulate the scientific value of this region’s paleoceanographic records, emphasizing that it is a key component of the Atlantic Ocean's large-scale circulation system. Finally, we revised the wording of the statement referring to “ensuring optimal efficiency” to improve clarity. This expression was rewritten to more accurately reflect that the goal is to improve the effectiveness of data acquisition and synthesis efforts by prioritizing regions with major observational gaps and high climatic relevance, rather than implying an undefined notion of “efficiency”. We believe these revisions substantially strengthen the Introduction by better situating the Southwest Atlantic within the broader context of global ocean–climate dynamics and by clearly articulating the scientific motivation for compiling and analyzing these records.
- 46-50: Missing from the introduction, but needed: an introduction to the ways in which SST was reconstructed from the various proxy observations made in the gathered sediment cores: MAT, UK'37, Mg/Ca; the uncertainties in the reconstruction of SST from these proxies and in this particular region, especially considering the relatively shallow water depths and regional oceanography from which the sediment cores were retrieved and the potential for the dataset to reflect terrestrial influences such as runoff, erosion, and anthropogenic accelerations of effects; and atmospheric influences on the coastal ocean setting. This will be useful for assessing and expanding results and discussion from analysis of the dataset. Some of the regional oceanographic background is at l. 102-110.
Author's response: We thank the reviewer for this helpful suggestion. We agree that the original Introduction did not sufficiently describe the proxy types used for SST reconstruction, nor the key sources of uncertainty and regional oceanographic/terrestrial influences that may affect these records. In response, we have included a clearer and more explicit description of the proxy approaches used in this compilation, including MAT, UK’37, and Mg/Ca-based SST reconstructions. We also added brief information on the uncertainties inherent to each proxy system, particularly in coastal and shallow-water environments, as represented in this dataset. Furthermore, we expanded the text to acknowledge the regional oceanographic and environmental factors that strongly modulate SST patterns in the SW Atlantic continental shelf. We also clarify that such processes may introduce additional uncertainty and variability in reconstructed SST signals and therefore must be considered when interpreting regional climate variability.
- Data
- 49-59: Provide original data source URLs in the main text. Describe the specific criteria, and give the rationale for inclusion of data in the database presented here: proxy observation, region, time coverage, time resolution, chronological control, other ... ? These would be important for replicating the work and for adding to the database in the future.
Author's response: We have provide URLs for the original data source in the Data Availability statement at the end of the paper and added hyperlinks to Supplementary Table 1. Also, between lines 102-105, we added “…for this study, record selection was based on data availability. Hence, all publicly accessible SST records from the SW Atlantic that covered at least part of this interval, regardless of whether they meet the PAGES 2k criteria…” to clarify that we included all publicly available SST records for the region.
- 60-71: exploratory visualization of SST anomalies. In this description of how the data were studied, I think the use of broad time windows and regional integration should be justified from a scientific standpoint. Then it should also be justified by the chronological control (see notes on revisions to Supplementary Material). Because the averaging is over time intervals between 150 and 550 years long, it might be better to use a historical climatology averaged over a larger spatial area and a longer time period as reference period. Instead of "summer" (l. 68), define the months within the year over which the average is performed. The rationale for a seasonal average should be described and sources cited, in the Introduction (see notes on l. 46-50).
Author's response: We thank the reviewer for this important comment. We clarify that the objective of this study is to provide an exploratory synthesis of available SST proxy records in the SW Atlantic and to highlight their temporal structure, spatial coverage, and inherent limitations. Aggregating the records into multi-centennial intervals enables more consistent comparisons among datasets with differing temporal resolutions, sampling densities, and chronological controls, while also allowing us to identify periods of lower data availability. We emphasize that the variability in the number of observations within each interval reflects the structure and limitations of the currently available dataset itself. Therefore, we have revised the manuscript to better justify the use of broad temporal windows from both scientific and chronological perspectives, explicitly noting that the adopted intervals are consistent with the temporal resolution and age uncertainties of many of the marine sediment records included in the compilation. We also revised the SST reference framework and now explicitly define the climatological reference period used in the anomaly calculations in both the main text and figure descriptions. Additionally, we clarified the treatment of seasonality throughout the manuscript by explicitly defining the months associated with the seasonal averages and revising the terminology where appropriate. We also expanded the Introduction and Methods sections to better explain the rationale for using seasonal SST estimates in proxy-based reconstructions, and we included relevant references on proxy seasonality and regional oceanographic influences.
- 74: Fig 2: (1) Please revise to indicate the differential lengths of the time periods studied: make boxplots the width of the named time periods, and give time in years CE on the horizontal axis in addition to the names. This will help readers unfamiliar with these names to understand the results. It also may provide a discussion point: since most of these names, if not all, originated with studies in the North Atlantic, to what extent does the exploratory visualization assume an Atlantic-wide common temporal expression of SST anomalies? This is not certain to be the case for nearshore/shallow water depth records from the southwest Atlantic (really - Brazil-Argentinian Margin might be a better description of the region sampled).
Author's response: We thank the reviewer for this helpful suggestion. We have revised Figure 3 (former Figure 2) to improve its interpretability by including the corresponding calendar year ranges (CE) alongside the period names on the x-axis. We note that, rather than modifying boxplot widths, we retained a constant width to preserve standard boxplot conventions and avoid potential misinterpretation of the statistical distributions. The differing durations of each period are now explicitly indicated through the labeled year ranges. We agree that, in fact, most of these climatic events, temperature expressions, and names, if not all, originated with studies in the North Atlantic; and that, as already highlighted in the literature, these are not necessarily observed in records elsewhere.
(2) Fig 2: Because the boxplots indicate nonnormal distributions within time intervals, I suggest you plot medians for each record within each period and boxplot.
Author's response: We thank the reviewer for this constructive comment. We clarify that the boxplots represent the full empirical distribution of individual annual SST anomaly observations within each time interval, defined by the 2.5th, 25th, 50th, 75th, and 97.5th percentiles. This percentile-based approach was adopted to better represent the non-normal distributions commonly observed in paleoclimate proxy data. The colored symbols represent core-level summaries, where each point corresponds to the median SST anomaly for a given core within each time interval. This provides a robust measure of central tendency that is less sensitive to non-normality and potential outliers within individual records. The vertical position of each symbol directly reflects this median value, as defined in the data processing workflow. We clarify that Panel B presents the number of observations per core within each climatic interval as a diagnostic view of sampling distribution and data coverage. This panel is intended solely to illustrate spatial and temporal heterogeneity in sampling density and is not used in any statistical calculations or in the construction of the SST anomaly distributions shown in Panel A, which are based exclusively on the full set of individual observations. Figure 3 (former Figure 2) has been revised to more clearly distinguish between the distribution of individual observations (boxplots) and the core-level summaries, improving the clarity of the hierarchical structure of the dataset. To further enhance transparency and reproducibility, we have added a supplementary table (Supplementary Table 2) reporting, for each core and period, the median SST anomaly and the number of contributing observations.
- 67-68: Explain the rationale for comparing the results to tropical global SST anomalies from Abram et al (2016). Background for this rationale should be in the Introduction (see notes on l. 38-43).
Author's response: We thank the reviewer for this important set of comments regarding Figure 4 (former Figure 3) and the associated methodological clarity. We agree that the original version of the figure and its description lacked sufficient information regarding data construction, comparability between datasets, and uncertainty representation, which could lead to misinterpretation. In response, we have substantially revised both the manuscript text and the figure caption to improve transparency and clarity. First, we clarified the origin and construction of the SST anomaly time series shown in the figure. The global and Southern Hemisphere SST records are now explicitly identified as derived from the gridded instrumental SST products of Kennedy et al. (2019), and this reference has been added to the manuscript and reference list. We also clarified that these datasets represent large-scale instrumental SST reconstructions and are therefore not directly comparable in resolution or origin to the proxy-based Southwest Atlantic composite developed in this study, but are included to provide a broader climatic context. Second, we revised the figure caption and main text to clearly describe how the Southwest Atlantic time series was constructed from the underlying proxy records. We now explicitly state the steps used to generate the composite, including the anomaly calculation for each record and the aggregation method applied to produce the regional series. Regarding uncertainty, we acknowledge the reviewer’s important suggestion; however, as this figure is presented as an exploratory comparison, we are not representing uncertainty. Finally, we have improved the rationale for including the SH and global SST series by explicitly stating that they are used as contextual benchmarks to situate regional variability within a broader climate framework, rather than as directly equivalent reconstructions.
- Results
- 73: I suggest creating a section 3, Results, which contains Figs 2-4 and the text describing the results of the analysis of the database.
Author's response: In response to the RC2 comment, we have added a new section, “SW Atlantic Common Era SST anomalies”. In this section, we have added a more detailed description of the database analysis results and discussion.
- 80-96 and Fig 2: Because the SST reconstructions are from shallow water depths, and from Fig 2 results, I suggest making more robust inference on boxplot medians rather than ranges. This might be also justified by introduction to the uncertainties in creating the SST reconstructions, and estimates of the SST reconstruction uncertainties (l. 46-50).
Author's response: We thank RC2 comment. We now describe the results and draw our inferences regarding SST anomalies in the multi-centennial bins of Figure 3b (former Figure 2), based on boxplot medians rather than anomaly ranges.
- 115 and Fig 4: I think this figure logically comes next, as it addresses the question of whether the Brazil-Argentinian Margin dataset introduced here shows similar Common Era variability as does the Ocean2k compilation of global SST. I think that data should be presented similarly to the boxplots as shown in Fig 2, for reasons that are detailed in the McGregor et al (2015) paper which is cited and sourced for the global mean reconstructed SST comparison. Specifically, all the gathered data for that 2015 paper were on different timescales, resolutions, and nominal chronologies, so were perhaps best considered as binned composites, similar to what is done in Fig 1. This also has the benefit of allowing the authors to make inference on medians and changes in medians between one time period and the next. It looks like the same resolvable features are in both composites: relatively flat in the first millennium, with a cooling in the 1300-1800 period.
Author's response: We thank the reviewer for this suggestion. We have changed the figures in the manuscripts and their order of presentation. Figure 4 has now been inserted as panel a into Figure 3, where we present the SW Atlantic SST anomaly distribution and sampling density in the Common Era as the compilation curve and the multi-centennial boxplots. In addition, we have clarified in the Methods section the rationale for using the Ocean2k approach in the construction of Figure 3a (former Figure 2)
121-122: "These findings demonstrate that the SW Atlantic lacks adequate records to represent Common Era SST anomaly changes...": This discussion argument seems to be contradicted by the results in Fig 3.
I understand that the authors argue that it is a change in the relative representation of warm vs cold current influenced regions in the regional SST reconstruction that creates the apparent agreement noted in the prior comment. This needs more support from data analysis. For instance, looking at Fig 1 for core locations, and then Fig 2 and Fig 5, it seems like:
Author's response: This sentence has been moved to the summary and conclusions section, which has been significantly revised to better support our findings.
- 110-112: The authors write: "Notably, the sedimentary
records north of 25°S span 778–2010 CE and are influenced by the Cabo Frio coastal upwelling system, which may amplify the regional SST cooling signal from the LIA to the present (Fig. 5)." The authors seem to be referring to record NAP62-2; other northern records are entirely within the LIA. Yet it seems from Fig 1 that at least 3 other records span this period and are within the current (no pun intended) influence of the warm Brazil Current (Fig 1) during the LIA.
Within the Little Ice Age period (Fig 2) there are 10 records available spanning a wide range of color codings. If those color codings are as indicated in Fig 5, then a wide range of sites along the coast, and in the influence of the warm Brazil Current, are sampled.
A comparison between record LaPAS-KF02 spanning nearly the entire time period, and NAP62-2, might indicate whether this is the case. Of course, if the data are composited as anomalies relative to modern means, then the effects of differences in mean SST near the upwelling region and across the transect will be removed from the composite already.
If the authors wish to argue that there has been a shift in SST along part of this transect, which might be related to the global or Atlantic basin SST reconstructions over the Common Era, as composited by McGregor et al (2015): this might be speculated using physical oceanographic arguments and changes in winds in the historical era, but evidence has not been presented to support that idea.
Author's response: We thank the reviewer for this observation regarding the interpretation of regional patterns and the spatial distribution of records, particularly within the LIA interval. We agree that the original wording may have overemphasized the role of a single regional mechanism (i.e., Cabo Frio coastal upwelling) and did not fully reflect the spatial heterogeneity of the dataset shown in Figure 3 (former Figures 1 and 5). In response, we have revised both the text and the underlying interpretation to better reflect the complexity highlighted by the reviewer. Specifically, we re-evaluated the spatial distribution of records across the study domain and clarified that SST variability during the Common Era, including the LIA interval, is not solely driven by discrete regional regimes. We appreciate this comment, as it helped improve the consistency between the spatial interpretation of the figures and the text, and led us to refine our discussion of regional variability and its drivers.
I am unable to further assess the authors' arguments by, for instance, making a plot of mean reconstructed SST vs site latitude, or a composite of selected subsets of the data by latitude, or comparing across subsets by proxy observation, because the authors have only provided a single value for each time point in the compilation in their Figshare repository (l. 150-152), rather than the SST reconstructions by core and providing their associated metadata. Please provide the complete database and all metadata in the repository, such that further exploration of the dataset is possible.
Author's response: We agree that the original repository structure did not fully allow independent exploration of the underlying reconstructions by individual cores. We have therefore revised both the supplementary material and the data availability to improve reproducibility and usability of the dataset. We now provide a completer and more detailed dataset, including SST anomaly values for each core (Supplementary Table 2), along with the associated metadata necessary for interpretation. This includes information on the proxy type for each record, the reference periods used for anomaly calculations, and the modern calibration context, indicating that all proxies reflect mean annual SST. In addition, we have added a new figure (Figure 3c) illustrating the number of records and their latitudinal distribution within each multi-centennial bin, allowing users to directly assess spatial coverage through time. We also clarified the construction of the composite and explicitly documented how individual records contribute to the final aggregated time series. Finally, we updated the repository description to ensure users can access both the compiled composite and the underlying core-level SST reconstructions, along with the metadata needed for further analysis (including latitude, proxy type, and reference period definitions). These changes were made to ensure full transparency and to enable independent evaluation and further exploration of the dataset by other researchers.
Unfortunately, given these comments (1-4), I find that the authors have not provided convincing support for their argument at l. 112-14, and their conclusion at l. 120-126.
(1) l. 194-196: "Overall, unlike the global trend of rising SSTs observed in several studies, especially highlighting the marked post-industrial warming (e.g., Neukom et al. 2019), the SW Atlantic shows an opposite pattern with a cooling trend from 1850 to 2020 (Fig. 2)." With only a single boxplot in Fig 2 for post-1850 data, this cannot be judged. The results in Fig 3 (blue line) seem to indicate that all values are cold relative to the 1961-1990 mean, and perhaps this is not surprising, as the 1961-1990 period postdates the vast majority of the data in the series. But if anomalies were calculated relative to the 1961-1990 mean, and these remain 2-3 degrees colder than that reference period through the temporal overlap period (Fig 3) - something seems wrong. please check if there might be a mistake in the figure?
Authors response: In the original version of the manuscript, SST anomalies were calculated relative to a fixed climatological reference period, which may indeed have contributed to the apparent cold bias highlighted by the reviewer. Following this comment, the anomaly calculation was revised. SST anomalies are now calculated as the difference between proxy-derived SST at each depth and the median proxy-derived SST value representing modern conditions within the same record, calculated individually for each core using all available data from 1800 yr BP onwards. Therefore, both the reconstructed SST values and the reference baseline are derived from the proxy dataset itself rather than from an instrumental climatology. To evaluate the spatial consistency of the proxy-derived reference baseline, the median proxy-derived SST values were compared with instrumental SST datasets (OISST and WOA), revealing significant relationships that indicate the proxy-derived SST estimates capture relative spatial SST variability across the study area.
This revised approach minimizes biases associated with uneven temporal coverage among records. The predominantly negative anomalies previously observed no longer reflect systematic offsets relative to a fixed modern instrumental reference period. The methodology, figures, and captions have been revised accordingly.
- 197 and Fig 3: I think this figure lacks rationale and I would differ in my interpretation of the results.
(2) Fig 3, time series: how was this constructed? Are there uncertainty estimates on the x and y axes? See also comments on introduction to the reconstructions and apply that information to the revision of this figure. The boxplot composite formulation used in Figs 2 and 4 might usefully enable representation of age and observational uncertainty here.
(3) It's unclear to me why a Southern Hemisphere and global average SST record would be comparable to the SW Atlantic anomaly series, but the legend gives me too little information to understand what is here. Are the SH and GL averages from historical gridded estimates? I don't find Kennedy et al (2019) in the reference list. I don't know how the averages were made. Earlier there was reference to "summer" values; is summer also used for creating these averages?
Author's response: We thank the reviewer for this important set of comments regarding Figure 4 (former Figure 3) and the associated methodological clarity. We agree that the original version of the figure and its description lacked sufficient information regarding data construction, comparability between datasets, and uncertainty representation, which could lead to misinterpretation. In response, we have substantially revised both the manuscript text and the figure caption to improve transparency and clarity. First, we clarified the origin and construction of the SST anomaly time series shown in the figure. The global and Southern Hemisphere SST records are now explicitly identified as derived from the gridded instrumental SST products of Kennedy et al. (2019), and this reference has been added to the manuscript and reference list. We also clarified that these datasets represent large-scale instrumental SST reconstructions and are therefore not directly comparable in resolution or origin to the proxy-based Southwest Atlantic composite developed in this study, but are included to provide a broader climatic context. Second, we revised the figure caption and main text to clearly describe how the Southwest Atlantic time series was constructed from the underlying proxy records. We now explicitly state the steps used to generate the composite, including the anomaly calculation for each record and the aggregation method applied to produce the regional series. Regarding uncertainty, we acknowledge the reviewer’s important suggestion; however, as this figure is presented as an exploratory comparison, we are not representing uncertainty. Finally, we have improved the rationale for including the SH and global SST series by explicitly stating that they are used as contextual benchmarks to situate regional variability within a broader climate framework, rather than as directly equivalent reconstructions.
(4) The scale of variation for the regional composite is a factor of 4x greater. If the imprint of forced climate change locally is roughly the same as on global SST (regress the global SST field on the global SST time series to find out), then is it possible that the unforced variability, for coastal locations in this region, especially for shallow water depths, is more atmospherically and therefore stochastically controlled? In which case you might expect to see what is in the blue line: no change over the historical period and a high level of interannual variability which is overlain by variability associated with changing sampling resolution for the reconstructions.
Authors response: We agree with the reviewer’s point that local and regional processes may exert a strong influence on the SST anomalies observed in this coastal setting, particularly in shallow-water environments where variability can be more directly and stochastically forced by atmospheric conditions. We also acknowledge that the magnitude of regional variability relative to the global signal can obscure the expression of the long-term forced climate trend in the reconstructions. In response, we have added a paragraph to the manuscript explicitly discussing the role of local/regional effects on SST anomalies, including the influence of enhanced interannual variability, atmospheric forcing, and potential impacts associated with changes in reconstruction sampling resolution through time.
- Summary and conclusions
- 128-137: I can agree with the first 2 sentences. The third is not supported by evidence presently provided in the manuscript, and might be contradicted by evidence that is presented. The fourth could be true but is not supported by data analysis, and is not currently accessible by data and metadata provided with the manuscript, but could be.
Author's response: We have significantly modified the summary and conclusions section, and both the third and fourth sentences have been removed.
- 138-148: A discussion of the extent to which high-resolution offshore sediment records could be obtained in the region would help support conclusions here. The authors could cite recent publications of datasets in the NOAA/NCEI repository (https://www.ncei.noaa.gov/access/paleo-search/), and also publications in ESSD, for instance:
https://doi.org/10.5194/essd-12-2261-2020
https://doi.org/10.5194/essd-12-2579-2020
https://doi.org/10.5194/essd-15-2081-2023
https://doi.org/10.5194/essd-16-1933-2024
https://doi.org/10.5194/essd-2025-812
https://doi.org/10.5194/essd-18-1185-2026
https://doi.org/10.5194/essd-18-1921-2026Author's response: We thank the reviewer for this valuable suggestion. We expanded the Summary and Conclusions section to include an assessment of the importance and feasibility of acquiring offshore sediment records from the region. We now emphasize that many of the currently available SST reconstructions are derived from shallow coastal environments strongly influenced by local oceanographic processes, which can obscure broader-scale climate signals. We therefore highlight that future efforts should prioritize offshore locations, particularly along the continental slope and in regions influenced by large-scale circulation systems such as the Brazil Current, where sedimentary archives may better preserve regional- to basin-scale SST variability. We also acknowledge that acquiring high-resolution offshore records in the South Atlantic remains challenging due to logistical limitations, uneven spatial coverage, sedimentation-rate variability, and the high costs of marine coring campaigns. Nevertheless, we emphasize that expanding offshore reconstructions represents a critical step toward improving the representation of Southern Hemisphere climate variability in Common Era paleoclimate syntheses and strengthening our understanding of South Atlantic ocean–climate dynamics. We also incorporated references to recent paleoclimate dataset publications in ESSD and acknowledged the role of NOAA/NCEI paleoclimate repositories and international data synthesis initiatives in promoting standardized, accessible paleoclimate datasets. These additions help better contextualize our compilation within ongoing global efforts to improve the representation of under-sampled regions in paleoclimate databases.
Supplementary Material:
Table 1: It would be helpful to have columns for 'bottom date' and 'top date' to allow the reader to see which age models might be best constrained by chronological control, which ones include chronological extrapolation (if any); additional information as mentioned previously would be valuable for dataset users.
Author's response: Following RC2 comments, we have modified Table 1.
Supplementary Figure 1: these data are overwhelming from very shallow water depths <150m. Therefore very close to present shoreline. Is there discussion of what this might mean for any interpretation of the results? Yes, there is! But see prior comments for further analysis and interpretation.
Author's response: This has been addressed; please refer to previous comments.
References
Please include links to original publicly accessible datasets, if available, and if not, provide the full dataset and all its metadata in the Figshare repository.
Citation: https://doi.org/10.5194/essd-2026-122-RC2
Author's response: We update the data in Supplementary Table 1, all the anomalies used allow temperatures to be calculated, even for those for which the data are not available in publicly available datasets. Also, we updated the repository with all data.
Citation: https://doi.org/10.5194/essd-2026-122-AC2
-
AC2: 'Reply on RC2', Renata Hanae Nagai, 29 May 2026
Data sets
SW Atlantic sea surface temperature anomalies J. dos Reis de França et al. https://doi.org/10.5281/zenodo.18392297
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 358 | 229 | 26 | 613 | 71 | 32 | 36 |
- HTML: 358
- PDF: 229
- XML: 26
- Total: 613
- Supplement: 71
- BibTeX: 32
- EndNote: 36
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1