the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Feb 2024
Probabilistic reconstruction of sea-level changes and their causes since 1900
Abstract. Coastal communities around the world are increasingly exposed to extreme events that have been exacerbated by rising sea levels. Sustainable adaptation strategies to cope with the associated threats require comprehensive understanding of past and possible future changes. Yet, many coastlines lack accurate long-term sea level observations. Here, we introduce a novel probabilistic near-global reconstruction of relative sea-level changes and their causes over the period 1900 to 2021. The reconstruction is based on tide gauge records and incorporates prior knowledge about physical processes from ancillary observations and geophysical model outputs. We demonstrate good agreement between the reconstruction and satellite altimetry and tide gauges (if local vertical land motion is considered). Validation against steric height estimates based on independent temperature and salinity observations over their overlapping periods shows moderate to good agreement in terms of variability, though with larger trends in three out of six regions. The linear long-term trend of the resulting global mean sea level (GMSL) record is 1.5±0.19 mmyr-1 since 1900, a value consistent with central estimates from the 6th Assessment Report of the Intergovernmental Panel on Climate Change. Multidecadal trends in GMSL have varied with enhanced rates in the 1930s, near-zero rates in the 1960s, and a persistent acceleration (0.08±0.04 mmyr-2) thereafter. As a result, most recent rates have exceeded 4 mmyr-1. Largest regional rates (>10 mmyr-1) over the same period have been detected in coastal areas near western boundary currents and the larger tropical Indo-Pacific region. Barystatic mass changes due to ice-melt and terrestrial water storage variations have dominated the sea-level acceleration at global scales, but sterodynamic processes are the most crucial factor locally, particularly at low latitudes and away from major melt sources. These results demonstrate that the new reconstruction provides valuable insights into historical sea-level change and its contributing causes complementing observational records in areas where they are sparse or absent.
- Preprint
(10455 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 09 May 2024)
-
RC1: 'Comment on essd-2024-46', Anonymous Referee #1, 01 Apr 2024
reply
In this paper, Dangendorf and colleagues present a reconstruction of regional and global mean sea-level change and its different drivers since 1900, based only on observations and statistical methods (to reconstruct sea level back in time). The dataset presented in this manuscript is a valuable addition to the field, and can be used in a range of scientific applications. The manuscript explains well the method and discusses some of the regional patterns of the reconstruction, displaying the multitude of processes that can be investigated with this dataset. My only concern here is that it should be made clearer to the reader that this is an update of a previous reconstruction, with some, albeit very important, alterations, so that the novelty of the dataset is not overstated. With this made it clear, and some small alterations, the manuscript deserves to be published to the scientific community.
Main comments:
- Novelty in relation to Dangendorf et al (2019).: The need to update the reconstructions available is well delineated in the introduction, which makes the issue and novelty of the study very clear. But I think it should be more explicit the differences and updates regarding the previous reconstruction from the lead author (Dangendorf et al., 2019). This is study is very briefly mentioned in the introduction, but it deserved more attention at the beginning to make it clear that this is an update/improvement of the previous reconstruction. From the introduction, that is not clear, and the reader is led to think that this is an update of Hay et al (2013; 2015). In the discussion, however, it becomes slightly more clear that this is an update of Dangendorf et al (2019) [Lines 608-610]. Only in the discussion [Lines 608-610] the authors say explicitly that this is an update of Dangendorf et al. (2019), extending the previous reconstruction by 6 years.
- Colormap of figures: In all sea-level literature, warm colors are related to positive trends, and cold colors to negative. The colormap used by all the authors is the opposite, which is very confusing. I really urge the authors to flip the colormap, so that blue will be negative and red positive.
Minor comments:
- Not sure if there is space in the abstract, but just by reading it, it was unclear what was the ‘novelty’ regarding the previous reconstructions.
- L21: make it clear that the reconstructions were larger (though with larger reconstructed trends in three out of six regions).
- L24: “most recent rates”… unclear to what period this refers. Please give a year.
- L49: Is there a more recent GMSL acceleration that could be mentioned (instead of a 2018 publication)?
- L127: I’m not so familiar with these methods, but how are the tide-gauge observations, which is in RSL, and satellite altimetry, which is ASL, combined in the state vector?
- L154: does this mean than that everything before 1955 is only based on statistical modelling?
- L161 (and L203): Why only one IBE estimate? Were there no other reanalysis available, to have a larger number, like for GRD and GIA…? And this IBE estimate is the combination of two reanalysis (one 1900-2015 and one 2016-2021), were there no single one that covered the entire study period? Do these two dataset have any significant differences (apart from period)?
- L191: Why only 5 years were considered to extrapolate the last year of GRD?
- L205: Use the grid resolution instead of number of grid cells (same for other occasions when the number of grids is mentioned (e.g., Section 4)). For example, L222: Does this mean that you at the end have a 4-degree resolution dataset?
- L213: Satellite altimetry was also used as input into the state matrix of the Kalman filter, no?
- Validation with satellite altimetry: If altimetry was used in the Kalman filter, is this a fair validation?
- L231: A suggestion, but you could use this value (r=0.89), or another as a reference contour line in the global map (Figure 2a).
- L241: Does this mean that there is a resolution limit to the reconstruction?
- L255: Is there a way of indicating which regions that have a significant residual? If hatching all the areas that are insignificant would make the figure “too dirty”, maybe you could mention in the text, since its such a small percentage, where the residuals are significant?
- Figure 3: The differences (last row) in the accelerations seems much larger, but this is probably an artefact of the color range. Maybe for these panels, both the trend and acceleration should be on similar color range, or the differences should be normalized.
- Sterodynamic estimates: There are some regional budgets, such as Wang et al (2021), Royston et al (2020) and Camargo et al (2023), that do have sterodynamic estimates. So these could be used to validate the sterodynamic of the reconstructions, instead of only independent steric estimates.
- L283: Here and also L287 you say ‘Northwest Pacific’, but on the table 1 and 2 is only ‘West Pacific”.
- L238: The agreement of the reconstruction is better with satellite altimetry than with tide gauge records, reflecting the dependence of the reconstruction on the quality of the altimetry in the coastal zone. In addition to the processes you mention here that altimetry measurements might misrepresent in the coastal zone, there are also a large number of correction issues in the coastal zone (see Vignudelli et al, 2019, for example). If the reconstruction is made with ‘standard’ altimetry product, and not dedicated coastal products (e.g., X-TRACT (Birol et al., 2017), ALES+ (Passaro, 2018)), it will inherit these issues. So, does this mean that the user should take care when using this product in the coastal zone (and say, not consider the most inshore grids?) In this case, a sentence saying this should be added. On the same note, I was wondering: (1) could the reconstruction be improved by including coastal altimetry products? (2) would the validation with the ‘virtual tide gauges’ (Cazenave et al., 2022) give better results than the validation with the standard tide gauges?
- L353: here you mention ‘64%’ and ‘33%’ of the tide gauges…what about the other 3%?
- L355-358: these are also locations with large non-linear (and non-GIA) VLM, such as earthquakes hot spots (U.S. west coast) and land subsidence (Gulf of Mexico). Could this also add to the large differences?
- L364: Could add a comment about why the differences with the tide gauges have so much larger uncertainties than the differences with satellite altimetry?
- L377-384: This part was very confusing for me, and required me to read several times to understand the different what trend values and number of tide gauges (90, 516, 119) were referring to. For example, the 0.27±1.67 and 0.1±1.8 mm/yr trends, are these average of the 90 and 516 locations? And the difference refers to KS reconstructed with GIA – tide gauges?
- L400-402: Please add the periods of Church & White (2011) and jevrejeva (2014), instead of just say ‘for shorter periods’. And add uncertainties to these trends, if provided.
- Table 3: You could also add the acceleration rates to the table (so similar to Table 1)
- L424 (and Fig 8): Why is the comparison made with D19 only from 1930-1939? And in the discussion (L612-614) the results are compared for 1993-2021?
- L428-429: This information should be given before, either in the introduction or methods.
- In Figure 10 you compare the accelerations with Palmer et al (2022), but this comparison never comes up in the text.
- L497: And to a smaller extent also due to GRD contribution
- L515: It seems like there is something missing in this sentence “The local maxima result from additional around…”
- L590: Are the codes ready for publication already? Or does this mean that there will be another publication for the code? It was unclear.
- L630: It’s a bit circular that the peak matches better with the ones from Frederikse et al (2020), since the barystatic GRD used in the reconstruction is from Frederikse et al (2020).
Editorial comments:
- L33: add comma after warming climate.
- L36: add comma after parenthesis.
- L50: add comma after parenthesis.
- L78: “unrelated through GIA”: should this ‘through’ be a ‘to’?
- L92: Add comma after observations.
- L102: I always see only ‘IB’ instead of ‘IBE’…
- L107: change ‘or’ to ‘and’
- L108: ‘MSL’ is introduced here, but not used again afterwards…
- Equation 2: first symbol on the right-hand side is not introduced/explained in the text.
- L128: modification instead of adjustment? (adjustment has a more negative connotation, seems like the previous work was wrong). And same for L129.
- Equation 3: Should the ‘and’ be in a new line?
- L166: use VLM, which was already introduced.
- L171: should it be 4000 (L162)?
- Figure 1a: Some of the symbols on the dark red and blue are a bit hard to see. Maybe using lighter shades for the background colors might improve it.
- L275: Reference should be Camargo (2020).
- L349: Change ‘Fig.’ to Figure (as it has been in all other instances)
- L381: Here you have ‘-0.14±1.00 mm/yr’, but Fig 6c says -0.13.
- L400: Here you have 1.5±0.19, but Table 3 is 1.5±0.20.
- L501: I would suggest using ‘scarcity’ instead of ‘paucity’, but that’s because I had to look up this word.
References
- Royston, S., Vishwakarma, B. D., Westaway, R., Rougier, J., Sha, Z., and Bamber, J.: Can We Resolve the Basin-Scale Sea Level Trend Budget From GRACE Ocean Mass?, J. Geophys. Res.-Oceans, 125, 1–16, https://doi.org/10.1029/2019JC015535, 2020.
- Wang, J., Church, J. A., Zhang, X., Gregory, J. M., Zanna, L., and Chen, X.: Evaluation of the Local Sea-Level Budget at Tide Gauges Since 1958, Geophys. Res. Lett., 48, 1–12, https://doi.org/10.1029/2021GL094502, 2021.
- Camargo, C. M. L., Riva, R. E. M., Hermans, T. H. J., Schütt, E. M., Marcos, M., Hernandez-Carrasco, I., and Slangen, A. B. A.: Regionalizing the sea-level budget with machine learning techniques, Ocean Sci., 19, 17–41, https://doi.org/10.5194/os-19-17-2023, 2023.
- Vignudelli, S., Birol, F., Benveniste, J. et al.Satellite Altimetry Measurements of Sea Level in the Coastal Zone. Surv Geophys 40, 1319–1349 (2019). https://doi.org/10.1007/s10712-019-09569-1
- Birol, F., N. Fuller, F. Lyard, M. Cancet, F. Niño, C. Delebecque, S. Fleury, F. Toublanc, A. Melet, M. Saraceno, F. Léger, 2017. “Coastal Applications from Nadir Altimetry: Example of the X-TRACK Regional Products.” Advances in Space Research, 2017, 59 (4), p.936-953. doi:10.1016/j.asr.2016.11.005
- Passaro, Marcello, Stine Kildegaard Rose, Ole B. Andersen, Eva Boergens, Francisco M. Calafat, Denise Dettmering, Jérôme Benveniste, ALES+: Adapting a homogenous ocean retracker for satellite altimetry to sea ice leads, coastal and inland waters, Remote Sensing of Environment, Volume 211, 2018, Pages 456-471, ISSN 0034-4257, https://doi.org/10.1016/j.rse.2018.02.074.
- Passaro, M. et al. ALES: A multi-mission adaptive subwaveform retracker for coastal and open ocean altimetry. Remote Sensi. Environ.145, 173–189 (2014).
- Cazenave, A., Gouzenes, Y., Birol, F. et al.Sea level along the world’s coastlines can be measured by a network of virtual altimetry stations. Commun Earth Environ 3, 117 (2022). https://doi.org/10.1038/s43247-022-00448-z
Citation: https://doi.org/10.5194/essd-2024-46-RC1
Data sets
Kalman Smoother Sea Level Reconstruction Sönke Dangendorf https://doi.org/10.5281/zenodo.10621070
Model code and software
Kalman Smoother Sea Level Reconstruction Sönke Dangendorf https://doi.org/10.5281/zenodo.10621070
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 405 | 161 | 34 | 600 | 16 | 16 |
- HTML: 405
- PDF: 161
- XML: 34
- Total: 600
- BibTeX: 16
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1