the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Reconstructing sea level rise at global 945 tide gauges since 1900
Abstract. Tide gauges record sea level changes along coast. They are widely used to determine the twentieth century global mean sea level (GMSL) rise. However, a major issue in tide gauge data is the presence of various, substantial, and sometime persistent data gaps, which hinder our understanding of sea level rise, especially at regional and local scales. Whilst the GMSL reconstructions have been provided by several influential studies, reconstructions at the exact sites of tide gauges are rarely available. Here, we present sea level reconstructions at global 945 tide gauges, covering the period over 1900 to 2022. Our approach relies on a data assimilation technique that integrates various physical sea level observations and predictions, including sea level simulations from 35 climate models. A prominent feature in our reconstruction is that it provides an ensemble of 35 reconstructions at each site of tide gauge, offering complete and refined sea level time series. This ensemble reconstruction allows for direct statistical assessment, e.g., average, median, spread, and percentile. The average of reconstructed sea level across 945 tide gauges reveals a GMSL rise of 1.75±0.05 mm/yr over 1900–2020, and shows strong agreements with other GMSL reconstructions for both the curves of time series and overall trends. At local scale, our reconstructions are comparable to an independent reconstruction, despite apparent rate differences at locations, it is suggested that our reconstructed sea level trends closely follow the raw records when they are available, emphasizing the importance of the observed sea level rise at tide gauges. Our sea level reconstructions offer a valuable resource for improving global and regional sea level projections, validating climate model performance, and informing coastal adaptation strategies. The reconstructed sea level is available at https://doi.org/10.5281/zenodo.15385035 (Mu, 2025).
- Preprint
(4638 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 27 Aug 2025)
-
RC1: 'Comment on essd-2025-300', Anonymous Referee #1, 20 Jul 2025
reply
The paper reconstructed a century-scale sea level rise at tide gauges, using a data assimilation approach that has been proposed by previous studies, but I saw some modifications or improvements, e.g., introducing a random process. The data assimilation is indeed driven by physical mechanisms, and therefore the reconstruction, or essentially the interpolation or extrapolation are physically interpretable. Authors considered 35 CMIP6 model output and, consequently, they gave 35 reconstructions, this ensemble obviously offers some useful statistical assessments, and this is really convenient, users can compute a desirable uncertainty estimate. Authors compared their reconstructions to observations from satellite altimetry, and other sea level reconstructions that are widely accepted by the community. The comparisons were performed on both global and local scales, and the results seems promising, although some differences were still seen, especially at the sites of tide gauges. The new global mean sea level reconstruction can serve as an independent estimate, users can get a better ensemble for average and spread. Overall, I think the dataset can be potentially applied to sea level studies, and the community would benefit from it. However, I have several comments, and I hope authors can address them before I see the paper published.
(1) In the method section, authors used the HP filter to compute the instantaneous rate for SDSL changes. My question is how the authors determine the parameter lambda (i.e., equation 14)? As a filter, HP might be sensitive to the changes of lambda. Based on my own understanding, the filtered or smoothed SDSL is perhaps related to the smoothed sea level curves seen in, e.g., Figure 6. Another reason that might explain the smoothed curves is that authors used the Kalman filter and smoother for the sea level fingerprints, so the total sea level would be much smooth. Authors need to prove how the curves would vary with parameter lambda.
(2) There might be confusing explanation in Table 3. Treu et al. (2024) used the low-frequency sea level reconstruction from Dangendorf et al. (2019), who employed a hybrid reconstruction. This hybrid reconstruction combined traditional EOF reconstruction and the data assimilation, the former provided sea level variability, the latter provided long-term trends. But why authors claimed that ‘Differences are possible between reconstructions and raw records’, is this because they observed apparent discrepancies in Figure 10 when they compared with Treu et al. (2024). If so, I think there might be another reason, that is Treu et al. (2024) considered different selection of tide gauges. Anyway, authors should add some more wording to clarify.
(3) In Figure 9, authors compared many GMSL reconstructions to justify theirs. I saw some differences in the overall rates. Authors attributed the differences to reconstruction methods and selections of tide gauges, this is true, and I agree. But authors overlooked another fact, that is, the GMSL curves represent the relative sea level or absolute sea level? This is of course highly related to the reconstruction methods, but I think author should add some comments to this point, and the paper Dangendorf et al. (2017) might be helpful (https://doi.org/10.1073/pnas.1616007114).
(4) In section 4, authors pointed out some limitations in their reconstructions, this is very important and useful. Authors claimed that some abrupt changes are not removed from the raw records, and those changes are induced by earthquakes, but have no significant effect on sea level reconstruction, this is understandable, as they account for only a small portion. I could agree to this. However, authors should elaborate a little bit more on the second limitation. The refined trends could be informative for long-term changes, despite we did not know how ‘long’ it is. The community usually employs a 30-year window for computing GMSL rise, so why not compare all the GMSL curves for the 30-year-rate curves, I would expect some differences, even significant ones, readers can glean some useful knowledge, and it would be better if authors add more wordings.
Minor suggestions:
Line 13, period over 1900 to 2022 -> period from 1900 to 2022
Line 17, assessment -> assessments
Line 18, GMSL rise -> GMSL rate
Line 22, observed sea level rise at -> observations from
Line 27, collection -> collected
Line 27, add the reference Holgate et al. (2013) after the website
Lines 33 and 35, reference Calafat et al. (2022) was not shown in the reference list, correct it
Line 43, cause ->causes
Line 62, at tide gauges ->at the sites of tide gauges
Line 69, physical -> physically
Line 75, error in citation of Calafat et al. (2022) as shown before
Line 76, Since -> since
Lint 97, at tide gauges -> at the sites of tide gauges
Line 113, raw tide gauge records -> raw records of tide gauges
Line 132, include -> includes
Line 176, its -> their
Line 189, Instantaneous -> instantaneous
Line 190, variation -> variations
Line 215, sea-level -> sea level, you should be consistent about the writing
Line 226, to address the ‘a zero global mean’, you add ocean mass increase, what about the global mean of thermosteric sea level, how you exactly treat this term?
Line 230, figure 3 was not even cited, you might want to add some more wordings to describe the changes shown in figure 3
Line 235, error in caption of figure 4, e.g., panel (e) was missing
Line 240, rises -> rise
Line 249, remove ‘for this’
Line 300, AVISO sea level observations -> AVISO sea level products
Line 316, assessment -> assessments
Line 333, influence of sea level observations in -> influence of sea level observations on
Line 355, ensemble of subset -> subsets
Line 379, the GMSL -> the GMSL curves
Line 401, The resulting GMSL curve with raw records exhibits -> The resulting GMSL curves with raw records exhibit
Line 425, figure 11, add a panel showing the difference between 95th and 5th percentiles
Line 429, sea level rate -> sea level rates
Line 440, T2024 who provide -> T2024 who provided or provides
Line 458, 3.3 Statistical assessment -> 3.3 Statistical assessments, this correction applies to others, e.g., line 459
Line 505, add a plot showing the 30-year rates for all GMSL reconstructions
Citation: https://doi.org/10.5194/essd-2025-300-RC1
Data sets
Sea level reconstruction at global 945 tide gauges Dapeng Mu https://doi.org/10.5281/zenodo.15385035
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
209 | 70 | 23 | 302 | 12 | 18 |
- HTML: 209
- PDF: 70
- XML: 23
- Total: 302
- BibTeX: 12
- EndNote: 18
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1