the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Jun 2022
Importance of non-stationary analysis for assessing extreme sea levels under sea level rise
Abstract. Coastal flooding caused by extreme sea levels (ESLs) is one of the major impacts related to the climate change. It is expected to increase in the future due to sea level rise and storm surge intensification. Estimates of return levels obtained under the framework provided by extreme events theory might be biased under climatic non-stationarity. Additional uncertainty is related to the choice of the model. In this work, we fit several extreme values models to a long-term (96 years) sea level record from the city of Venice (NW Adriatic Sea, Italy): a Generalized Extreme Value distribution (GEV), a Generalized Pareto Distribution (GPD), a Point Process (PP), and the Joint Probability Method (JPM) under different detrending strategies. We model non-stationarity with a linear dependence of the model’s parameters from the mean sea level. Our results show that non-stationary GEV and PP models fit the data better than stationary models even with detrended data. The non-stationary PP model is able to reproduce the rate of extremes occurrence fairly well. Actualized estimates of the return levels for non-stationary models are generally higher than estimates from stationary models. Thus, projections of return levels in the future might be significantly different from those calculated using stationary models. Overall, we show that non-stationary extremes analyses can provide more robust estimates of return levels to be used in coastal protection planning.
-
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
(1317 KB)
-
Supplement
(75 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1317 KB) - Metadata XML
-
Supplement
(75 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-347', Anonymous Referee #1, 30 Jun 2022
The manuscript compares the return levels obtained from four different extreme value analyses accounting for non-stationarity in the Punta Della Salute tide gauge, Venice. The extreme value distributions include: (i) generalized extreme value (GEV) applied to a block maxima sampling (BM); (ii) generalized Pareto distribution (GPD) and (iii) a point process (PP) method, both applied to peak over a threshold (POT); and (iv) a joint probability method (JPM). In addition, the authors tested the implications of using three different detrending techniques, including (i) removing the annual mean sea level from the time series (MSL); (ii) removing the last 19 years’ mean sea level (MSL_L); and (iii) not detrending the data before fitting the distributions.
I find the topic important for coastal flood risk assessment as traditional designs have been based on analyses using direct methods that ignore the non-stationarities, which can lead to an underestimation of the risk, as found in previous studies. In addition, having a comprehensive analysis of the different non-stationary extreme value methodologies would help on the way to obtaining a more standardized analysis, facilitating the comparison of the results between the studies. Thus, the results of the manuscript are relevant for the scientific community and coastal risk stakeholders after improving the work in some aspects, mainly related to
- the level of comprehensiveness of the analysis: it will be interesting for the scientific community to include one of the most utilized indirect methods: the revised joint probability method. In doing so, the authors will also reduce the existing overlapping with previous studies,
- the level of applicability to other study areas by including more tide gauge records in the analysis, and
- the level of replicability (some information relevant for reproducibility is missed).
-
AC1: 'Reply on RC1', Damiano Baldan, 13 Jul 2022
Thank you for the constructive comments. In the revised submission we will perform some additional analyses and edit some sections of the manuscript to meet the comments from referee #1. See below a more detailed overview:
Reviewer’s comment
Authors' reply
Improving the work in the level of comprehensiveness of the analysis: it will be interesting for the scientific community to include one of the most utilized indirect methods: the revised joint probability method. In doing so, the authors will also reduce the existing overlapping with previous studies
We will include the revised joint probability method to the list of tested methods.
Improving the work in the level of applicability to other study areas by including more tide gauge records in the analysis
We acknowledge the sea level data from the case study in Venice might not be representative of other geographic locations given their high degree of non-stationaries (which in turn was useful for detecting the non-stationary pattern). We will add this point to the discussion.
If the editor feels it would be an added value to the manuscript, we could replicate the same analysis with data from another tide gauge station with a lower degree of non-stationarity, present the results in a dedicated figure, and include a short discussion of the results.
Improving the level of replicability (some information relevant for reproducibility is missed).
We will include more details in the methods section based on the specific reviewer’s comments.
Citation: https://doi.org/10.5194/egusphere-2022-347-AC1
-
RC2: 'Comment on egusphere-2022-347', Anonymous Referee #2, 08 Jul 2022
In this submission, authors study the value of nonstationary analysis of coastal flooding under sea level rise. They have fitted various extreme value models (e.g. GEV, GPD, PP and JPM) to a long-term record of observational water level record at Venice, Italy and quantified the difference in estimated flood risk with and without consideration of nonstationarity. The idea is very interesting and timely, the research is well designed, and the manuscript is well written. I have no significant comment, except a few very minor suggestions below:
- "extreme event theory" has been used in multiple occasions (i.e. Page 1, lines 8 and 25); I've ween usually called "extereme value theory" in the literature. Are the authors referring to a different concept/theory?
- Some recent studies on nonstationary extreme value analysis must be cited and their contribution be acknowledged
1. Cheng et al., (2014),Non-stationary extreme value analysis in a changing climate, Climatic Change volume 127, pages353–369, doi:10.1007/s10584-014-1254-5.
2. Ragno E., AghaKouchak A., Cheng L., Sadegh, M., 2019, A Generalized Framework for Process-informed Nonstationary Extreme Value Analysis, Advances in Water Resources, 130, 270-282, doi: 10.1016/j.advwatres.2019.06.007
Well done!
Citation: https://doi.org/10.5194/egusphere-2022-347-RC2 -
AC2: 'Reply on RC2', Damiano Baldan, 13 Jul 2022
Thank you for the overall appreciation of our work. We will correct the terminology as suggested and we will include the suggested references in the revised submission.
Citation: https://doi.org/10.5194/egusphere-2022-347-AC2
-
AC2: 'Reply on RC2', Damiano Baldan, 13 Jul 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-347', Anonymous Referee #1, 30 Jun 2022
The manuscript compares the return levels obtained from four different extreme value analyses accounting for non-stationarity in the Punta Della Salute tide gauge, Venice. The extreme value distributions include: (i) generalized extreme value (GEV) applied to a block maxima sampling (BM); (ii) generalized Pareto distribution (GPD) and (iii) a point process (PP) method, both applied to peak over a threshold (POT); and (iv) a joint probability method (JPM). In addition, the authors tested the implications of using three different detrending techniques, including (i) removing the annual mean sea level from the time series (MSL); (ii) removing the last 19 years’ mean sea level (MSL_L); and (iii) not detrending the data before fitting the distributions.
I find the topic important for coastal flood risk assessment as traditional designs have been based on analyses using direct methods that ignore the non-stationarities, which can lead to an underestimation of the risk, as found in previous studies. In addition, having a comprehensive analysis of the different non-stationary extreme value methodologies would help on the way to obtaining a more standardized analysis, facilitating the comparison of the results between the studies. Thus, the results of the manuscript are relevant for the scientific community and coastal risk stakeholders after improving the work in some aspects, mainly related to
- the level of comprehensiveness of the analysis: it will be interesting for the scientific community to include one of the most utilized indirect methods: the revised joint probability method. In doing so, the authors will also reduce the existing overlapping with previous studies,
- the level of applicability to other study areas by including more tide gauge records in the analysis, and
- the level of replicability (some information relevant for reproducibility is missed).
-
AC1: 'Reply on RC1', Damiano Baldan, 13 Jul 2022
Thank you for the constructive comments. In the revised submission we will perform some additional analyses and edit some sections of the manuscript to meet the comments from referee #1. See below a more detailed overview:
Reviewer’s comment
Authors' reply
Improving the work in the level of comprehensiveness of the analysis: it will be interesting for the scientific community to include one of the most utilized indirect methods: the revised joint probability method. In doing so, the authors will also reduce the existing overlapping with previous studies
We will include the revised joint probability method to the list of tested methods.
Improving the work in the level of applicability to other study areas by including more tide gauge records in the analysis
We acknowledge the sea level data from the case study in Venice might not be representative of other geographic locations given their high degree of non-stationaries (which in turn was useful for detecting the non-stationary pattern). We will add this point to the discussion.
If the editor feels it would be an added value to the manuscript, we could replicate the same analysis with data from another tide gauge station with a lower degree of non-stationarity, present the results in a dedicated figure, and include a short discussion of the results.
Improving the level of replicability (some information relevant for reproducibility is missed).
We will include more details in the methods section based on the specific reviewer’s comments.
Citation: https://doi.org/10.5194/egusphere-2022-347-AC1
-
RC2: 'Comment on egusphere-2022-347', Anonymous Referee #2, 08 Jul 2022
In this submission, authors study the value of nonstationary analysis of coastal flooding under sea level rise. They have fitted various extreme value models (e.g. GEV, GPD, PP and JPM) to a long-term record of observational water level record at Venice, Italy and quantified the difference in estimated flood risk with and without consideration of nonstationarity. The idea is very interesting and timely, the research is well designed, and the manuscript is well written. I have no significant comment, except a few very minor suggestions below:
- "extreme event theory" has been used in multiple occasions (i.e. Page 1, lines 8 and 25); I've ween usually called "extereme value theory" in the literature. Are the authors referring to a different concept/theory?
- Some recent studies on nonstationary extreme value analysis must be cited and their contribution be acknowledged
1. Cheng et al., (2014),Non-stationary extreme value analysis in a changing climate, Climatic Change volume 127, pages353–369, doi:10.1007/s10584-014-1254-5.
2. Ragno E., AghaKouchak A., Cheng L., Sadegh, M., 2019, A Generalized Framework for Process-informed Nonstationary Extreme Value Analysis, Advances in Water Resources, 130, 270-282, doi: 10.1016/j.advwatres.2019.06.007
Well done!
Citation: https://doi.org/10.5194/egusphere-2022-347-RC2 -
AC2: 'Reply on RC2', Damiano Baldan, 13 Jul 2022
Thank you for the overall appreciation of our work. We will correct the terminology as suggested and we will include the suggested references in the revised submission.
Citation: https://doi.org/10.5194/egusphere-2022-347-AC2
-
AC2: 'Reply on RC2', Damiano Baldan, 13 Jul 2022
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 304 | 80 | 6 | 390 | 21 | 4 | 2 |
- HTML: 304
- PDF: 80
- XML: 6
- Total: 390
- Supplement: 21
- BibTeX: 4
- EndNote: 2
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| China | 1 | 135 | 39 |
| United States of America | 2 | 72 | 20 |
| Italy | 3 | 29 | 8 |
| Germany | 4 | 18 | 5 |
| France | 5 | 12 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 135
Damiano Baldan
Elisa Coraci
Franco Crosato
Maurizio Ferla
Andrea Bonometto
Sara Morucci
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1317 KB) - Metadata XML
-
Supplement
(75 KB) - BibTeX
- EndNote
- Final revised paper