| 02 Jul 2025
A geospatial database of coastal characteristics for erosion assessment of Europe’s coastal floodplains
Abstract. Coastal erosion and flooding are known to be linked, with erosion potentially exacerbating flood extents and risk, but analysis of the combined hazards is limited. This paper describes the CoasTER geographic database designed as a first step in integrating existing information on erosion and other relevant characteristics for Europe’s coastal floodplains to support flood assessment and climate services. The CoasTER database updates and builds on earlier erosion research and data sources. It also includes a coastal geomorphological typology which incorporates human modification in the form of hard engineering and infrastructure.
Almost 80 % (25,000 km) of shorelines associated with coastal floodplains are composed of erodible sediments, with coastal wetlands being the most prevalent geomorphological type. While accretion is the dominant historical trend for these shorelines, approximately 27 % are currently classed as eroding over the last 40 years. The majority of floodplain shorelines are associated with either developed or agricultural areas and human structures are visible along almost 8,000 km of shoreline, restricting morphodynamic response to sea level rise. If the erosive trend continues for developed areas, over 2,500 km of shoreline will require further management to maintain current protection levels and nearly 1,000 km will require new management. The CoasTER database reveals the potential magnitude of erosion-flood interactions in the future defining where mobile sediments and coastal floodplains are co-located. It demonstrates that episodic and/or long-term erosion and coastal flooding is a Europe-wide issue that deserves the attention of local to European decision-makers in order to define a coherent management strategy.
General comments :
This paper presents a new Europe-wide database combining both coastal erosion and flooding information to enable identifying the interactions between these two hazards at large spatial scales. The database builds upon previously existing databases (e.g. Eurosion (2004); Corine Land Cover (2018), etc.), and improves, simplifies, and extends them, while also adding additional information about the presence of structures and other features that constrain coastal evolution. The overall structure of the paper may be clarified by stating clearly the different elements that are considered necessary in such a database. The manuscript describes the different elements and then summarizes the overall statistics. Several details about the different data sources could be added (succinctly!) to clarify the contributions of each data set and to clearly identify the work done here. The paper is well written overall and presents an interesting new data set that may be useful for coastal managers. The specific comments listed below highlight a few questions and a few details that could be added to clarify some points throughout the manuscript.
Specific comments:
This paper presents the new database, while also relying on a contributions from previously existing databases. This work is worthy of publication, and several points can be addressed to clarify and expand the applications of this work.
Firstly, the introduction suggests the importance of quantitative analyses of erosion and flooding, and the need to generate improved databases to support efforts (e.g. in the CoCliCo project) to map current and future coastal flooding. I would suggest extending the bibliography here by citing other potentially relevant European-wide databases/studies, such as Le Gal et al (2023), Souto-Ceccon et al. (2025), among others. Many additional data sets exist at local and regional scales, so perhaps it is worthwhile (L35) to emphasize the need to produce these datasets on European scales.
The end of the introduction presents the work completed here, and it would be worthwhile to describe more clearly the database that will be presented (e.g. the major components and how this improves upon existing databases). For example, only erosion from Eurosion (2004) and the shoreline position from EEA (2017) are mentioned in L44-48. Table 1 later presents a more complete summary of the different databases used, but it would be useful to present the essential elements in such a database earlier. One approach may be to identify the necessary elements here or to suggest that the necessary elements will be identified in the background section that follows. This may help to structure the following sections of the article and to create a smooth transition to the background section. (The background section almost feels like it could integrate well in the introduction, but this is a choice to be made by the authors and/or journal concerning the overall format.) Perhaps even a diagram indicating the general elements in the database would be useful at the end of the background so that the overall themes of the database are described early on before additional details are provided. Finally, I would also suggest emphasizing the open source nature of the data (which is implicit, but it doesn’t hurt to be explicit about this!)
At the very end of the introduction, it would be worthwhile to be more specific in the statement “possible developments are also considered.” (e.g. limitations are identified in the discussion and suggestions for future improvements are made regarding certain specific aspects?)
One of the contributions of this work is improvements to the shoreline position contour defined by EEA (2017). The “cleaning” is referenced in L94: can you provide more details about what kind of cleaning was completed? what percentage (or spatial distribution or type of elements (e.g. port contours, wetlands, etc.) were improved?) This coastline was then cut into “homogeneous sections” (L99): please list the criteria that were used to achieve this. The coastline characteristics were then assigned to a zone extending approximately 100m landward of the EEA shoreline (L101): how was this value of 100m chosen? I would recommend being more specific about these aspects since they are some of the contributions of your work!
Section 3 could benefit from a more detailed presentation of the elements integrated in the database (in particular in 3.2). While the goal of this work is not to present thoroughly the previously existing databases, it would be useful to the reader to provide (succinctly) additional details (e.g brief description and estimated errors, which is an important point again in the discussion when identifying the limitations of the current shoreline data used (e.g. L323-324)) about the useful parameters that have been extracted from previous data sets. For example, by stating that the Corine Land Cover database includes 44 classifications that have been regrouped (Table A1/A2) and simplified down to XX classifications. In L115, is the visual interpretation completed using satellite images? (similar to what was indicated for structures in L141?)
In section 3.2.4, it would be useful to the reader to define more explicitly the definition of the floodplain zones, including the 2 types of floodplain connections to the coastal segments. In L143-4, the coastal segments are connected to the 1 in 100 year coastal floodplain. Although it is mentioned in Table 1, it is worthwhile repeating here that these values come from the CoCliCo flood units (Lincke & Hinkel, 2023), and to describe in a few words the method used to obtain the 1 in 100 yr levels. Are the direct connections between coastal segments and floodplains made by evaluating all floodplains that are connected to coastal zones below 2m + the 1 in 100 yr extreme event elevation? Then, L152 refers to “non-local, more indirect implications”. Please clarify what is meant by this. Does this refer to zones that are not connected using the above-mentioned cutoff?
When describing the erosion data set that is used in this database, please clarify what is meant by the data limitations that led to using 0.5 m/yr as a cutoff. You state that (following Luijendijk et al., 2018) this cutoff is used, so is this the error threshold or a given significance level?
Section 4 presents the statistics of erosion and floodplain interactions around Europe. In L224: can you clarify what is meant “However, erosion ... has more impact on individual country shorelines.”? (Is the goal to emphasize the importance identifying of erosion zones? I’m not sure I undertand the reference to “individual countries” here) Also, the term “natural” is used throughout this section, when referring to both wetlands (e.g. is the goal to distinguish between natural wetlands and wetlands that have been impacted by human interventions?) and evolution (e.g. is the goal to distinguish between natural evolution and climate change-induced evolution?). One general suggestion for this section is to refer consistently to the percentages in addition to the number of kilometers (e.g. L234, L244, L249+) since the percentages help to clarify a broader interpretation of the results. This may also be useful in Table 4 (not necessarily for each country, but perhaps for the totals?)
In L237, how is the “total floodable shoreline” defined? It seems like all shorelines are floodable during extreme events, but this depends on the definition of the shoreline.
In L268-269, the concept of commitment is introduced, stating that “Figure 7 shows that Italy has the highest commitment to maintaining open coast major port infrastructure in floodplains followed by France, Spain and the UK”. The use of the term “commitment” merits some discussion because this implies something about the political/management decisions that may not be well represented in the database if countries have existing defences but are not necessarily committed to continuing to upgrade them. This concept is addressed again in L314-315, but it seems difficult to discern commitment (implying the future) from presence (current state) with the data sets used here, but perhaps you can clarify this in the text.
When discussing the current erosion, stability, and accretion trends in L280-281, it is suggested that trends may change in the future, which merits a reference or a short explanation of how/why (even though it seems obvious!). In the following sentence, statistics are cited for the case “if erosion is universal along developed coasts”: Do you mean if a constant erosion trends is applied everywhere? I’m not sure to understand properly this sentence.
The discussion summarizes the advantages and limitations of the database, highlighting the importance of the interactions between erosion and floodplains, which is often implicitly considered but not quantitatively addressed in independent studies of coastal erosion or flooding. One important detail to emphasize in the discussion may be to identify who is the primary beneficiary/user of this database and for what kind of applications? coastal managers? Is the goal that this data can be used to make policy or management decisions? or to allow a preliminary detection of sites needing observations/detailed modeling to explore further local hazards related to combined erosion and flooding? Perhaps the reference to French et al. (2016) in the conclusion could be moved here and elaborated upon (describe briefly their case study?).
Finally, L365-377 bring up new points that are often placed more in the discussion than in the conclusions of an article, which usually focuses on summarizing what has been stated in the paper.
Technical corrections:
Several small suggestions are listed here :
L19: « European coastal floodplains »
L34: insert « spatial » with scales ?
Figure 1: explain where « Belle Henriette » is in the picture (e.g refer to the blue arrow ?) and perhaps also the additional impacts caused by the erosion in this location (increased flooding volume or speed or duration of flooding ?)
L71: « waves »
L76: remove « ; » (or replace with « : » ?)
L95: write out UK the first time it is cited?
L200: « .. »
L222: what areas do not have shoreline movement data ?
Figure 4: One suggestion is to make the total (last column) stand out better by making it wider or setting it apart, or making thin horizontal dashed lines across the plot to show when percentages are above or below each threshold (red/gray and gray/green transitions), simply to help guide the eye of the reader.
L260: « are a ... » remove « a »
L271: remove « to » ?
L301: « . » remove extra spaces
L320: new paragraph ?
L321: add comma after « analysis, »
L322: « land claim » => « reclaimed land »
L336: « coastlines »
L351: « is a developing » => remove « a »