[Coastal flood hazard for Lecce, Italy, from breaches in the dunes] - paper

Coastal flood hazard for Lecce, Italy, from breaches in the dunes Gianandrea Mannarini a,*, Mario Leonardo Salinas a, Giorgia Verri a, Vladimir Santos Da Costa a, Riccardo Barzaghi band Daniela Carrion b a CMCC Foundation –Euro-Mediterranean Center on Climate Change, Italy b Dipartimento di Ingegneria Civile e Ambientale, Politecnico di Milano, Milano, Italy *Corresponding author. E-mail: [email protected] GM, 0000-0001-9205-7765; MLS, 0000-0002-4045-4790; GV, 0000-0002-7561-618X; VSDC, 0000-0001-7704-1861; RB, 0000-0002-2847-3841; DC, 0000-0003-1262-9394 ABSTRACT The territory of the municipality of Lecce encompasses a long coastal strip along the Adriatic Sea, which, to date, lacks high-resolution flood hazard mapping. We used a 1-m-resolution digital terrain model (DTM), as well as extreme sea level (ESL) projections from the literature, including storm surge, wave setup, tide, and mean sea level rise. A geodetic transformation was used to refer ESL to the DTM vertical datum. The outcome was used in a bathtub model, validated via in situ photos and satellite observations. The modest elevation and gentle slope of the terrain lead to a potential long-range flooding. The northern sectors, including Spiaggiabella, Torre Chianca, and a specific site in Frigole, are identified as the areas most susceptible to coastal flooding. Gaps and damage to the dune belt, specific roads, and reclamation canals can function as ingress points for seawater during flooding events. While the results are affected by several uncertainties, we find an appreciable coastal hazard for several urban and agricultural settlements in Lecce. The study underscores the importance of territorial monitoring in shaping effective, context-specific adaptation strategies for coastal resilience along the Lecce shoreline. Key words: bathtub, coastal, dunes, flooding, geoid, sea level rise HIGHLIGHTS •First hazard mapping for coastal flooding along the coastal strip of Lecce. •Based on a high-resolution terrain model and extreme sea level projections. •Damage to the dune belt as entry points for flooding. •The floodable territory has been characterized by land cover type. •Preliminary validation via in situ photos and satellite imagery. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/). © 2025 The Authors Journal of Water and Climate Change Vol 00 No 0, 1 doi: 10.2166/wcc.2025.764 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 GRAPHICAL ABSTRACT GLOSSARY AdriaClim ‘Climate change information, monitoring and management tools for adaptation strategies in Adriatic coastal areas’, a project of the 2014–2020 Interreg V-A Italy-Croatia CBC Programme AdriaClimPlus follow-up of the AdriaClim project BTM bathtub model CLC Corine land cover CORDEX Coordinated Regional Climate Downscaling Experiment DTM digital terrain model ESL extreme sea level FEMA Federal Emergency Management Agency FES finite element solution HDM hydrodynamic model MedCORDEX contribution to CORDEX for the Mediterranean region MSL mean sea level NDWI Normalized Difference Water Index POI places of interest (a set of landmarks from the Lecce PUG’s draft) PUG Piano Urbanistico Generale (Urban Master Plan) qGIS formerly ‘quantum GIS’, a geographic information system RCP representative concentration pathway RMSE root mean square error S2GLC Sentinel-2 global land cover SP133 Strada Provinciale (provincial road) no. 133 1. INTRODUCTION Coastal urban communities face rising risks exacerbated by climate change. Rapid urbanization, pressure from tourism and maritime transport, pollution, coastal erosion, sea level rise, storm surges, and saltwater intrusion are stressors affecting both the coastal environment and human activities (Davenport & Davenport 2006). Several studies have underscored the complex interplay between sea level variability and its cascading effects on coastal communities, emphasizing the importance of dataJournal of Water and Climate Change Vol 00 No 0, 2 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 driven analyses in understanding and mitigating coastal flooding risks. In Venice, Ferrarin et al. (2022), from the analysis of over 140 years of in situ observations, found a significant increase in nuisance flooding, with tide and long-term fluctuations, attributed to both subsidence and sea level rise, becoming dominant factors. Li et al. (2021) reported a sharp rise in tidal flooding events along the United States coastline since 1950, driven not only by tidal changes but also by rising mean sea levels. Indeed, an enhancement of coastal flood risk is expected because of sea level rise (Hinkel & Klein 2009). The Japan Meteorological Agency 1 offers an operational mapping service for various types of geophysical risks, including floods. In the United States, the Federal Emergency Management Agency (FEMA) publishes static maps 2 of flood risks. In the United Kingdom, the ‘Check for flooding’ 3 short-term, operational service provides both alerts (‘flood is possible’) and warnings (‘flood is expected’), including real-time observations. The European Union floods directive aims to manage the flood risk in Europe, including both rivers and coastal risk, by updating flood hazard and risk maps and taking measures to reduce risk. 4 While the utility of operational alert services is undeniable for saving lives and reducing immediate loss and damage during emergencies, properly addressing the sea level rise driven by climate change necessitates both national adaptation plans (Perini et al. 2016) and long-term territorial planning (IPCC 2023). For example, Narendr et al. (2022) employed an analytical hierarchical process to delineate zones of vulnerability and risk indices for Sagar Island, aimed at informing governmental decision-making and enhancing resilience. In Italy, territorial planning is articulated at various administrative levels, ranging from the regional to the urban one. Urban planning is conducted through a master plan, known in the Apulia Region as Piano Urbanistico Generale (PUG). The municipality of Lecce, in Apulia, includes a 20-km-long coastal strip on the Adriatic Sea (Figure 1). This territory is characterized by a diversity of natural environments and human presence (Figure 2). Sandy beaches are bordered by dunes and back-dune wetlands. Swamp lands reclaimed in the early 1900s were rapidly and irregularly urbanized in the 1970s and 1980s, exploiting the infrastructure and land grid created by reclamation works (Curci et al. 2022, cf. places of interest (POI) #20, #30, #36 and #37 in Supplementary Material file poi.xlsx). This exacerbated the existing geological hazards from both sinkholes and flooding (Margiotta & Parise 2019). Climate change, with related mean sea level (MSL) rise, may represent an additional threat, with an enhancement of existing hazards (Hinkel & Klein 2009). Adaptation strategies identified in the scientific literature span a wide array of options, primarily shaped by governance archetypes (Bongarts Lebbe et al. 2021;Galluccio et al. 2024). These range from coastal protection to adaptation and from infrastructure-based to integrated approaches. Hard interventions –such as concrete seawalls –were traditionally favoured, yet they often generate significant ecological impacts and typically fail to offer long-term protection. In contrast, soft adaptations like beach nourishment enhance coastal resilience to extreme sea level (ESL) events, though they rely heavily on the availability of large volumes of suitable sand. Accommodation measures may involve building-specific protections, the development of drainage infrastructure, as well as contingency planning and early warning systems. Managed retreat, involving the relocation of buildings or entire communities inland, sometimes represents the safest option. Despite social, psychological, and financial barriers to acceptance, it is being implemented in pilot projects. In Ravenna, where a vast area is exposed to coastal flooding, two alternative scenarios have been envisioned: one that seeks to preserve the current territorial configuration at all costs, despite rising sea levels, and another that allows the anthropogenic landscape and natural ecosystems to evolve more freely (Lobosco & Mencarini 2020). Adapting a territory to sea level rise is a key aspect of long-term urban planning, closely tied to the aims of a PUG. Longterm master plans for Lecce were adopted in 1934, 1960, and 1989. Throughout the last three City Council terms, various efforts were undertaken to revise the plan. A comprehensive version of the PUG was finalized in 2023, receiving a favourable assessment from the Basin Authority in 2024. However, following the most recent elections, the newly appointed Municipal Administration opted to withdraw the PUG to carry out further revisions. The scientific work discussed in this manuscript originates from the public consultation phase conducted during the preceding Council term. To our knowledge, no similar scientific investigations had been carried out in this specific area before. This lack of information could hinder efforts to develop a comprehensive, long-term strategy for safe and sustainable urban development along the coastal strip of Lecce. 1 https://www.jma.go.jp/jma/indexe.html. 2 https://www.fema.gov/flood-maps. 3 https://check-for-flooding.service.gov.uk/. 4 https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:32007L0060. Journal of Water and Climate Change Vol 00 No 0, 3 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 To tackle this gap, we conducted a thorough review of the existing data and methodologies, as detailed in Section 2. Based on this review, we identified methods that could be rapidly implemented to provide an initial estimation, as described in Section 3. The significance, constraints, and outlook of our findings are discussed in Section 4, while our conclusions are drawn in Section 5. 2. METHODS To estimate flooded areas in coastal hazard mapping, two key components are essential: accurate sea level data and a robust methodology for flood assessment. Both components are elaborated upon in the following. 2.1. Sea level information When characterizing the sea level of a specific coastal region, two primary aspects hold significance: the long-term, time-averaged sea level, commonly termed as MSL, and the episodic, ESL, resulting from high tides, storm surges, and MSL variations. Below, we outline historical observations for MSL and projections pertinent to both MSL and ESL. The projected values for the domain of interest are then provided in Supplementary Table S2. Figure 1 |Case study area. (a) The Apulia region (dark grey) in the context of Europe. (b) Close-up with locations of some tide gauges of the Rete Mareografica Nazionale marked by black circles and cells of the gridded products (IPCC, Copernicus, and AdriaClim) used in this work. (c) The territory of the municipality of Lecce, with a highlight of the DTM coverage for its coastal part and the DIVA point (black triangle) used for extracting ESL information and several relevant locations (empty circles) in the coastal territory of Lecce. Journal of Water and Climate Change Vol 00 No 0, 4 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 2.1.1. Mean sea level In this work, historical sea level data for the coast of Lecce could be obtained from both tide gauges and satellite altimetry. Tide gauges are part of the Rete Mareografica Nazionale, overseen by ISPRA (Istituto Superiore per la Protezione e la Ricerca Ambientale). They include those located in Bari (145 km away from Lecce) and Otranto (50 km away, as shown in Figure 1(b)). Annual-mean data from these observing stations was obtained from the Permanent Service for Mean Sea Level (PSMSL), which processed the raw data by ISPRA via a Demerliac filter and a weighted average protocol (Holgate et al. 2013). Additionally, satellite altimetry gridded data from the Copernicus Marine Service was used in this study. We processed the absolute dynamic topography (ADT variable) of the SEALEVEL_EUR_PHY_L4_MY_008_068 product. This is a daily product with 0.125-degree spatial resolution, which we aggregated on an annual basis. For reducing coastal land contamination of the remotely sensed sea level, the field was averaged on around 50 offshore pixels from the area shown in Figure 1(b). However, we found that the resulting signal does not change much with the extent of the selected region. The MSL global projections were extracted from datasets of the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR5) by Church et al. (2013). This gridded product features a one-degree resolution, and the utilized grid point is visualized in Figure 1(b). The AR5 information was downloaded at annual resolution from CEN-UH, 5 for both the RCP4.5 and 8.5 emission scenarios. The current observed rate of global MSL rise is estimated to be between 3.2 and 4.2 mm/year (IPCC 2023). However, coastal sea level dynamics differ from those of the open ocean due to factors such as bathymetry, coastal geometry, and river discharge (Piecuch et al. 2018;Verri et al. 2024). Therefore, wherever feasible, global MSL projections should be compared against regionally resolved products. An MSL projection for the coastal region of Lecce is available as an outcome of the AdriaClim project (Verri et al. 2024). AdriaClim is a limited area downscaling from the Med-CORDEX initiative. It includes six model components: atmosphere, land surface, hydrology, marine hydrodynamics, waves, and biochemistry. Rivers ending in the Adriatic Sea and with a catchment area of at least 500 km 2 were prognostically solved, and their discharges entered the marine hydrodynamics model, while tides and wave feedback were neglected. The hydrodynamics products were provided for the whole Adriatic Sea on Figure 2 |Overview map with indication of the POI of PUG’s draft (photos for the POIs are linked in the related Supplementary Material file). The colour patches represent the CLC land cover classes and the black lines represent the network of reclamation canals. Local toponyms are represented by horizontally typed names in italics. Each of the eight white-edged sectors, with names angled at 45°, represents an area of 2km 2 or 200 hectares (ha). 5 https://www.cen.uni-hamburg.de. Journal of Water and Climate Change Vol 00 No 0, 5 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 a grid with a 2 km horizontal resolution. The AdriaClim climate simulation started in 1992 and ended in 2050, making use of the RCP8.5 emission scenario. Furthermore, the MSL equation by Pinardi et al. (2014) was used to estimate the total sea level. In particular, the sea level from AdriaClim was post-processed by adding both a thermosteric and a halosteric contribution accounting for both temperature and salinity anomalies with respect to the reference year (1992). In this work, the resulting total sea level field was averaged on 24 by 24 elemental cells, covering an area comparable to the one used for satellite altimetry, cf. Figure 1(b). The resulting, annual averaged field was considered to represent the regional MSL relevant for Lecce. Its comparison with the global projections is provided in Section 2.1.3. 2.1.2. Extreme sea level We used ESL projections from Kirezci et al. (2020). They refer to episodic sea level extremes due to a synchronous occurrence of high tide, storm surge, and wave setup. The authors made use of astronomical tides, reanalysis of storm surge, and modelling of wave setup. Long-term stationarity in both waves and storm surge was assumed. The historical extremes were validated by the authors versus observations from the GESLA-2 network, revealing an average root mean square error (RMSE) of about 20 cm. To derive the historical ESL, Kirezci et al. (2020) performed an extreme value analysis. They used a peak-over-threshold approach, fitting with a generalized Pareto distribution and a 98th percentile threshold. For future ESL projections, they incorporated MSL values from the IPCC AR5 projections into their historical estimates. We selected the resulting ESL information at the Dynamic and Interactive Vulnerability Assessment (or DIVA, Hinkel & Klein 2009) point closest to the region of our interest, shown in Figure 1(c). The data is available for the years 2010, 2050, and 2100, and we interpolated them to an annual resolution using quadratic regression. A centennial return period was used, which equates to a 1% annual probability of occurrence. According to FEMA’s classification, this level is deemed ‘high-risk’, signifying at least a one-in-four chance of flooding during a 30-year mortgage period. Since Lecce’s past masterplans remained in place for a similar duration, there is a high probability of experiencing centennial flood events during the time of the next one. To estimate the confidence interval for future ESL, we used the RMSE found by Kirezci et al. (2020) in their validation phase. As in their dataset the wave climate was assumed to remain constant, the primary driver of long-term changes was the rise in MSL. 2.1.3. Comparing sea levels Figure 3 displays the time series of annual-mean, sea level observations, and projections utilized in this study. The ESL signals are referenced to the vertical datum of the ITALGEO-05 geoid (see Section 2.2.2); the MSL to the 1986–2005 baseline period of the IPCC AR5 projections. As observed in Martínez-Asensio et al. (2014), both tide gauge and satellite MSL observations, along with AdriaClim projections, exhibit significant (order of a few centimetres) interannual variability. Moving on to the Figure 3 |ESL values from Kirezci et al. (2020) for the RCP4.5 scenario as the dashed black line with a grey band relative to their RMSE; 5– 95th quantile range for the IPCC AR5 MSL projections as coloured solid lines (light blue: RCP4.5; light red: RCP8.5); annual-mean MSL from AdriaClim (solid black line), from satellite altimetry (dashed red line), and from tide gauges in Bari and Otranto (yellow and blue markers, respectively). See the text in Section 2.1.3 for the vertical datums. Journal of Water and Climate Change Vol 00 No 0, 6 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 MSL projections, both their central values and the 5th–95th percentile plumes are considered. They indicate that, depending on the greenhouse gas emission scenario (RCP4.5 or RCP8.5), the study area in Lecce is projected to experience an additional sea level rise of approximately 17–21 cm by 2060, relative to 2020 levels (cf. supplementa.xlsx in the Supplementary Material). This increase is expected to proportionally intensify the magnitude of ESL events. AdriaClim exhibits interannual variability with an amplitude comparable to that observed in satellite altimetry but shows limited temporal coherence with observational datasets. This is expected, as a free-run oceanographic model was employed, meaning that it was not nudged towards observed states. During the 2010–2050 period, the AdriaClim data suggests an average MSL growth rate at Lecce of approximately 3.5 mm/year. This aligns closely with the IPCC AR5 projections, particularly those associated with the RCP4.5 scenario. However, the regional projection of AdriaClim was derived for RCP8.5. The relatively lower trend observed in AdriaClim compared with the global projections for RCP8.5 may be explained by variability in the freshwater budget (Verri et al. 2024). Given this observation and recognizing that the ESL by Kirezci et al. (2020) incorporates the IPCC MSL projection, we have opted to use their low-emission (RCP4.5) projection. The AdriaClimPlus 6 project will tackle the challenge of providing more reliable regional-to-local sea level projections for various climate scenarios. As shown in Figure 3, the ESL at Lecce exceeds the MSL by approximately 60 cm. According to the FES2014 7 model output, the astronomical tidal range at Lecce is around 26 cm. Therefore, storm surge and wave setup account for most of the episodic sea level rise observed during extreme events. Furthermore, the ESL MSL value remains constant throughout the projection period, as Kirezci et al. (2020) assumed a stationary wave climate, with no anticipated future changes in wave conditions. 2.2. Coastal flooding and modelling The relevant literature for coastal flooding and modelling is reviewed in Section 2.2.1, before the method used for the estimation of coastal flooding in Lecce is presented in Section 2.2.2. 2.2.1. Literature Numerical methods for coastal flooding include both hydrodynamic models (HDMs) and bathtub models (BTMs). HDMs are numerical solvers of the shallow water equations, which represent the conservation of mass and momentum in two or three dimensions. In the past three decades, numerous HDMs have been developed or updated, including: TELEMAC2D, solving the De Saint Venant equations of free surface flow on an unstructured grid (Galland et al. 1991); MIKE 21, a suite simulating hydrodynamics, advection–dispersion, short waves, sediment transport, water quality, eutrophication, and heavy metals (Warren & Bach 1992); XBeach, which includes a non-stationary wave driver for wave-generated surf and swash motion (Roelvink et al. 2009); LISFLOOD-FP, coding new equations from 1D shallow water theory, with a sensitivity to soil friction (Bates et al. 2010); FUNWAVE, solving a set of fully non-linear Boussinesq equations and which accurately predicted wave runup against a suite of benchmark test data (Shi et al. 2012); Delft3D, a multi-dimensional hydrodynamic and transport simulation programme which calculates non-steady flow and transport phenomena resulting from tidal and meteorological forcing; and WRF-Hydro-CUFA, based on the hydrologic model WRF-Hydro and including a hydraulic flow solver, to consider flood controls of stormwater drainage (Son et al. 2023). Instead, BTMs are a class of modelling approaches that compute the intersection between a water surface corresponding to the sea elevation and a three-dimensional terrain model. BTMs do not explicitly incorporate hydrodynamic effects and are therefore considered stationary or static models. They provide a simpler method than HDMs to estimate flood-prone areas and can be computed even within a geographic information system (GIS), at the cost of neglecting water mass conservation and momentum dissipation due to soil friction. There exists a substantial body of literature discussing the application of BTMs in coastal flood estimation. The paper by Breilh et al. (2013) presents a comparison of three flooding methods during the well-documented flooding caused by the Xynthia storm in France. They used BTMs without or with spatially varying sea elevations along the shore and found good predictions for distances from the coastline of the order of hundreds of metres and overestimations at distances of the order of 10 km. Spatial variations of maximum sea level elevation had a limited impact on the estimation of the flooding. In Seenath et al. (2016), a BTM was compared with two HDMs (LISFLOOD-FP and TELEMAC-2D) in a tropical sandy coast 6 https://www.italy-croatia.eu/web/adriaclimplus. 7 https://www.aviso.altimetry.fr/en/data/products/auxiliary-products/global-tide-fes.htm. Journal of Water and Climate Change Vol 00 No 0, 7 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 with coral reefs, finding that all three models were consistent (less than 5% difference in flood predictions). However, when hydraulic connectivity was not enforced, the BTM overestimated flooding. The pan-European study by Vousdoukas et al. (2016) considered a BTM, LISFLOOD-FP, and a flood intensity index across all coasts of Europe, using a coarse (90 m) terrain model. They found that a BTM with connectivity can overestimate flood extents by 56% compared with LISFLOOD-FP, especially for coasts with low-lying and mild-slope terrain. This approach operates under the assumption of a constant hydraulic head loss, irrespective of terrain slope and roughness. An impressive observational dataset for a specificflood event for a 20 km coastline along an open estuary in eastern Canada was used by Didier et al. (2019) to compare the performance of a BTM versus the XBeach model (Roelvink et al. 2009). Both models correctly predicted flooded areas, with BTM overpredicting their extent by 36% and the floodwater depth by 0.5 m. The floods that occurred in North Carolina during Hurricane Florence were computed through three variants of BTM in Rucker et al. (2021). When compared with a high-resolution, full-physics model, they found that a basic BTM method overpredicts water level extents, while one corrected with a parametrization for head loss (based on Manning’s equation) performed better, with close matches to the true flooded areas. In Hong et al. (2024), a BTM not accounting for shoreline connectivity and an HDM were compared for a flood event in the Great Lakes. The flooded areas were found to be consistent, with BTM overestimating flooding in not-connected areas. Accordingly, a prevailing consensus in the literature highlights the critical importance of ensuring connectivity for the successful implementation of a BTM. The following subsection outlines the approach adopted to incorporate this principle into our analysis. 2.2.2. Implementation Given the limited availability of high-resolution input data required for HDMs –such as high-resolution topography, land cover, and soil characteristics –we employed a BTM approach. Despite its simplified nature, this method enabled the efficient identification of flood-prone areas across a substantial stretch of coastline, spanning approximately 15 km and covering an area of 10 km 2 . This extensive spatial coverage offers valuable input for preliminary flood risk assessment and supports strategic planning efforts along the Lecce coast. To represent the land topography of the coastal strip of Lecce, a digital terrain model (DTM) is used in this work. The DTM is based on data collected from an airborne LiDAR sensor flown in summer 2015. The horizontal resolution was 1 m and the vertical accuracy 30 cm. Besides the DTM, orthorectified aerial photos were used. Land elevations were given in terms of orthometric heights with respect to the ITALGEO-05 geoid (Barzaghi et al. 2007). This surface is termed IT-05 in the following. Additionally, land cover data was employed in this work to provide qualitative insights into the nature of areas susceptible to future flooding. Distinguishing between urban, agricultural, and wetland areas can be pivotal for decision-makers aiming to devise effective adaptation strategies. Notably, Leal-Alves et al. (2020) specifically differentiated urban areas from the overall inundated region. In perspective, land cover data can significantly contribute to improving flood modelling methodologies, especially by aiding in the estimation of terrain roughness (Vousdoukas et al. 2016;Rucker et al. 2021). The land cover classification utilized in the present work is inherited from the original provider, such as Corine or Sentinel-2 (cf. Supplementary Material Section 2). This implied that, for example, the term ‘salt marsh’may encompass not only brackish water lagoons but also inland freshwater swamps (e.g. ‘Rauccio swamp’in Figure 2) located further from the direct sea influence. Other differences between the land cover datasets are discussed in Supplementary Material Section 2. The ESL values used in this work were referenced to the Earth Geopotential Model 1996 (EGM-96) geoid by Kirezci et al. (2020) by adding a mean dynamic topography (MDT). The used procedure does not mention the possibility that EGM-96 and MDT refer to different ellipsoids. Nevertheless, we assumed that the provided ESL values consistently represent orthometric heights on the EGM-96 geoid. Instead, the topographic information for our region of interest is provided in terms of orthometric heights on the IT-05 geoid. Then, to make ESL and DTM values comparable, a procedure to transform across systems of reference is needed and is documented in Supplementary Material Section 1. In essence, the difference in undulation between the geoids of the two datasets (ESL and DTM) was linearized. Then, the ellipsoidal heights of ESL were referred to the IT-05 system of coordinates. In the coastal strip of Lecce, this local variation resulted in an upward adjustment of the ESL by approximately 2.7 cm. This constant offset was added to all ESL values provided by Kirezci et al. (2020). The orthometric height of ESL, once referred to the IT-05 system, can be directly compared with the terrain elevation from the DTM. Given the sea level ηat a specific year tand the terrain elevation z, we conducted the following geospatial data processing via the qGIS open-source software: Journal of Water and Climate Change Vol 00 No 0, 8 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 (a) computation of isolines via the command: gdal_countour -flη(t); (b) manual pruning of isolines not topologically connected to the shore, as shown in Figure 4; (c) creation of a raster field of the flood depth η(t) –z; (d) use of the isolines from step (b) for clipping the raster of step (c); (e) overlay of both aerial photos and road graph using the OSMnx package. To demonstrate that this procedure remains both reproducible and objective, despite involving some manual intervention, consider the following. Once qGIS computes the isolines, the operator selects the one adjacent to the 0 m contour –typically aligned with the current shoreline and open-ended. Closed isolines, in contrast, indicate isolated areas of lower elevation, disconnected from the shoreline, and thus not susceptible to seawater flooding. Related isolines are manually pruned. Fundamentally, the outcome of this procedure is not driven by significant subjective decisions. The pivotal step –selecting the isoline that borders the coastline –is inherently unambiguous. Although some minor variation may arise in how operators handle some small isolines distant from the shore, the overall methodology remains consistent and replicable, minimizing subjectivity in the results. As a final methodological note, although this article primarily focuses on the hazard dimension, we include flood depth information (η–z) as a widely used indicator for estimating direct tangible losses (Chen et al. 2024) and, thus, one key component in a vulnerability assessment. 2.3. Validation To establish a ground truth for coastal flooding of the near-present period, both in situ and remote observations are, to some extent, viable options and are described in the following. The use of HDMs is out of the scope of this work. Figure 4 |Isoline pruning methodology: just flooded areas connected to the shoreline (yellow isolines) are retained, while the remaining ones (magenta) are pruned. The black solid line represents the shoreline for a sea level of 0.00 metres. Journal of Water and Climate Change Vol 00 No 0, 9 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 Davenport, J. & Davenport, J. L. (2006) The impact of tourism and personal leisure transport on coastal environments: a review,Estuarine, Coastal and Shelf Science,67 (1–2), 280–292. Didier, D., Baudry, J., Bernatchez, P., Dumont, D., Sadegh, M., Bismuth, E., Bandet, M., Dugas, S. & Sévigny, C. (2019) Multihazard simulation for coastal flood mapping: bathtub versus numerical modelling in an open estuary, eastern Canada,Journal of Flood Risk Management,12, e12505. doi: 10.1111/jfr3.12505. Ferrarin, C., Lionello, P., Orlic, M., Raicich, F. & Salvadori, G. (2022) Venice as a paradigm of coastal flooding under multiple compound drivers,Scientific Reports,12, 5754. doi: 10.1038/s41598-022-09652-5. Galland, J.-C., Goutal, N. & Hervouet, J.-M. (1991) TELEMAC: a new numerical model for solving shallow water equations,Advances in Water Resources,14 (3), 138–148. doi: 10.1016/0309-1708(91)90006-A. Galluccio, G., Hinkel, J., Fiorini Beckhauser, E., Bisaro, A., Biancardi Aleu, R., Campostrini, P., Casas, M. F., Espin, O. & Vafeidis, A. T. (2024) Sea level rise in Europe: adaptation measures and decision-making principles. In: van den Hurk, B., Pinardi, N., Kiefer, T., Larkin, K., Manderscheid, P. & Richter, K. (eds) Sea Level Rise in Europe: 1st Assessment Report of the Knowledge Hub on Sea Level Rise (SLRE1). State of the Planet, Copernicus Publications, ch. 6. doi: 10.5194/sp-3-slre1-6-2024. Hinkel, J. & Klein, R. J. T. (2009) Integrating knowledge to assess coastal vulnerability to sea-level rise: the development of the DIVA tool, Global Environmental Change,19 (3), 384–395. doi: 10.1016/j.gloenvcha.2009.03.002. Holgate, S. J., Matthews, A., Woodworth, P. L., Rickards, L. J., Tamisiea, M. E., Bradshaw, E., Foden, P. R., Gordon, K. M., Jevrejeva, S. & Pugh, J. (2013) New data systems and products at the Permanent Service for Mean Sea Level,Journal of Coastal Research,29 (3), 493– 504. doi: 10.2112/JCOASTRES-D-12-00175.1. Hong, Y., Kessler, J., Titze, D., Yang, Q., Shen, X. & Anderson, E. J. (2024) Towards efficient coastal flood modeling: a comparative assessment of bathtub, extended hydrodynamic, and total water level approaches,Ocean Dynamics,74 (5), 391–405. doi: 10.1007/ s10236-024-01610-1. IPCC (2023) Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. (Core Writing Team, Lee, H. & Romero, J. (eds)) Geneva, Switzerland: IPCC, pp. 35–115. Kirezci, E., Young, I. R., Ranasinghe, R., Muis, S., Nicholls, R. J., Lincke, D. & Hinkel, J. (2020) Projections of global-scale extreme sea levels and resulting episodic coastal flooding over the 21st century,Scientific Reports,10, 11629. doi: 10.1038/s41598-020-67736-6. Leal-Alves, D. C., Weschenfelder, J., Albuquerque, M. d. G., de Almeida Espinoza, J. M., Ferreira-Cravo, M. & Almeida, L. P. M. (2020) Digital elevation model generation using UAV-SfM photogrammetry techniques to map sea-level rise scenarios at Cassino Beach, Brazil,SN Applied Sciences,2, 2181. doi: 10.1007/s42452-020-03936-z. Li, S., Wahl, T., Talke, S. A., Jay, D. A., Orton, P. M., Liang, X., Wang, G. & Liu, L. (2021) Evolving tides aggravate nuisance flooding along the US coastline,Science Advances,7, eabe2412. doi: 10.1126/sciadv.abe2412. Lobosco, G. & Mencarini, V. (2020) Landscape and climate change: a resilient strategy for the adaptation plan of the Ravenna area in Italy, Convergences –Journal of Research and Arts Education,13 (26), 103–117. doi: 10.53681/c1514225187514391s.26.39. Margiotta, S. & Parise, M. (2019) Hydraulic and geomorphological hazards at wetland geosites along the eastern coast of Salento (SE Italy), Geoheritage,11, 1655–1666. doi: 10.1007/s12371-019-00363-4. Martínez-Asensio, A., Marcos, M., Tsimplis, M. N., Gomis, D., Josey, S. & Jordà, G. (2014) Impact of the atmospheric climate modes on Mediterranean sea level variability,Global and Planetary Change,118,1–15. doi: 10.1016/j.gloplacha.2014.03.007. McFeeters, S. K. (1996) The use of the Normalized Difference Water Index (NDWI) in the delineation of open water features,International Journal of Remote Sensing,17 (7), 1425–1432. doi: 10.1080/01431169608948714. Narendr, A., Vinay, S., Aithal, B. H. & Das, S. (2022) Multi-dimensional parametric coastal flood risk assessment at a regional scale using GIS,Environment, Development and Sustainability,24 (7), 9569–9597. doi: 10.1007/s10668-021-01839-6. Perini, L., Calabrese, L., Salerno, G., Ciavola, P. & Armaroli, C. (2016) Evaluation of coastal vulnerability to flooding: comparison of two different methodologies adopted by the Emilia-Romagna region (Italy),Natural Hazards and Earth System Sciences,16 (1), 181–194. doi: 10.5194/nhess-16-181-2016. Piecuch, C. G., Bittermann, K., Kemp, A. C., Ponte, R. M., Little, C. M., Engelhart, S. E. & Lentz, S. J. (2018) River-discharge effects on United States Atlantic and Gulf coast sea-level changes,Proceedings of the National Academy of Sciences,115 (30), 7729–7734. doi: 10.1073/ pnas.1805428115. Pinardi, N., Bonaduce, A., Navarra, A., Dobricic, S. & Oddo, P. (2014) The mean sea level equation and its application to the Mediterranean Sea,Journal of Climate,27, 442–447. doi: 10.1175/JCLI-D-13-00139.1. Roelvink, D., Reniers, A., van Dongeren, A., van Thiel de Vries, J., McCall, R. & Lescinski, J. (2009) Modelling storm impacts on beaches, dunes and barrier islands,Coastal Engineering,56 (11–12), 1133–1152. doi: 10.1016/j.coastaleng.2009.08.006. Rucker, C. A., Tull, N., Dietrich, J. C., Langan, T. E., Mitasova, H., Blanton, B. O., Fleming, J. G. & Luettich, R. A. (2021) Downscaling of realtime coastal flooding predictions for decision support,Natural Hazards,107, 1341–1369. doi: 10.1007/s11069-021-04634-8. Seenath, A., Wilson, M. & Miller, K. (2016) Hydrodynamic versus GIS modelling for coastal flood vulnerability assessment: which is better for guiding coastal management? Ocean & Coastal Management,120,99–109. doi: 10.1016/j.ocecoaman.2015.11.019. Shi, F., Kirby, J. T., Harris, J. C., Geiman, J. D. & Grilli, S. T. (2012) A high-order adaptive time-stepping TVD solver for Boussinesq modeling of breaking waves and coastal inundation,Ocean Modelling,43–44,36–51. doi: 10.1016/j.ocemod.2011.12.004. Son, Y., Di Lorenzo, E. & Luo, J. (2023) WRF-Hydro-CUFA: a scalable and adaptable coastal–urban flood model based on the WRF-Hydro and SWMM models,Environmental Modelling & Software,167, 105770. doi: 10.1016/j.envsoft.2023.105770. Journal of Water and Climate Change Vol 00 No 0, 16 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025 Verri, G., Furnari, L., Gunduz, M., Senatore, A., Santos da Costa, V., De Lorenzis, A., Fedele, G., Manco, I., Mentaschi, L., Clementi, E., Coppini, G., Mercogliano, P., Mendicino, G. & Pinardi, N. (2024) Climate projections of the Adriatic Sea: role of river release,Frontiers in Climate,6, 1368413. doi: 10.3389/fclim.2024.1368413. Vousdoukas, M. I., Voukouvalas, E., Mentaschi, L., Dottori, F., Giardino, A., Bouziotas, D., Bianchi, A., Salamon, P. & Feyen, L. (2016) Developments in large-scale coastal flood hazard mapping,Natural Hazards and Earth System Sciences,16 (8), 1841–1853. doi: 10. 5194/nhess-16-1841-2016. Warren, I. R. & Bach, H. K. (1992) MIKE 21: a modelling system for estuaries, coastal waters and seas,Environmental Software,7(4), 229– 240. doi: 10.1016/0266-9838(92)90006-P. First received 14 October 2024; accepted in revised form 23 July 2025. Available online 14 August 2025 Journal of Water and Climate Change Vol 00 No 0, 17 Downloaded from http://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2025.764/1591855/jwc2025764.pdf by guest on 17 September 2025