Sea-level rise due to climate changes (SLR) combined with land subsidence, namely relative sea-level rise (RSLR), is one of the processes that is threatening most severely the coastal systems around the world. Flooding, saltwater contamination of farmlands and aquifers, coastal erosion, loss of tidal morphologies along with severe damages to infrastructures are among the major effects of RSLR.

Although at different rates, the ocean levels increased over the last decades up to a few mm yr-1 and are expected to rise even more quickly by the end of the century (e.g., IPCC, 2014). Despite this general rise, land subsidence in many coastlands worldwide has exceeded the SLR up to a factor of ten (Erkens et al., 2015).

At regional scales, SLR can be considered spatially homogeneous whereas ground dynamics is often characterized by a considerable variability. During the last decade, the growing use of satellite SAR interferometry over large areas reveled that land subsidence in low-lying coastlands often depicts uneven distribution patterns, primarily controlled by shallow sediment compaction, especially the Holocene deposits (Tosi et al., 2010; Teatini et al., 2011; Jankowski et al., 2017), and the hydro-morphological setting. These are key driving mechanisms in the susceptibility of the low-lying coastal plains to RSLR

Vulnerability analysis is a powerful method for studying the interaction of the various environmental features that determine the exposure of coastlands to RSLR (e.g., Gornitz, 1991). However, previous investigations neglected or poorly considered the variability of the coastal subsidence, generally assuming a quite homogeneous sinking trend also over large areas, or at least with a regional pattern.

Here we present a step forward in the analysis of the vulnerability to RSLR of the Venice coastal system (Fig. 1), considering the actual pattern of the coastal subsidence (e.g., Teatini et al., 2011; Strozzi et al., 2013; Tosi et al., 2016).

The analysis is conducted at the regional scale by combining high spatial resolution ground movement data obtained through multi-temporal SAR interferometry with information on the hydro-morphological setting of the coastal system and the present and projected sea-level rise trends.

The coastal vulnerability to RSLR was assessed by means of an index-based model that considers the following input dataset: high spatial resolution land subsidence; hydro-morphological setting, including the distance of emerged land from potential pathways for inundation due to RSLR, i.e. the coastline, watercourses, lagoons and wetlands; surface elevation of emerged and submerged sectors; present and projected sea-level rise.

The set-up of the SUBS, RSUBS, DEM and WD thematic layers was the first step of the data analysis. The maps developed on the 200 m regular grid are shown in Fig. 2.

The SUBS thematic layer clearly depicts the variability of the ground displacements although the SAR data are interpolated on the 200 m grid. In general, this pattern is primarily controlled by compaction of the Holocene deposits, which thickness increases from less than 1 m in the mainland to more than 20 m in the northern and southern coastal sectors (Tosi et al., 2010). The presence of buried geomorphological features, such as sandy paleo-channel and littoral ridge systems contributes in increasing the local variabilities patterns. To the author knowledge, induced land subsidence due to groundwater exploitation is limited to local sectors in the northern area.

The thematic layer has been classified based on the statistical distribution of the RSUBS, WD, and DEM values (Table 2).

The values 0 and 4 mean a low and a high contribution to vulnerability, respectively. Considering the peculiarity of the Venice area, whose territory lies at an elevation below or slightly above the mean sea-level, also a few mm yr-1 of relative sea-level rise seriously damage the coastal system. Therefore, the class with null value has been limited to RSLR less than 1 mm yr-1. In addition, since RSLR differently affects the ecosystem of transitional environments, the DEM thematic layer has been classified separately for the mainland (littoral strips included) and lagoon areas (islands included), hereinafter referred as DEMma and DEMla, respectively. Specifically, in the lagoon areas, RSLR threats the morphologies with a surface elevation within the tidal range (e.g., salt marshes) and above sea level (e.g. islands and historical centers) more than the lagoon bottom deeper than -1 m a.m.s.l. (e.g., channels). Conversely, in the mainland, the most threaten sectors are those with the lowest elevation with respect to the sea level.

The adopted classification is the same for the three scenarios, since the threat-classes are preserved even if subsidence and SLR change.

The classified thematic layers, with scores from 0 to 4 (from low to high contribution to vulnerability), are shown in Fig. 3. The colour-coding highlights the rating in accordance with the classification shown in Table 2.

The Index of Coastal Vulnerability (CVI) to RSLR was computed with a weighted linear regression for both the ongoing and future scenarios (see Eq. 2) and the values reclassified into five classes, i.e. negligible, marginal, moderate, critical, and extreme.

The maps showing the vulnerability to RSLR of the Venice coastland for the ongoing long-time series scenario, ongoing short-time series scenario and future scenario in 2050 are presented in Fig. 4.

The use of high spatial resolution land subsidence maps obtained by Satellite SAR interferometry makes it possible to take a step forward in the analysis of the vulnerability to RSLR with respect to previous works that generally considered land subsidence quite homogeneous or with a region trend.

The vulnerability analysis of Venice's coastland to RSLR presented here uses thematic layers at 200 m spatial resolution (Figs. 2 and 3).

Regarding the outcome of the vulnerability analysis (Fig. 4), the percentages of the Venice coastland in each class of the CVI under the ongoing long-time series, ongoing short-time series and future scenarios are summarized in Table 3.

Under the ongoing long-time series scenario, the areas with negligible vulnerability to RSL cover only 13.5 %. Areas with marginal, moderate and critical vulnerability are more extended, amounting to 30.5 %, 23.5 % and 22 %, respectively. About 10 % of the coastland is in an extreme vulnerability condition. Using the ongoing short-time series and future scenarios the areas with negligible class vanish and 27 %, 33 %, 23.5 % are the percentages of areas with moderate, critical and extreme vulnerability, respectively.

By 2050, under the future scenario, 60 % of the coastland will be in a critical and extreme vulnerability condition.

At the coastland scale, the high variability of vulnerability to RSLR reflects the variability of land subsidence, which superposes that due to the hydro-morphology setting.

Land subsidence plays a key role resulting in the variability of the vulnerability to RSLR. Land subsidence rates vary from less than 1 to about 5 mm yr-1, with local peaks reaching 10 mm yr-1. Thus also under future scenario in some sectors of the Venice coastland land subsidence and sea-level rise of the Adriatic Sea contribute equally by 2050.

It is important to highlight that the specific indices used here do not exhaustively cover all of the various factors influencing the Venice coastland in view of the expected RSLR and various assumptions based on the scientific knowledge of the study area have been made to simplify the analysis. Assumptions, along with classification and weighting criteria inevitably lead to some degree of subjective judgments. The choice of the SLR scenarios to be used in the analysis is a critical issue. Regarding the present state, notice that the yearly average value of sea level changes is quite variable and is characterized by pseudo-cyclicity and oscillations (e.g., Carbognin et al., 2010). Time series shorter than 50–60 years often yield fuzzy and contradictory predictions. The 1.3 mm yr-1 average SLR over the 1875–2018 time interval, i.e. the ongoing long-time series scenario, is certainly significant and robust for the Adriatic Sea. However, the 3.8 mm yr-1 resulting from the 1992–2018 period, i.e. the ongoing short-time series scenario, shows a fairly uniform trend that could be more representative for the present state. Regarding the future scenario, the uncertainty characterizes the quantification of both the expected SLR and land subsidence evolution. To this end, the expected scenario here considered is limited to 2050, being an evolution model of land subsidence unavailable for the study area and the large difference of the SLR projections by 2100. In any case, it is possible to assume that land subsidence will continue in accordance with the present trend because the groundwater regulation adopted in Venice and the long-term secondary consolidation of the Holocene deposits will remain almost unchanged over a relatively short period. Regarding the global sea-level rise by 2050, a rate of 5 mm yr-1 computed from the IPCC projections is adopted (IPCC, 2014).

In relation to the shoreline erosion and the sinking of the embankments, it was assumed that over the next 30 years, beaches nourishment and land reclamation will be able to keep pace with the RSLR.

The vulnerability to RSLR exhibits substantial variability among individual areas reflecting the heterogeneity of land subsidence and the hydro-morphology setting. The key results of this study are the following:

Land subsidence rates vary from less than 1 to more than 5 mm yr-1. Thus, it means that also under future scenario land subsidence in some sectors of the Venice coastland contributes equally with the sea-level rise of the Adriatic Sea in the vulnerability analysis.

Under the ongoing long-time series scenario, 30.5 %, 23.5 % and 22 % of the coastland area is classified with marginal, moderate and critical vulnerability, respectively.

Under the future scenario, i.e. by 2050, the areas with critical and extreme vulnerability are expected to cover 60 % of the Venice coastland.

Although this study represents a step forward in the assessment of the vulnerability to RSLR of the Venice coastland, the authors are aware that further studies will have to face the uncertainties not addressed in this analysis.