3.2.2. Coastal Vulnerability Associated to Risk of Land Flooding

The analysis of coastal vulnerability associated with flood risk of river origin is even more interesting. In this case, the reverse phenomenon occurs, with the interior coastal perimeter of the Mar Menor being the area where the impact is concentrated (Figure 10). This is a consequence of the existing orographic distribution with numerous wadis that flow into that stretch of coast. In this case, the natural areas are not actually the most affected (despite the existence of crypto-wetlands theoretically within these protected areas), but the areas of greatest concentration are urban and agricultural ones. Due to this, a detailed DTM has been incorporated into these urban areas using LIDAR technology to obtain more precise results.

**Figure 10.** Modeling of the danger by fluvial flooding on a LIDAR DTM of the coastal strip for a return period T = 100 years. Three detailed examples of the impact are attached: on the urban plot in the western urban settlements (**top**), in the protected natural areas (**middle**) and in the urbanized area of the southern perimeter (**below**), (data source for GIS analysis: see Reference [56]).

In the West side we find large urban areas subject to an intense phenomenon of flooding throughout its urban surface (particularly close to the coastal town of Los Alcazares). This phenomenon is mainly a consequence of the urban configuration and the growth patterns of the town during recent decades, which have not taken the risks associated with the orography of the land into account. This growth pattern and urban structure of medium-low density have only aggravated existing problems. However, it should be noted that the highway (built in the 1990's) that surrounds the town exerts a "dam effect". This new local configuration greatly protects the town, being possibly related to the fact that the recent floods in the town have been less catastrophic than the last ones of 60 years ago. Even so, this configuration is a double-edged sword, since the possibility of occurrence of a maximal event of greater dimensions (for example, with a return period T = 500 years) could suppose the blockage of the drainage elements of the highway, thus making this dam overflow directly into the population, which would significantly aggravate the consequences for the town.

In the Southern area, we find a very different problem. In this case, the agricultural areas are the most affected. However, the current orographic configuration of the agricultural lands allows easy access of the waters to several small coastal towns, flooding them. As we have already seen, this orographic configuration has not always been thus, but has been subjected to various land processes of transformation of use and structure over the last decades. We find transformations throughout these last decades that may be susceptible to worsening as well as improving the status quo of the territory analyzed. Therefore, in the last section the extent to which human action has contributed to worsen or improve the existing situation in this field will also be discussed, based on the results of the geo-statistical analysis.

#### *3.3. Geostatistical Bivariate Analysis of GIS Indicators*

In the first place, we have made the two GIS indicators of the distribution of flood risks of marine and land origin dimensionless, as explained in the methodology section. Then, we have integrated them into a single statistically homogenous GIS indicator, called index of global coastal vulnerability *IGCV*. This indicator has been spatially correlated with the different UTD, CIR, and DLT impact indicators through the Global Moran's I statistic. This bivariate analysis assesses the relationship two to two between the indicators at the level of two-dimensional autocorrelation in the spatial plane. The aggregate result of this first global correlation can be summarized for the scope of study in Table 4.


**Table 4.** Bivariate Global Moran's I statistics for spatial autocorrelation between *IGCV* global coastal vulnerability index and the three GIS impact indices UTD, CIR, and DLT.

The results show positive global statistical autocorrelation between indices, but higher levels of correlation for the *DLT*-*IGCV* and *UTD*-*IGCV* couples than the *CIR*-*IGCV* one. This difference may be due to the more direct incidence at a spatial level of the processes of urbanization and direct transformation of the coastal edge, rather than the execution of marine and port infrastructures. The incidence of the latter can be derived indirectly at a spatial level far from the focus of the impact through the coastal dynamics, so a more local geo-statistical analysis is necessary in order to analyze this relationship.

The level of local correlation between the different pairs of indices was evaluated by means of the Anselin Local Moran's I (LISA) statistic. This statistic allows us to transpose at a local spatial level the so-called hot spots (high impact-high vulnerability), cold spots (low impact–low vulnerability) and significant outliers (Low-High and High-Low cross-links) to avoid the limitations of global geo-statistical analysis. A varied catalog of different situations can be observed; the most significant results of this statistic are summarized in the tessellated mesh distribution of Figure 11.

**Figure 11.** Tessellated mesh distribution of Local Anselin Moran's I statistic for aggregated impacts and vulnerability correlation analysis with hot (HH) and cold (LL) spots and outliers (LH and HL).

We can here observe how the processes of urbanization in the inner perimeter of the Mar Menor have notably increased the coastal vulnerability of land origin in the territory, despite the theoretical beneficial impact of some infrastructures such as the highway. Similarly, the transformation of agricultural orography in the Southern perimeter has favored this type of vulnerability, increasing the risk of flooding of land origin. Nevertheless, it is paradoxical in this case to observe how the impact of the so-called island-ports in the Mar Menor has increased the protection against the marine flooding of the Mar Menor in some villages such as Los Nietos.

On the other hand, the progressive dislocation of beaches in the Northern area of La Manga as a result of artificial dredging and the development of port infrastructures has increased the coastal vulnerability of the Northern half of La Manga due to flooding of marine origin coming from the Mediterranean. A different situation is found in the Southern half. There, the same dense and compact urbanization process that mitigates the impact of the coastal vulnerability of marine origin from the Mediterranean on the urban plot is responsible for the indirect vulnerability that occurs on the side of the Mar Menor by the disappearance of several beaches due to the effect on coastal dynamics.

#### **4. Discussion**

The Mar Menor, La Manga and its neighboring coastal territory have long been a common object of social and scientific controversy. As mentioned before, there are numerous studies in the scientific bibliography related to the Mar Menor [57–64] and the problems derived from its unique configuration and high environmental value. Most of them converge in associating the environmental impacts of the territory individually, mainly with the intensive urbanization of the coastal perimeter, the construction of marine infrastructures and the anthropization of the environment with actions such as intensive agriculture or the artificial widening of the "golas". Nevertheless, practically none of them address the problems of this territory in a global way from the perspective of the physical impact on the territory. They focus instead on analyzing the existing segmented consequences from the point of view of water quality [65,66], marine fauna [37,61], seabed vegetation cover [38,60], birds [39,67], landscape [47], etc.

In addition, the main biological, chemical, or ecological approach of most of the existing studies in the area has traditionally focused on variables with little connection with the coastal vulnerability of the territory as impact factors (nitrates from agriculture, heavy metals from old mining areas, etc.). Therefore, except for a number of specific and segmented studies in the field of coastal hydrodynamics [68,69], sea level [70], or territorial anthropization [34,47] there was no analysis from the physical perspective of the coastal territory that allows a diagnosis of the global vulnerability in this area.

This study presents a different approach to all the existing analyses to date. The different factors related to the current coastal vulnerability of the territory at a global level have been analyzed spatially. Additionally, through an innovative methodology based on the retrohistoric GIS analysis of anthropization processes derived from human activity, it has been possible to geo-statistically correlate the link between the impacts of phenomena such as the construction of ports or the change of land use, with the current coastal vulnerability. This type of diagnosis may prove to be of great interest in order to develop strategies for mitigating coastal vulnerability in a complex territory, since it not only analyzes existing risks through hazard maps, but takes into account how human activity contributes to current problems.

In this sense, to correctly implement these strategies to mitigate the existing vulnerability, it is very important to know which elements have a negative impact and which have a positive or negative impact, both in the current scenario and in their trend trajectory. In addition, it must be taken into account that complex environments such as the one analyzed can have cross-linked impacts, generating a negative effect in one place, while generating a positive effect in another. This mapping context with different scenarios of the couple human impact-vulnerability (HH, LL, HL, and LH) would allow us to implement a segmented mitigation strategy with different sub-strategies depending on how different impacts of human activity affect the current coastal vulnerability at a spatial level.

For example, in the case study analyzed, in the inland Western area of the Mar Menor, one should preferably act on the orography of the wadis, making it compatible with the current urban layout and preventing future urban developments from being carried out in the areas of greatest risk. In the Southern perimeter, attention must also be paid to the flood risk of land origin. However, where it is really necessary to act is in the configuration of the agricultural areas, since the current urban settlements have a fairly stabilized growth and it is the transformation of the agricultural orography that fosters the greatest risk of flooding at present. On the other hand, in the La Manga area, the need to implement at least two different sub-strategies is observed. In the Northern zone, it would be necessary to act in the urban plot configuration (since it is observed that it is continuing to be urbanized at present) against the risk of flooding of marine origin from the Mediterranean (in this case aggravated by the impact of the sedimentary dynamics of the area linked to the port infrastructures located further North). A very different strategy would be needed in the Southern area, where the configuration of the urban plot mitigates the risk of flooding of marine origin from the Mediterranean, while indirectly promoting negative effects on coastal vulnerability on the Mar Menor side. In this

case it would be necessary to implement a strategy to rebalance the current situation between the two sides of the ancient dune cord.

This approach is especially interesting for a coastal vulnerability study given that in territories such as this one, issues like the impact of rising sea levels as a result of climate change have been socially very controversial. In fact, approaches in this territory have traditionally existed that lack a solid scientific basis; these have generated some social alarm about the impact of the risk of a sea level rise due to climate change (Figure 12). In complex environments such as this one, approaches with great scientific rigor, such as can be seen in Reference [70], may even be insufficient to assess coastal vulnerability at a comprehensive level, since it is also necessary to take into account the impacts and cross-links of other variables related to the anthropization processes of human activity and to develop detailed digital models of the terrain and urban plot able of simulating the physical reality at the local level.

**Figure 12.** Simulation carried out in 2007 of the estimated sea level rise in different sites of La Manga for the year 2050 as a result of climate change. Source: Greenpeace.

In this sense, possible lines of future research in this area for this territory to formulate more precise and segmented mitigation strategies could be carried out to deepen the study at the local level of certain elements whose performance was positive or negative depending on the context of analysis. For example, retrospective GIS analysis and geo-statistical correlation could be used to determine in greater detail which seaside dikes are detrimental or beneficial to the stability of sedimentary dynamics, or to assess the long-term effect of dredging on the golas at an environmental level. Both issues have traditionally generated important social debates in the area: In the first case in relation to the need to eliminate or not the marine dikes to improve the quality of beaches, and in the second, on the possibility of further expanding the golas with the aim of improving the quality of the waters in the Mar Menor, despite the risk of deepening the process of "mediterraneanisation" of the lagoon and increasing the risk of marine flooding.

Another important issue concerns the limitations of the methodology used. In the analyzed case, a very accurate geo-referenced cartography was available for different dates, as well as a DTM model that enables the orographic natural and urban reality to be represented in a reliable way. Additionally, the existence of an orographically simple terrain and an urban configuration without great constructive singularities have both considerably facilitated the work. It was only necessary to calibrate the mathematical models to represent the "dam effect" of the AP-7 highway, and the marine flooding in the golas as a result of the dredging presents some minor uncertainties. However, the model at the global level can be considered very robust, with the data used as input to the geo-statistical analysis being considered very reliable. Even so, this statement clearly cannot be extended to every study, and it is thus necessary to maximize the precision in the DTM model and take caution in mathematical models when we are faced with complex natural orography or urban configurations with several constructive singularities.

Consequently, it is clear that this diagnostic methodology can be very beneficial to improve current coastal vulnerability assessment systems and implement more accurate mitigation strategies. It is true that it is a methodology that requires an important geo-referenced database of historical character and with a great deal of precision. In this context, it cannot be said that this new technique can currently be applied in a generalized manner for analysis of coastal vulnerability in any territory. Nevertheless, we must bear in mind that national GIS databases and international Spatial Data Infrastructures (SDI) systems are becoming generalized throughout the world and perfecting their level of accuracy. Therefore, it may be an approach with a promising future that in the coming years will allow for the opening of increasingly sophisticated new lines of research in the field of coastal vulnerability.

#### **5. Conclusions**

This study has presented an innovative methodology for analyzing the coastal vulnerability of a territory based on the GIS evaluation of the spatial statistical correlation of long-term anthropic impacts and the distribution of current risks. The geo-statistical analysis carried out for the case of the Mar Menor Mediterranean lagoon reveals that the urbanization processes being developed in the last decades have generated imbalances. On the one hand, they provoke the retraction of the coastline in the old dune cord called La Manga, increasing coastal vulnerability due to flooding of marine origin. On the other hand, the inadequate urban sprawl has notably increased the vulnerability in the interior coastal perimeter due to flooding of fluvial origin in some coastal towns.

In the case of infrastructures, we find a catalog of very heterogeneous situations. The construction of some ports (although not all) have severely affected the balance of beaches, making them expand or almost disappear (and therefore increasing the risk of marine flooding). The role of several motorways, whose barrier effect theoretically mitigates the risk of terrestrial flooding in some coastal towns, should also be highlighted. This situation is actually a double-edged sword because, in the event of a flood that overcomes that dam effect, the consequences could be catastrophic. Finally, land use transformations or dredging and earthmoving in the coastal strip have also caused different impacts on the vulnerability of the territory. Changes in agricultural use in the Southern fringe have increased the risk of land-based flooding, while dredging in the La Manga golas increases that of marine origin.

**Supplementary Materials:** The KML file including the area of analysis is available online at http://www.mdpi. com/2073-4441/10/11/1642/s1.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The author declares no conflict of interest.
