Quantify the Contribution of Nature-Based Solutions in Reducing the Impacts of Hydro-Meteorological Hazards in the Urban Environment: A Case Study in Naples, Italy

: Urban areas are vulnerable to multiple risks associated with hydro-meteorological hazards (HMHs). The assessment of the climate beneﬁts of implementing nature-based solutions (NBSs) in urban areas, especially in open spaces, is widely recognised and discussed within the scientiﬁc literature; however, the quantiﬁcation of these beneﬁts, in terms of the HMHs reduction, human safety and human well-being, is still a subject of debate. In this context, this contribution proposes a methodological approach that, starting from the analysis of the impacts of coastal ﬂooding and in terms of the potential direct and tangible economic damages, heatwave events and vulnerability of open spaces, proposes the application and assessment of NBSs in terms of the reduction in these impacts. The process was developed in the GIS environment based on the processing of open-source data. The test was conducted in the case study of Naples’ waterfront to identify the potentialities and limitations of the approach. The results showed the contribution of NBSs in reducing the economic damages due to coastal ﬂooding and the improved vulnerability conditions to heatwave events.


Introduction
Climate change has severe impacts on humans and their habitats. Among natural hazards, Hydro-Meteorological Hazards (HMHs) are naturally occurring meteorological, climatological and/or hydrological phenomena that can induce impacts on both natural and human environments. In the future, HMHs, including cyclones, coastal storm surges, flash floods, heatwaves and droughts, will be increasingly frequent and intense, and will be concentrated mostly in urban areas [1,2].
In urban areas, the implementation of Nature-Based Solutions (NBSs) to reduce the impacts associated with HMHs, both in terms of the mitigation of the causes and adaptation to the effects, is widely recognised and discussed within the scientific literature [3][4][5][6][7]. NBSs are defined by the European Commission as solutions that "are inspired and supported by nature, they are cost-effective, simultaneously provide environmental, social and economic benefits and help build resilience" [8]. These measures include multiple approaches such as ecosystem-based adaptation, ecosystem-based mitigation, eco-disaster risk reduction and green infrastructure [9][10][11].
Urban coastal settlements, in Europe and worldwide, are among the areas most at risk due to a high exposure (of people, activities, goods and services) and due to the intrinsic vulnerability of territories. These areas are subject to multiple HMHs, in addition, considering the long-term climate scenarios, the direct relationship between the rising of global average temperatures and the rising of sea levels will further compromise the equilibrium of territories that are already fragile [12][13][14][15][16][17].
Concerning sea-related HMHs, traditional grey and highly engineered defence systems, such as groins, seawalls, breakwaters and storm surge barriers, have over the years protected coastal territories. Nowadays the negative impacts of this "grey approach" on • the approach is aimed at quantifying the positive effects of NBSs in reducing the impacts associated with HMHs (coastal floods and heatwaves) in urban public open spaces; in fact, these hazards will increase in both intensity and frequency in all urban and metropolitan settlements along the coast; • the proposed GIS-based framework represents a useful support tool for decisionmakers to enable the simulation and assessment of climate-resilient multi-hazards solutions (NBSs); • the framework is replicable in different urban settlements, as it uses mainly opensource data that is independent from the specificity of the urban study area.
In Section 2 are introduced the methodologies and approaches applied to quantify impacts due to coastal flooding and heatwaves (HMHs). Moreover, our GIS-based framework is presented. In Section 3 the proposed framework is tested on a waterfront area in the municipality of Naples, Italy, selected as a critical area based on recent storm surge events. In Section 4 the potentialities and the limitations of our approach are discussed. Conclusion and research perspectives are in Section 5.

The PLANNER and Coast-RiskBySea Methodological Frameworks
To assess the effectiveness of NBSs in reducing the impacts related to hydro-meteorological hazards, a comprehensive knowledge of the risks is required, starting from interventions that can be simulated and their effectiveness assessed. To analyse the potential risks by 2100, considering coastal floods due to extreme sea level events and heatwave scenarios, two models were proposed that were developed within the environmental design and computer science research group of the Department of Architecture, University of Naples Federico II: the PLANNER model, implemented starting from the methodological framework developed by [41] and the Coast-RiskBySea model [42].
Both models were developed in the GIS environment and provided an analysis based on medium and long-term climate scenarios: 2050 and 2100.
The PLANNER model allowed for the assessment of the heatwave impacts on the population in the urban environment. The model used census zones, processed by ISTAT (Italian National Institute of Statistics) and remotely sensed data as the base data. The results were several impact scenarios derived from the combination of the exposed population type, intrinsic vulnerability of the urban system and hazard scenarios to the heatwave.
The PLANNER model workflow was schematized in Figure 1.
To analyse the potential risks by 2100, considering coastal floods due to extreme sea level events and heatwave scenarios, two models were proposed that were developed within the environmental design and computer science research group of the Department of Architecture, University of Naples Federico II: the PLANNER model, implemented starting from the methodological framework developed by [41] and the Coast-RiskBySea model [42].
Both models were developed in the GIS environment and provided an analysis based on medium and long-term climate scenarios: 2050 and 2100.
The PLANNER model allowed for the assessment of the heatwave impacts on the population in the urban environment. The model used census zones, processed by ISTAT (Italian National Institute of Statistics) and remotely sensed data as the base data. The results were several impact scenarios derived from the combination of the exposed population type, intrinsic vulnerability of the urban system and hazard scenarios to the heatwave.
The PLANNER model workflow was schematized in Figure 1. The Coast-RiskBySea model allowed for the assessment of coastal flooding impacts due to extreme sea level events on the built environment. The model used homogeneous hexagonal atomic units and remotely sensed data as the base data. The results were several risk scenarios derived from the combination of the exposed built environment type, vulnerability of coastal areas and hazard scenarios to the extreme sea level. The Coast-RiskBySea model workflow is schematized in Figure 2. The Coast-RiskBySea model allowed for the assessment of coastal flooding impacts due to extreme sea level events on the built environment. The model used homogeneous hexagonal atomic units and remotely sensed data as the base data. The results were several risk scenarios derived from the combination of the exposed built environment type, vulnerability of coastal areas and hazard scenarios to the extreme sea level. The Coast-RiskBySea model workflow is schematized in Figure 2.
The workflow of both models was based on the risk assessment framework proposed by the IPCC in Reports AR5 and AR6 [2], in which the climate risk on urban settlements is assessed, combining hazard, vulnerability and exposure. In both models, the vulnerability meant the extent to which the physical system, open spaces and buildings were sensitive to or unable to cope with the negative effects of climate change [2].
In the PLANNER model, the impacts were generated by referring to a specific hazard scenario produced by the HW phenomenon that acts on an urban physical system evaluated in terms of the vulnerability to the phenomenon. The vulnerability was measured through indicators that summarised the fundamental characteristics of the physical system in reference to the HW phenomenon. These characteristics referred to the construction, morphological, technological and environmental aspects.  The workflow of both models was based on the risk assessment framework proposed by the IPCC in Reports AR5 and AR6 [2], in which the climate risk on urban settlements is assessed, combining hazard, vulnerability and exposure. In both models, the vulnerability meant the extent to which the physical system, open spaces and buildings were sensitive to or unable to cope with the negative effects of climate change [2].
In the PLANNER model, the impacts were generated by referring to a specific hazard scenario produced by the HW phenomenon that acts on an urban physical system evaluated in terms of the vulnerability to the phenomenon. The vulnerability was measured through indicators that summarised the fundamental characteristics of the physical system in reference to the HW phenomenon. These characteristics referred to the construction, morphological, technological and environmental aspects.
Meanwhile, in the Coast-RiskBySea model, the impacts on the built environment due to coastal floods were assessed according to the potential direct and tangible economic damages. The risk depended on the vulnerability of the territories as a function of the mean elevation, and the exposed value of the built environment as a function of land-use characteristics.
Both models were developed in a GIS environment and were characterised by the use of open-source data, a condition that allowed for the replicability of the analyses even in spatial contexts different from those tested.

The Proposed Framework
Starting from the presented methodologies and input data source, a new GIS-based framework was proposed to test the performance effectiveness of Nature-Based Solutions (NBSs) in the requalification project of urban public open spaces, oriented towards the risk and impact reductions to Coastal Flooding (CF) due to extreme sea level events, and to heatwaves (HWs).
The joint use of the two models provided the opportunity to approach multi-hazard simulations by simultaneously assessing the conditions of vulnerability to CF (assessed as a function of the mean elevation value above sea level), the associated potential risks (assessed as a function of the potential direct and tangible economic damages on the built environment) and the vulnerability to heatwaves (assessed as a function of Albedo, NDVI, Hillshade and Sky View Factor indicators). After analysing the risk and vulnerability con- Meanwhile, in the Coast-RiskBySea model, the impacts on the built environment due to coastal floods were assessed according to the potential direct and tangible economic damages. The risk depended on the vulnerability of the territories as a function of the mean elevation, and the exposed value of the built environment as a function of land-use characteristics.
Both models were developed in a GIS environment and were characterised by the use of open-source data, a condition that allowed for the replicability of the analyses even in spatial contexts different from those tested.

The Proposed Framework
Starting from the presented methodologies and input data source, a new GIS-based framework was proposed to test the performance effectiveness of Nature-Based Solutions (NBSs) in the requalification project of urban public open spaces, oriented towards the risk and impact reductions to Coastal Flooding (CF) due to extreme sea level events, and to heatwaves (HWs).
The joint use of the two models provided the opportunity to approach multi-hazard simulations by simultaneously assessing the conditions of vulnerability to CF (assessed as a function of the mean elevation value above sea level), the associated potential risks (assessed as a function of the potential direct and tangible economic damages on the built environment) and the vulnerability to heatwaves (assessed as a function of Albedo, NDVI, Hillshade and Sky View Factor indicators). After analysing the risk and vulnerability conditions, it will be then possible to simulate one or more climate-adaptative design solutions, in this case by NBSs, to verify and prefigure the effectiveness of projects and so to provide decision support in the evaluation of design alternatives.
As anticipated, the framework focused only on the design of public open spaces. The purpose of the GIS-based framework was to support decision-makers in the planning and design of NBSs. The support provided is in assessing the performance effectiveness of NBSs in terms of the damage reduction to CF and the vulnerability reduction to HW. The framework was based on the following steps: 1.
the input data collection and definition of reference atomic units; 2.
the development and implementation of computational processes to assess potential economic damages related to CF, applying the Coast-RiskBySea methodology, and vulnerability to HW and applying the PLANNER methodology; 3.
the identification of critical areas of where to plan the design of the NBS; Land 2023, 12, 569 5 of 20 4. the simulation of one or more NBSs and the evaluation of their effectiveness on CF and HW.
Thus, the first step was data source collection, which, to ensure the replicability of processes in multiple contexts, were mainly open-source.
The framework, in synthesis, was based on data that can be categorised mainly in three categories: topographic data, satellite data and derived from pan-European opensource databases.
Topographic data (DBT) can be found on the online and open-source platforms of local and national governments; these data are institutionally recognised. By means of such data, it was possible to structure in GIS the "knowledge base" based on qualitative and quantitative data to identify the functional-spatial and dimensional characteristics of open spaces.
Jointly with DBT, satellite data were essential to implement the knowledge base, concerning indicators related to the CF and HW vulnerabilities.
Having collected the data, atomic reference units were identified and processed for simultaneous data management in GIS. Based on the two models (PLANNER and Coast-RiskBySea), new atomic reference units were created, as the two models were structured starting from different reference units. The first model used a hexagonal reference grid by which the territory was divided into n parts of the same surface extension, and the second model used ISTAT census zones to have a homogeneous distribution over the whole area of the section of census data.
Thus, considering both phenomena, the elevation above sea level was identified as a key factor in identifying and structuring the new atomic reference units.
Based on the Digital Terrain Model (DTM), the heights Above Mean Sea Level (AMSL) were reclassified through a process of spatial analysis by assigning a unique class value to each elevation interval. The threshold intervals used for the classification were the same as those used in the Coast-RiskBySea model [42]; the intervals were fixed starting from those defined in the "global depth-damage functions" [43] in order to calculate the potential economic damages on the built environment. Values were classified through a manual method, assigning to each range an AMSL class (Table 1). Once the AMSLs were identified on the reclassified raster file, the data were converted to the polygonal shapefile format in order to be processed by spatial intersection with the Topographic Database (DBT) of the Municipality under study. This intersection provided data regarding functional-spatial and dimensional characteristics of the areas, useful for calculating vulnerabilities and potential risks. Considering that the purpose was to provide support in the decision-making process for the design of public open spaces, only public open areas were considered.
The output of the process consisted of polygons derived from the DBT, which retained data on the type of open space and the relative use of them, but which were decomposed into multiple polygons according to the ASML classes derived from the DTM. Contiguous features, belonging to the same ASML class and with the same land use, were then dissolved into a single polygon. The second step involved the implementation and the application of computational processes to calculate the risk related to CF and vulnerability related to HW.
The HW vulnerability of open spaces was assessed by considering the interaction between the 4 basic indicators: Albedo, Hillshade, NDVI and Sky View Factor, following the methodological framework [41] adopted in the PLANNER model. To each indicator a weight was given with which the vulnerability could be summarised through a weighted average.
Once the value was achieved, the indicator was then classified into the 5 classes shown in Table 2. As anticipated, the CF vulnerability, following the Coast-RiskBySea methodological framework [42], was assessed by considering the height Above Mean Sea Level (AMSL) of each open space, more precisely by the mean elevation polygon position above the sea level.
Starting from the basic feature of the AMSL and extrapolated from the DTM raster, the values were divided and classified into 8 classes. Values were always approximated by default to avoid any underestimation of the potential climate consequences.
To each interval was assigned a vulnerability class, as defined in the table (Table 3). Once the vulnerability to CF was also determined, the CF long-term risk scenarios were calculated through the integration of exposure and hazard. This processing allowed for us to obtain the business-as-usual scenarios, i.e., the scenarios where no climate mitigation/adaptation solutions are introduced.
Damages were assessed through the depth-damages function by the EU Joint Research Centre [43]; in this function were assigned a given specific water depth and normalised index, based on the six General Land Use classes (GLU) and a maximum cost in €/sqm based on national estimates.
Once the maps have been elaborated, decision-makers can then identify the critical areas where they can intervene, based on the potential risks to CF and potential vulnerabilities to HWs, with NBSs. Considering the lifecycle of projects, it is important to perform simulations related to the long-term scenarios to plan and design solutions that are resilient over time.
After identifying the focus area, decision-makers can test the effectiveness of the chosen solutions by simulating the interventions. The identified design alternatives can be evaluated based on the relative intensity of the transformation and on the improvement increase in terms of the HW vulnerability reduction and CF risk reduction.

Test and Case Study
After identifying the methodological framework, the proposed GIS-based approach needed to be tested on a real case study to verify its potential and limitations.

Study Area
The city of Naples, located in the Campania Region in Southern Italy, was identified as a case study. As anticipated in the Introduction, urban-coastal areas would be subject to multiple risks associated with the direct, as well as indirect, effects of climate change.
European coastal cities are extremely vulnerable to sea-related hazards and these territories are, in fact, already impacted by natural hazards, which in the future will be increased due to the gradual rise in the sea level as a result of the increase in global mean temperature [12][13][14][15][16]. In addition to these risks, urban areas, due to dense urbanisation, scarce vegetation, and air pollution, are also susceptible to heatwave events [1,2,27,28].
Recent studies highlight that among Italian coastal cities, Naples will be subject to the impacts of HMHs, requiring the implementation of climate mitigation and adaptation strategies and solutions [17,44].
Once selecting Naples as a case study, the area of Via Partenope ( Figure 3) was highlighted as a critical area because of the recent storm events that have affected the area [45]. evaluated based on the relative intensity of the transformation and on the improveme increase in terms of the HW vulnerability reduction and CF risk reduction.

Test and Case Study
After identifying the methodological framework, the proposed GIS-based approa needed to be tested on a real case study to verify its potential and limitations.

Study Area
The city of Naples, located in the Campania Region in Southern Italy, was identifi as a case study. As anticipated in the Introduction, urban-coastal areas would be subj to multiple risks associated with the direct, as well as indirect, effects of climate chang European coastal cities are extremely vulnerable to sea-related hazards and these t ritories are, in fact, already impacted by natural hazards, which in the future will be creased due to the gradual rise in the sea level as a result of the increase in global me temperature [12][13][14][15][16]. In addition to these risks, urban areas, due to dense urbanisatio scarce vegetation, and air pollution, are also susceptible to heatwave events [1,2,27,28] Recent studies highlight that among Italian coastal cities, Naples will be subject the impacts of HMHs, requiring the implementation of climate mitigation and adaptati strategies and solutions [17,44].
Once selecting Naples as a case study, the area of Via Partenope ( Figure 3) was hig lighted as a critical area because of the recent storm events that have affected the area [4 In terms of the urban settlement, the area is highly significant for its dense residential and commercial fabric and historical and cultural emergencies; the area is also relevant because of the high number of users (residents and tourists) who live in the public open spaces.
Once the case study area was identified, the first step was the input data collection and the creation of the atomic reference units.

Data Source
The input data source to implement the proposed model were the topographic, satellite and derived from pan-European open-source databases. In synthesis, the data were listed by defining the spatial scale, the main characteristics and the source:  [48].
The input data, although derived from institutional sources, were not homogeneous. To use the data in GIS computational processes, once acquired, it was necessary to convert the data on the same coordinate system to reproduce a correct coding of the fields.

Experimental Tests
The proposed methodological framework was summarised schematically in Figure 4 to understand the relationship between the components of the exposed elements, vulnerabilities and hazards.
Once the case study area was identified, the first step was the input data collection and the creation of the atomic reference units.

Data Source
The input data source to implement the proposed model were the topographic, satellite and derived from pan-European open-source databases. In synthesis, the data were listed by defining the spatial scale, the main characteristics and the source: • topographic database (DBT) in shapefile format regarding the land use type and its characteristics, the scale is 1:5000, data were provided by the Campania Region geodatabase [46]; Extreme Sea Level climate projections by the EU JRC, data were characterised by a municipal scale [48].
The input data, although derived from institutional sources, were not homogeneous To use the data in GIS computational processes, once acquired, it was necessary to conver the data on the same coordinate system to reproduce a correct coding of the fields.

Experimental Tests
The proposed methodological framework was summarised schematically in Figure  4 to understand the relationship between the components of the exposed elements, vulnerabilities and hazards. After the data collection, the second step was the creation of the new atomic units Starting from the Naples DBT [46], considering that the purpose of the framework was After the data collection, the second step was the creation of the new atomic units. Starting from the Naples DBT [46], considering that the purpose of the framework was oriented decision support in the design of public open spaces, all shapefiles that pertained to this category were selected and uploaded in GIS (Table 4). Through a merging operation, polygons pertaining to the same class were grouped into a single shapefile; meanwhile the DTM raster was reclassified according to the intervals shown in Table 1 related to the AMSL.
Through a spatial intersection operation between the result of the DBT merge and the new shapefile obtained from the DTM, it was possible to obtain the shapefile composed of polygons differentiated by the ASML field. Finally, a dissolve operation was performed to merge all contiguous polygons with the same ASML class (elevation) and land use derived from the DBT. These steps enabled the creation of the atomic reference units of the case study area.
In addition to the creation of the atomic reference units, the vulnerability of open spaces to the Coastal Flood (CF) phenomenon was assessed with the ASML field. The ASLM values, according to the intervals presented in Table 3, were divided into 8 vulnerability classes. The CF vulnerability map was shown in Figure 5, jointly with the percentage distribution of vulnerability classes.
oriented decision support in the design of public open spaces, all shapefiles that pertained to this category were selected and uploaded in GIS (Table 4). Through a merging operation, polygons pertaining to the same class were grouped into a single shapefile; meanwhile the DTM raster was reclassified according to the intervals shown in Table 1 related to the AMSL.
Through a spatial intersection operation between the result of the DBT merge and the new shapefile obtained from the DTM, it was possible to obtain the shapefile composed of polygons differentiated by the ASML field. Finally, a dissolve operation was performed to merge all contiguous polygons with the same ASML class (elevation) and land use derived from the DBT. These steps enabled the creation of the atomic reference units of the case study area.
In addition to the creation of the atomic reference units, the vulnerability of open spaces to the Coastal Flood (CF) phenomenon was assessed with the ASML field. The ASLM values, according to the intervals presented in Table 3, were divided into 8 vulnerability classes. The CF vulnerability map was shown in Figure 5, jointly with the percentage distribution of vulnerability classes. As shown in Figure 5, the most vulnerable areas are those located along the coast and along the piers, which have low elevation levels. In terms of the percentage distribution, As shown in Figure 5, the most vulnerable areas are those located along the coast and along the piers, which have low elevation levels. In terms of the percentage distribution, more than 60% of the open spaces fell into the Low or Very low vulnerability classes, while only 16% were classified as High, Very high or Extremely high.
After identifying the CF vulnerability, to calculate the exposure, to each polygon is assigned a land use class. Following the Coast-RiskBySea models [42], the exposed value was calculated as a function of the potential direct and tangible economic damages on the built environment, thanks to the introduction of the JRC damage functions by the EU [43]. These functions were differentiated by the General Land Use (GLU) classes, so DBT classes were associated with GLU classes. Table 5 showed the correspondence between DBT land use classes, GLU labels and GLU classes. The result of the process generated the General Land Use map shown in Figure 6.
These functions were differentiated by the General Land Use (GLU) classes, so DBT cl ses were associated with GLU classes. Table 5 showed the correspondence between DBT land use classes, GLU labels a GLU classes. The result of the process generated the General Land Use map shown in Figure 6. The maximum potential exposed values were calculated on each atomic unit throu a tabular joint that allowed for the correspondence between the potential economic da ages defined by the global depth-damage functions and the GLU classes. The maximum potential exposed values were calculated on each atomic unit through a tabular joint that allowed for the correspondence between the potential economic damages defined by the global depth-damage functions and the GLU classes.
According to the EU JRC depth-damage functions [43], to each GLU class was assigned the corresponding maximum economic damage €/sqm, based on the Italian national averages (Table 6). These functions enabled only the assessment of the direct and tangible economic damages on the built environment, not considering, therefore, intangible and indirect impacts. Combining the exposed value with the vulnerability and hazard scenarios, it was possible to assess the potential economic damage scenarios to CF due to ESL events.
Hazards are considered for the ESL climate projections, expressed as functions of the water dept, measured in meters. In the case study area, considering the ESL projections by 2100, RCP 4.5 scenario, return period of 100 years and the 95th percentile, according to [48], events were expected characterised by a water depth of about 3.22 m.
Combining the exposure, vulnerability and hazard, approximating the water depth to 3.00 m, on the reference atomic units the potential maximum economic damages to the CF were calculated. The risk map is shown in the Figure 7. These functions enabled only the assessment of the direct and tangible econom damages on the built environment, not considering, therefore, intangible and indirect i pacts.
Combining the exposed value with the vulnerability and hazard scenarios, it w possible to assess the potential economic damage scenarios to CF due to ESL events.
Hazards are considered for the ESL climate projections, expressed as functions of water dept, measured in meters. In the case study area, considering the ESL projectio by 2100, RCP 4.5 scenario, return period of 100 years and the 95th percentile, according [48], events were expected characterised by a water depth of about 3.22 m.
Combining the exposure, vulnerability and hazard, approximating the water dep to 3.00 m, on the reference atomic units the potential maximum economic damages to CF were calculated. The risk map is shown in the Figure 7. The damages to the atomic units were classified in 5 classes of risk elaborated b stakeholder, based on the calibration that emerged from an economic investigation a subdivided according to the intervals identified in Table 7.  The damages to the atomic units were classified in 5 classes of risk elaborated by a stakeholder, based on the calibration that emerged from an economic investigation and subdivided according to the intervals identified in Table 7.

Range of Value [€]
Risk Classes Labels D > 300,000.00 € 1 High 100,000.00 € < D ≤ 300,000.00 € 2 Medium-high 30,000.00 € < D ≤ 100,000.00 € 3 Medium 5000.00 € < D ≤ 30,000.00 € 4 Medium-low 0.00 € < D ≤ 5000.00 € 5 Low After identifying the risk scenarios to CF, the vulnerability of open spaces to HW phenomena was then calculated. Through zonal statistical operations, with the various satellite rasters values have been extracted and summarised for the four indicators that will synthesise the vulnerability to HWs: the Albedo, Hillshade, NDVI and Sky View Factor [41].
The The results of the processes are shown in Figure 8. To summarise the vulnerability to HWs, each indicator is given a weight, and a weighted average was calculated. The output of the process is shown in Figure 9. To summarise the vulnerability to HWs, each indicator is given a weight, and a weighted average was calculated. The output of the process is shown in Figure 9. To summarise the vulnerability to HWs, each indicator is given a weight, and a weighted average was calculated. The output of the process is shown in Figure 9. As can be seen from the histogram (Figure 9b), the study area had predominantly a medium-high level of vulnerability to the heatwave phenomenon reaching 71.6%, which was mainly distributed along the waterfront road infrastructure.
Comparing the map of the impacts to CF and the map of HW vulnerability, it can be seen that there was a certain homogeneity in terms of the critical areas. The areas with the highest risk were mainly located along the waterfront road axis. As can be seen from the histogram (Figure 9b), the study area had predominantly a medium-high level of vulnerability to the heatwave phenomenon reaching 71.6%, which was mainly distributed along the waterfront road infrastructure.
Comparing the map of the impacts to CF and the map of HW vulnerability, it can be seen that there was a certain homogeneity in terms of the critical areas. The areas with the highest risk were mainly located along the waterfront road axis.
Once identified, the critical areas to intervene (illustrated by the workflow identified in Figure 4) were re-processed starting from the introduction of the NBSs. In this new workflow, NBSs constituted new input data and the workflow is presented in Figure 10. Once identified, the critical areas to intervene (illustrated by the workflow identified in Figure 4) were re-processed starting from the introduction of the NBSs. In this new workflow, NBSs constituted new input data and the workflow is presented in Figure 10. To highlight the potential use of the proposed decision support framework in the planning and design of NBSs, two solution NBSs were identified, based on the transformation scenario: soft and medium-hard. The first solution (soft) involved the design of the road with a line of trees on both sides of the lane, while the second (medium-hard) included both the permeable paving and planting of trees in flower beds ( Figure 11). To highlight the potential use of the proposed decision support framework in the planning and design of NBSs, two solution NBSs were identified, based on the transformation scenario: soft and medium-hard. The first solution (soft) involved the design of the road with a line of trees on both sides of the lane, while the second (medium-hard) included both the permeable paving and planting of trees in flower beds ( Figure 11). To highlight the potential use of the proposed decision support framework in the planning and design of NBSs, two solution NBSs were identified, based on the transformation scenario: soft and medium-hard. The first solution (soft) involved the design of the road with a line of trees on both sides of the lane, while the second (medium-hard) included both the permeable paving and planting of trees in flower beds (Figure 11). This simulation test was intended as a demonstration exercise and did not consider all the environmental, regulatory, cultural-historical, social and economic constraints that characterise the area. The aim was in fact to test the approach as decision support systems to stakeholders in the planning and design of NBSs oriented towards multi-risk reduction, and not to propose a design solution.
Having identified the NBSs, a statistical parameterisation of the average values of each indicator was carried out to calculate the new HW vulnerability. The objective was to attribute to each NBS a relative average value of Albedo, Hillshade, NDVI, Sky View Factor and, consequently, of vulnerability, according to the analysis of significant samples of open spaces in the territory of the Municipality of Naples.
For the soft NBSs, roads were considered with rows of trees on both sides, a recurring condition in several areas of the city; while for the medium-hard NBSs, urban parks and green urban areas, characterised by permeable surfaces and by the presence of flower beds and trees were selected. This simulation test was intended as a demonstration exercise and did not consider all the environmental, regulatory, cultural-historical, social and economic constraints that characterise the area. The aim was in fact to test the approach as decision support systems to stakeholders in the planning and design of NBSs oriented towards multi-risk reduction, and not to propose a design solution.
Having identified the NBSs, a statistical parameterisation of the average values of each indicator was carried out to calculate the new HW vulnerability. The objective was to attribute to each NBS a relative average value of Albedo, Hillshade, NDVI, Sky View Factor and, consequently, of vulnerability, according to the analysis of significant samples of open spaces in the territory of the Municipality of Naples.
For the soft NBSs, roads were considered with rows of trees on both sides, a recurring condition in several areas of the city; while for the medium-hard NBSs, urban parks and green urban areas, characterised by permeable surfaces and by the presence of flower beds and trees were selected.
Once an adequate number of open spaces existed, such that a representative sample could be obtained, the average values of each parameter were summarised by an arithmetic means. Considering the standard deviation, the class values were assigned to each indicator. In Table 8 are illustrated the results of the parameterisation. As can be seen in the table, for each indicator, the standard deviation did not compromise the change in class of the respective indicator. This indicates that there was a certain level of homogeneity of the values of each parameter for the open spaces sampled for both the soft and medium-hard solutions.
To obtain the new vulnerability values for the critical areas subject to transformation, the original values of the four indicators were replaced with the new values that emerged from the parameterisation. The results, expressed in terms of the HW vulnerability, are shown in Figure 12.
Concerning the CF risk, the soft NBS was excluded from the test as it did not imply any change in the GLU class because its type of transformation does not imply any change in the land use function. The medium-hard NBS on the other hand, in line with the type of intervention, resulted in the change in the original GLU from class 4 (Infrastructures) to class 3 (Commercial).
To re-calculate and obtain the new risk values by 2100, expressed in terms of potential economic damages and considering the same potential risk scenario, in most critical areas was implemented the design of NBSs.
As in the case of HWs, it is evident that the simulation did not take into account the regulatory, economic and cultural-historical value constraints. The original GLU values were replaced with the new values that resulted from the application of the medium-hard NBS. The output map is illustrated in Figure 13.

Soft
0.12 ± 0.01 2 167.00 ± 14.54 2 0.27 ± 0.04 3 0.51 ± 0.17 5 3 Medium-hard 0.12 ± 0.02 2 141.15 ± 8.22 3 0.38 ± 0.11 4 0.51 ± 0. 10 5 4 As can be seen in the table, for each indicator, the standard deviation did not compromise the change in class of the respective indicator. This indicates that there was a certain level of homogeneity of the values of each parameter for the open spaces sampled for both the soft and medium-hard solutions.
To obtain the new vulnerability values for the critical areas subject to transformation, the original values of the four indicators were replaced with the new values that emerged from the parameterisation. The results, expressed in terms of the HW vulnerability, are shown in Figure 12. Concerning the CF risk, the soft NBS was excluded from the test as it did not imply any change in the GLU class because its type of transformation does not imply any change in the land use function. The medium-hard NBS on the other hand, in line with the type To re-calculate and obtain the new risk values by 2100, expressed in terms of potent economic damages and considering the same potential risk scenario, in most critical are was implemented the design of NBSs.
As in the case of HWs, it is evident that the simulation did not take into account t regulatory, economic and cultural-historical value constraints. The original GLU valu were replaced with the new values that resulted from the application of the medium-ha NBS. The output map is illustrated in Figure 13.

Results and Discussion
The test results show that in the study area between the identified solutions, consi ering both phenomena, the medium-hard nature-based solution appeared as the most e fective, taking into account multi-hazard conditions. In fact, this solution allowed for bo

Results and Discussion
The test results show that in the study area between the identified solutions, considering both phenomena, the medium-hard nature-based solution appeared as the most effective, taking into account multi-hazard conditions. In fact, this solution allowed for both the CF risks and the HW vulnerability reductions.
Concerning the reduction in the CF risks, a comparative analysis of ex-ante and ex-post scenarios showed an economic decrease in approximately € 1,123,000. Data are reported in the table below highlighting the corresponding GLU classes (Table 9). In percentage terms, the economic damages were due to CF decrease in approximately 48%. Damage reduction was concentrated in the areas subjected to environmental redevelopment through the introduction of NBSs.
At the same time, in relation to the HW vulnerability, it was possible to note a significant improvement compared to the ex-ante scenario, both for the entire study area and especially for the areas subject to transformation ( Figure 14). In percentage terms, the economic damages were due to CF decrease in approximately 48%. Damage reduction was concentrated in the areas subjected to environmental redevelopment through the introduction of NBSs.
At the same time, in relation to the HW vulnerability, it was possible to note a significant improvement compared to the ex-ante scenario, both for the entire study area and especially for the areas subject to transformation (Figure 14). In Table 10 the change in vulnerability class between the ex-ante and ex-post scenarios are compared, highlighting with each class the percentage of surface area. As shown in the table, the areas that in the ex-ante scenario were classified as Medium-high showed the greatest HW vulnerability reduction, with a decrease in about 24%, followed by the areas classified as Medium with a smaller reduction in about 6%, and finally the High class with a decrease in 0.2%.
In Table 11 is shown a decrease in the HW vulnerability classes in regards to the In Table 10 the change in vulnerability class between the ex-ante and ex-post scenarios are compared, highlighting with each class the percentage of surface area. As shown in the table, the areas that in the ex-ante scenario were classified as Medium-high showed the greatest HW vulnerability reduction, with a decrease in about 24%, followed by the areas classified as Medium with a smaller reduction in about 6%, and finally the High class with a decrease in 0.2%.
In Table 11 is shown a decrease in the HW vulnerability classes in regards to the Pedestrian, Sidewalks and Roadway areas; these changes allow for an overall improvement of HW vulnerability in the case study area. Table 11. Detailed analysis of vulnerability comparison ex-ante and ex-post NBS application.

Benefits and Limits of the Approach
The aim of this study was to propose a GIS-based framework to evaluate the climate benefits of the implementation of NBSs in urban open spaces. The effectiveness of NBSs is assessed in terms of reducing the CF risks and reducing the HW vulnerability. However, there are limitations in the application of the framework related mainly to the quality and quantity of input data.
As mentioned above, the proposed model used as input open-source data that, while enjoying institutional recognition, have editing errors, including the incorrect classification of land use typology, a key feature for the assessment of impacts related to CF. The choice to use open data was aimed at the possibility of replicating the study in different urban settlements.
The quantity limit of the data affected the results of the parameterisation of the indicators of the HW vulnerability (Albedo, Hillshade, NDVI and SVF) for the NBSs and consequently for the simulations of interventions. NBSs were developed from the study of similar cases within the study area. Although the sample of cases was consistent, a greater number of examples could be analysed and used to improve the parameterisation and thus the accuracy and precision of simulations.
The primary goal of the model was to test NBSs in a multi-risk perspective to verify the climate benefits at an urban scale (1:5000), without delving into the design aspect at a local scale of detail. Since the aim was to define a support tool for the decision-maker in order to verify the NBSs efficiency in terms of the risk to CF and HW vulnerability reduction, the framework dis not explore the detail design of the NBSs; in order to provide a good trade-off between the portability and accuracy about the application of NBSs. This framework limit will be overcome using large scale detail data allowing for the accuracy in the design solutions.

Conclusions and Research Perspectives
Multi-risk climate conditions and the complexity that characterises urban systems require the implementation of design solutions that include both adaptive and mitigation strategies. Nature-based solutions are well-suited in projects on public open spaces and offer their contribution by bringing both climate benefits and numerous environmental and social co-benefits.
In this scenario, the proposed GIS-based framework was aimed at supporting decisionmakers to assess the climate benefits associated with the design and implementation of NBSs in urban open spaces oriented towards reducing the impacts associated with coastal flooding and the improvement of heatwave vulnerability conditions.
The model can thus be used by decision-makers as a decision support tool, in the planning and design of climate-resilient multi-hazards design solutions. Thanks to the simulations performed in the GIS-based framework, it is possible to highlight the critical areas in terms of the coastal flooding risk and heatwave vulnerability, then the NBSs design alternatives and their benefits can be assessed.
The use of institutional and open data released on a 1:5000 scale makes the process replicable in other urban settlements; future tests will include the application of our framework in other national and European case studies.
The tests were performed by considering coastal flooding and heatwave hazard scenarios; in the future, we intend to apply our framework including other types of climate change phenomenon.
The test on the case study area, in the Municipality of Naples, shows how the implementation of NBSs is able to improve resilient multi-risk conditions of urban open spaces. It also shows the different degree of efficacy of the proposed strategies in terms of the CF risk and HW vulnerability reduction based on soft or medium-hard design approaches.
In the future, implementing the framework with detailed approaches will be possible, to define the transformation limits of each open space, providing better decision support for local stakeholders. For this purpose, in the future we intend to integrate the framework with a detail component that can define a larger scale design to evaluate the transformation project in terms of the damage reduction, optimizing the economic investment necessary for the realisation (cost/benefit ratio).