Current and Future Multirisk Analysis in Climate Change Scenarios with Riskcoast WebGIS

Abstract

Several regions in Europe are exposed to multiple climate hazards, although their integrated understanding is still limited. The Riskcoast WebGIS platform, developed in the context of the project with the same name, aims to identify the exposed elements and carry out a current and future multirisk mapping assessment in climate change scenarios, for a set of climate hazards: landslides, flash flooding, estuarine flooding, coastal flooding, and coastal erosion. The main objective of this chapter is to present the main functionalities of the Riskcoast WebGIS platform and the multi-risk assessment capabilities for different future risk scenarios arising from climate change applied to the case study of the municipality of Setúbal, Portugal.

1. Introduction

Human-caused climate change is already generating weather and climate extremes in every region across the globe, leading to widespread adverse impacts on food and water security, human health, and on economies and society and related losses and damages to nature and people [1]. The increase in the frequency, intensity and geographical extent of hazardous events associated with climate extremes has been analyzed in various studies [2, 3], denoting a growing concern in understanding the processes of climate evolution and adaptation. In recent years, research on the adverse impacts associated with extreme events has linked climate change to the amplification of the risk of various phenomena such as floods, coastal floods, heat waves, droughts, storms, or forest fires, studying this relationship for single specific climate or weather hazard [2, 4, 5]. Moreover, the development of comprehensive approaches to assess natural disaster risks, and in particular those related to climate change with a multiple risk perspective, has been addressed by few authors [4, 6, 7], considering all aspects that contribute to increased hazard, exposure, and vulnerability. Recognizing and integrating the manifold hazards linked to climate change is crucial, given that threats become more pronounced in regions exposed to various climate risks [2]. Effective risk management can only be achieved through the integration of this comprehensive knowledge.

According to the “Report of the open-ended intergovernmental expert working group on indicators and terminology relating to disaster risk reduction” [8], the term “multi-hazard” means the selection of multiple major hazards that a region faces and the specific contexts where hazardous events may occur simultaneously, cascadingly, or cumulatively over time, taking into account the potential interrelated effects. Despite the importance of analyzing multiple hazards for describing future disaster events in terms of their magnitude and probability [5], such an approach presents several challenges. According to Forzieri et al. [2], the study of multiple hazards poses two major challenges: hazards are not directly comparable, as their processes and description metrics differ, and hazards can interact, triggering cascading effects and coupled dynamics. The significance of examining multiple hazards is especially pronounced in coastal regions, as they are exacerbated by climate change [9, 10]. These areas are particularly exposed and susceptible to climate-related threats.

As with research centered on a single specific climate or weather hazard, the development of stand-alone applications for modeling risks was also common in recent years, particularly for earthquakes, tsunamis, landslides, and floods. More recently, there has been the development of a new generation of platforms in the field of natural risks based on open source code and WebGIS technologies [11]. The evolution of platforms toward open-source solutions is justified in research and development projects due to the principles of open research, where it is possible to access freely modifiable source code [12], while the adoption of WebGIS platforms has become commonplace because this technology deals with geographic information, including geospatial analysis, within the online environment [13].

To overcome all these challenges, a WebGIS application called Riskcoast was developed as part of the project with the same name. The project aimed to develop risk prevention cartographic tools to be applied in spatial planning and emergency planning. The main tools developed included the creation and updating of hazard, vulnerability and risk mapping in the SUDOE coastal regions, adapted to different future risk scenarios arising from climate change. The WebGIS platform aims to identify the exposed elements and carry out a current and future multirisk mapping assessment in climate change scenarios, for a set of processes that respond to climate drivers: landslides, flash flooding, estuarine flooding, coastal flooding, and coastal erosion. The main objective of this chapter is to present the main functionalities of the WebGIS Riskcoast platform and the multirisk assessment capabilities for different future risk scenarios arising from climate change applied to the case study of the municipality of Setúbal, Portugal.

This chapter is divided into five sections, besides this introduction. Firstly, the case study is presented, followed by the data used in the WebGIS platform. The methodology section includes the conceptual approach and multirisk workflow analysis, the conceptualization of climate change scenarios, and the platform architecture description. The results section illustrates its application to the municipality of Setúbal, demonstrating its value to address current and future multirisk analysis in climate change scenarios. The chapter ends with conclusions and notes for future developments.

2. Case study: Setúbal, Portugal

The study area used to test the WebGIS tool is the municipality of Setúbal, Portugal (Figure 1). The municipality of Setúbal is located in the Lisbon Metropolitan Area, Portugal. With a total area of 230 km², the municipality comprises five parishes: Azeitão, Gâmbia-Pontes-Alto da Guerra, Sado, Setúbal, and São Sebastião.

Figure 1.

Case study location.

The municipality of Setúbal is located on the north bank of the mouth of River Sado, bordering the Atlantic Ocean and flanked to the west by the Arrábida Mountain. In 2021, the municipality of Setúbal had around 123,000 inhabitants, about 16% of the total population of the Setúbal Peninsula. Due to its riverside location, its high population density (536 inhabitants/km²), the existence of a historic urban area, and strong industrial development, it is particularly vulnerable to the consequences of climate change.

3. Datasets

The variables used in the calculations of the Multihazard Risk Index for implementing current and future multirisk mapping assessment in climate change scenarios are listed inTable 1 and are available in the GitHub repository at https://github.com/nmileu/riskcoast/tree/main/data. The geographical unit of analysis was the census block, obtained from the National Statistics Institute [15]. All layers used in the application are in geoJSON format.

Dataset	Description	Source
Susceptibility | Flash floods	The layers corresponding to the climate hazards analyzed (flash floods, landslides, estuarine flooding, coastal flooding, and coastal erosion) were obtained in Zêzere et al.’s study [14] and constitute single layers in the system according to the current climate scenario (present-day; 2100 – RCP 4.5 and 2100 – RCP 8.5 and have been converted to GeoJSON format.	Zêzere et al. [14]
Susceptibility | Landslides
Susceptibility | Estuarine flooding
Susceptibility | Coastal flooding
Susceptibility | Coastal erosion
Exposure | Census blocks	The exposure analysis in the Multihazard Risk Index analysis component is based on calculating the population and number of buildings for each census block. These layers were also converted in the system to GeoJSON format.	National Statistics Institute - INE [15]
Exposure | Building centroids
Social vulnerability	The vulnerability component is a layer defined from the census blocks. This layer was also converted in the system to GeoJSON format.	Zêzere et al. [14]; Santos and Ferreira [16]

Table 1.

Source datasets used in Riskcoast WebGIS application. All data are available at https://github.com/nmileu/riskcoast/tree/main/data.

4. Methods

4.1 Risk analysis approach

The risk analysis carried out is based on the Multihazard Risk Index that combines the three main components of risk: susceptibility, exposure, and vulnerability. Zêzere et al. [14] presented a comprehensive description of susceptibility mapping in the technical report “Actual and Future Risk in Climate Change Scenarios Setúbal Test Site,” including detailed methodological expositions for the considered set of climate hazards: landslides, flash flooding, estuarine flooding, coastal flooding, and coastal erosion. As the above report includes all technical details, we will present here only the definitions and methodological outline of each component of the current and future multirisk mapping assessment model in climate change scenarios, describing, when appropriate, how the calculations were adapted for the WebGIS framework.

The analysis of climate hazards, now and in the future, was carried out for the set of physical processes with a relevant impact on the municipality of Setúbal and was supported by technical-scientific methods, adjusted to the municipal scale and to the available data. The susceptibility assessment was carried out independently for each type of climate hazard [14]. The mapping of actual hazards was carried out based on the direct delimitation of the areas affected by the hazardous processes considered or using indirect zoning methods, that is, quantitative or semi-quantitative methods based on analyzing the causes of the hazardous processes. Future climate hazards were estimated, whenever possible, quantitatively, based on the territorial incidence of the current hazard and its foreseeable evolution obtained from the projections adjusted to two state-of-the-art climate scenarios (Representative Concentration Pathways in case of limited (4.5) or no action (8.5) until 2100).

The exposure to hazards was considered to be the situation of people, infrastructure, housing, production capacities, and other tangible human assets located in hazard prone areas [8]. The exposure was considered on the platform through roadways, railways, residential buildings, resident population, and strategic, vital, and/or critical equipment. However, for calculating multihazard exposure, it was an option to consider only the number of buildings and the resident population. The buildings are represented geometrically by their centroid, corresponding to all buildings with total or partial residential function. The resident population per building was estimated by dasymetric analysis between the building layers and the census blocks of the 2011 census [17]. The assessment of current exposure in the calculation of the Multihazard Risk Index was carried out by intersecting the buildings and resident population with the hazardous areas corresponding to landslides, flash flooding, estuarine flooding, coastal flooding, and coastal erosion. In the context of assessing future exposure and calculating the Multihazard Risk Index, we decided to intersect the existing exposed elements (buildings and resident population) with the areas identified as hazardous at the end of the twenty-first century, in both the RCP 4.5 and RCP 8.5 scenarios. The results obtained through this approach emphasize the exposure of the elements currently present in the study area to the climate projected for 2100. This underscores the potential consequences for hazardous processes driven by climate factors if such conditions were to occur in the present day.

With regard to the vulnerability component, we chose to consider only the Criticality of Social Vulnerability. Criticality of Social Vulnerability is defined as the set of characteristics and behaviors of individuals that condition their propensity to suffer damage after the occurrence of a disastrous event [14]. These characteristics can contribute to the breakdown of the system and community resources that enable them to respond to or deal with catastrophic scenarios. In this work, we use the results obtained for the complete Metropolitan Area of Lisbon [16], using data from the 2011 census for the statistical census block disaggregation. Following the original work, it was possible to define an initial set of 45 variables, from which population density and building density were removed as they reflect exposure rather than vulnerability, leaving a total of 43 variables [14], representing the following domains: demographics, social support, the condition of built heritage, the economy, education, housing, family structure, employment, and health. In order to ensure comparison between territorial units of analysis, most of the variable data was expressed as a proportion, and social vulnerability was calculated using a principal component analysis (PCA).

4.2 Multihazard Risk Index

The Multihazard Risk Index (MRI) is dimensionless and results from the product of susceptibility (S), exposure (E), and vulnerability (V), using Eq. (1).

$$MRI=S^{\frac{1}{a}}\times E^{\frac{1}{a}}\times V^{\frac{1}{a}}$$E1

where ^a corresponds to the exponent of the Multihazard Risk Index.

The MRI, with due regard for differences in scale, risk components, and input data, is based on the INFORM risk index, which is an international reference risk index that combines data from 16 components describing hazards, exposure, vulnerability, and lack of coping capacity [18]. The formulation adopted in this work has recently been successfully applied at the municipal scale in Portugal, for flood risk [19], landslides [20], and wildfires [21].

The risk analysis, with the calculation of the MRI, was carried out for the present day and for the climate at the end of the twenty-first century, considering the RCP 4.5 and RCP 8.5 scenarios. The Territorial Unit (TU) used for analysis was the census block, as defined in the Geographical Base for Referencing Information of the National Statistics Institute [21]. Multihazard susceptibility was calculated using the intersection of the TUs with the areas prone to be affected by each of the considered climate hazards: landslides, flash flooding, estuarine flooding, coastal flooding, and coastal erosion. The exercise was carried out for three scenarios, corresponding to the actual situation (scenario 1) and the climate at the end of the twenty-first century, for RCP 4.5 (scenario 2) and RCP 8.5 (scenario 3). For each UT, and for each scenario, multihazard susceptibility was calculated by adding up the percentage of area affected by each of the processes considered, using Eq. (2).

$$STUi=\sum _{i=1}^{n}ZiCHi$$E2

where CHi is climate hazard i (landslides, flash flooding, estuarine flooding, coastal flooding, and coastal erosion) and Zi is the percentage of Territorial Unit i intersected by each hazard.

As the processes involved in landslides and coastal erosion along coastal cliffs are the same, the corresponding values in Eq. (2) have only been counted once in cases where there is a spatial overlap between these two types of hazard.

Multihazard exposure was calculated from the intersection of the TUs with the buildings exposed to each of the climate hazards considered landslides, flash flooding, estuarine flooding, coastal flooding, and coastal erosion. The population living in these buildings was estimated by dasymetric mapping based on data from the 2011 censuses. As in the case of susceptibility, the exercise was carried out for three scenarios: current situation (scenario 1) and the climate at the end of the twenty-first century, for RCP 4.5 (scenario 2) and RCP 8.5 (scenario 3).

For each TU, and for each scenario, multihazard exposure was calculated by adding up the resident population exposed to each of the physical processes considered, according to the Eq. (3).

$$ETUi=\sum _{i=1}^{n}PiCHi$$E3

where Ri is the resident population in the buildings in Territorial Unit i intersected by each climate hazard.

The vulnerability considered in Eq. (1) corresponds to criticality and was calculated for each TU independent of susceptibility and exposure, as described in Section 2.1.

In the end, before integration into the Multihazard Risk Index, the three components of the MRI (susceptibility, exposure and vulnerability) were scaled to the interval [0, 1] using the min-max method (4).

$$X_{i,norm}^m=\frac{X_i^m-X_{i,\min }}{X_{i,\max }-X_{i,\min }}$$E4

The values of 𝑋_{𝑖,m𝑎𝑥} and 𝑋_{𝑖,𝑚𝑖𝑛} were determined for each component of the MRI, taking into account the full range of values obtained for the three scenarios considered.

Finally, it should be noted that, as the integration process is multiplicative, the MRI is equal to zero whenever any of the three components that define it (susceptibility, exposure, and vulnerability) is equal to zero.

4.3 Combined hazard selection

Multihazard susceptibility and exposure are calculated by intersecting the TUs with the areas prone to be affected by each of the climate hazards: landslides, flash flooding, estuarine flooding; coastal flooding; and coastal erosion. As this is a multi-hazard approach, the selection of the processes to be considered in the MRI calculation is a user option in the WebGIS framework, with several possible combinations. The user can calculate the MRI using a single climate hazard or any combination of several climate hazards. The number of possible combinations without repetition for the five processes was determined from expression (5).

$$C_r^n=\frac{n!}{r!\left(n-r\right)!}$$E5

where $C_r^n$ = number of combinations, n = total number of objects in the set, and r = number of choosing objects from the set.

In this way, the user can choose between 32 combinations of processes for one actual and two future climate change scenarios, making it possible to obtain a total of 96 multirisk mapping assessments.

4.4 Riskcoast WebGIS development and architecture

The Riskcoast WebGIS application was written in Javascript, using HTML, CSS, and Javascript files (Figure 2) directly from a repository on GitHub (https://github.com/nmileu/riskcoast), with instructions in this version only in Portuguese. The mapping component was implemented using Leaflet open-source JavaScript library, with base maps provided by OpenStreetMap. The spatial analysis tasks, namely, the spatial intersections between hazard prone-areas and exposed elements, were carried out using the Turf Javascript library. This is a modular library for geospatial analysis in browsers, using the GeoJSON data format, which has the advantage of speed and does not require to send data to a WebGIS or database server.

Figure 2.

Riskcoast WebGIS architecture.

For the Multihazard Risk Index calculation, the processing is done through Javascript functions. The choice of Javascript in the processing was guided by the ease of implementation and the processing speed as most of the code runs in the user’s browser. To represent geographic layers with their non-spatial attributes, GeoJSON was the chosen format. Despite the limitations, the format allows the representation and parsing of simple geographical features like the layers used in Riskcoast project in a simple and human-readable format.

The WebGIS tool can be accessed via a web browser on a desktop or mobile device at https://nmileu.github.io/riskcoast/. The site has the main map area where the user can view the layers available in the Tree Layers Control, namely, the layers of social vulnerability, exposure, hazard (susceptibility), and risk. In the main map, the user can zoom or pan around the map and visualize the layer classification for each census block (Figure 3).

Figure 3.

Riskcoast WebGIS interface.

The main map has four main tabs (‘Project description’, ‘Exposure analysis’, ‘Multi-Hazard Risk analysis’, and ‘Legends’). The first tab gives the description of the “Riskcoast” research project. The next tab, the Exposure analysis tab, controls the exposition scenarios. The first step is to select the climate scenario (current or future), followed by one of the climate hazards for which the user want to intersect with the exposed elements (roads, railways, buildings, population, and strategic, vital, and/or critical equipment). In the next tab, the user can select the climate scenario (current or future, RCP 4.5, RCP 8.5) and the combination of climate hazards for which the MRI is computed at the census block scale. The last tab is dedicated to showing the legends for the layers.

5. Results

At this point, we present two applications (current and future multirisk analysis in climate change scenarios) in which we have calculated the risk using different process combinations: climate hazards associated with precipitation (landslides and floods), climate hazards associated with the rise of sea level (estuarine flooding, coastal flooding, and coastal erosion), and all climate hazards.

5.1 Current multirisk analysis in climate change scenarios

Figure 4 shows the mapping of the Multihazard Risk Index in the municipality of Setúbal, obtained using the Riskcoast WebGIS application, for current landslides and flash floods. In this figure, it is possible to identify a large number of territorial units with a high risk (MRI > 0.4), all located in the Setúbal city center and essentially exposed to the flash flooding hazard.

Figure 4.

Current multirisk mapping assessment for landslides and flash floods.

Figure 5 shows the mapping of the Multihazard Risk Index in the municipality of Setúbal, obtained using the Riskcoast WebGIS application, for current estuarine flooding, coastal flooding, and coastal erosion. In this figure, it is possible to identify a group of territorial units with a medium risk (MRI > 0.2), located in the parish of Sado (Mitrena Industrial Zone) and essentially exposed to estuarine flooding.

Figure 5.

Current multirisk mapping assessment for estuarine flooding, coastal flooding, and coastal erosion.

Figure 6 shows the mapping of the Multihazard Risk Index in the municipality of Setúbal, obtained using the Riskcoast WebGIS application, for current landslides, flash floods, estuarine flooding, coastal flooding, and coastal erosion, where the city center and the Mitrena industrial zone stand out as risk areas.

Figure 6.

Current multirisk mapping assessment for all climate hazards.

5.2 Future multirisk analysis in climate change scenarios

Figure 7 shows the mapping of the Multihazard Risk Index in the municipality of Setúbal, obtained using the Riskcoast WebGIS application, for landslides and flash floods for the end of the twenty-first century and considering the worst climate scenario (RCP 8.5). The figure highlights a large number of territorial units with a high risk (MRI > 0.4), located in Setúbal city center and essentially exposed to the flash flooding hazard.

Figure 7.

Future (RCP 8.5) multirisk mapping assessment for landslides and flash floods.

Figure 8 shows the mapping of the Multihazard Risk Index in the municipality of Setúbal, obtained using the Riskcoast WebGIS application for estuarine flooding, coastal flooding, and coastal erosion for the end of the twenty-first century and considering the worst climate scenario (RCP 8.5). The figure highlights a group of territorial units with a high risk (MRI > 0.4), located in the riverside area and in the inner area of the Sado River estuary.

Figure 8.

Future (RCP 8.5) multirisk mapping assessment for estuarine flooding, coastal flooding, and coastal erosion.

Figure 9 shows the mapping of the Multihazard Risk Index in the municipality of Setúbal, obtained using the Riskcoast WebGIS application, for the end of the twenty-first century (RCP 8.5) and accounting the complete set of considered climate hazards: landslides, flash floods, estuarine flooding, coastal flooding, and coastal erosion. The multihazard risk analysis for the end of the century, with the climate conditions defined by the RCP 8.5 scenario and considering the elements currently exposed in the municipality, shows an increase in risk. There was no reduction in the Multihazard Risk Index in any of the municipality’s territorial units. In addition to those that already stand out in Setúbal’s city center on the map showing the current situation, there are two more territorial units, located in Setúbal’s riverside area and in the inner area of the Sado river estuary, reflecting the increased susceptibility and exposure to the coastal and estuarine flooding processes.

Figure 9.

Future (RCP 8.5) multirisk mapping assessment for all climate hazards.

6. Conclusions

The development of the WebGIS tool was based on a simple architecture and open-source components that made it possible to implement a solution that does not require sending data to a WebGIS server or database server. Considering the large number of possible combinations of climate hazards in the definition of multirisk scenarios arising from climate change, a critical aspect considered in the development of the platform was ease of use, requiring no advanced knowledge in geographic information systems. The ability to visualize and explore current multirisk mapping assessment models in climate change scenarios regarding the different risk components, namely, susceptibility, exposure, and vulnerability, allows highlighting areas with lower or higher risk. On the other hand, the ease in quantitatively assessing risk at the census block by creating future multirisk mapping assessment models in climate change scenarios through a structured and systematic process provides opportunities for the evaluation of different strategies and actions to minimize the adverse impacts of any hazardous event.

Analyzing the multihazard risk, quantified using the Multihazard Risk Index for the present-day and the future (for the RCP 8.5 scenario), it shows that the risk at the end of the twenty-first century will increase significantly. The current risk is maximum in the downtown area of Setúbal as a result of flash floods and will continue to be so in the future. However, the risk will increase in the entire area of the Sado estuary, particularly on the waterfront of the city of Setúbal and in the inland surroundings of the Sado estuary, particularly in the parish of Praias do Sado.

Although the WebGIS Riskcoast platform is associated with a European project and has a defined time-bound, it is possible to identify a number of improvements and future developments. Assessing exposure at the end of the twenty-first century is a complex exercise, mainly due to the enormous uncertainty about population numbers and their distribution. One of the developments in the multirisk assessment model is the possibility of including demographic projections, despite the complexity of spatializing the demographic projection at the census block scale. In functional terms, the interactive delimitation of the geographical area for carrying out the risk analysis, the comparative analysis of scenarios, and the implementation of advanced reports is another evolution of the platform that will make it possible to improve the use multirisk mapping assessment models in climate change scenarios.

To conclude, despite the limitations and possibilities for improvement identified for future development of the WebGIS platform, the implementation of the prototype in the context of the research project allowed to demonstrate that a multirisk assessment can be obtained quickly and easily for any combination of climate hazards for a study region particularly susceptible to climate change, providing a high added-value contribution to supporting local adaptation strategies to increase municipalities resilience to climate change.

Acknowledgments

This work is part of the project Riskcoast—Development of tools to prevent and manage geological risks on the coast linked to climate change (SOE3/P4/EO868, Interreg Sudoe).

Conflict of interest

The authors declare no conflict of interest.