A coastal flooding database from 1980 to 2018 for the continental Portuguese coastal zone

Applied Geography 135 (2021) 102534 Available online 18 August 2021 0143-6228/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). A coastal flooding database from 1980 to 2018 for the continental Portuguese coastal zone Alexandre Oliveira Tavares a , Jos´ e Leandro Barros b , * , Paula Freire c , Pedro Pinto Santos d , Luís Perdiz b , Andr´ e Bustorff Fortunato c a Earth Sciences Department and Centre for Social Studies, Universidade de Coimbra. Rua Sílvio Lima, P´ olo II, 3030-790, Coimbra, Portugal b Centre for Social Studies, Universidade de Coimbra, Portugal. Col´ egio de S. Jer´ onimo, Apartado 3087 3000-995, Coimbra, Portugal c Estuarine and Coastal Zones Unit/National Civil Engineering Laboratory, Av. do Brasil 101, 1700-075, Lisboa, Portugal d Centre for Geographical Studies, Institute of Geography and Spatial Planning, Universidade de Lisboa, Lisboa. Edifício IGOT, Rua Branca Edm´ ee Marques, Cidade Universit´ aria, 1600-276, Lisboa, Portugal ARTICLE INFO Keywords: Coastal zone Coastal flooding Database Impacts Occurrences ABSTRACT Continental Portugal presents an extensive and diversified coastal zone which concentrates the main public and private infrastructures of the different economic sectors, as well as the main critical infrastructures. This area is also characterized by a high population density, being a differentiated territory in geophysical, biological and landscape terms. The wave regime is highly energetic, and storms are frequent. In the last decades, the coast of continental Portugal has been affected numerous times by overtopping and coastal flooding processes. Identifying the critical coastal typologies affected by flooding can contribute to a comprehensive flood risk management framework for the Portuguese coastal zones. Hence, a historical database of coastal flooding occurrences was created for the period 1980–2018 based on national and regional newspapers. For this period 650 occurrences were identified as well as 1708 impacts associated with them. In terms of impacts, the typologies associated with public areas, human impacts, the natural system, environmental degradation and buildings stand out. Results provide relevant temporal and spatial information about coastal historical flood occurrences related to extreme storm events and associated impacts, and contribute to the design of a risk framework. 1. Introduction Coastal flooding and its associated impacts have become a growing concern in recent decades, as a result of the increasing exposure and changes in the hazard forcers throughout the 20th and 21st centuries (Nicholls & Cazenave, 2010; Weisse et al., 2014; Silva et al., 2017). Among other natural hazards, coastal flooding is responsible for some of the worst human and economic losses worldwide (Kron, 2013). According to the Intergovernmental Oceanographic Commission of UNESCO (IOC/UNESCO et al.,), 40% of the world’s population and most economic activities are located within 100 km from the coastline, while in the European Union (EU) 86 million people (19%) live within 10 km of the coastline (EEA, 2006). In Portugal, the coastal zone stands out as an extremely important area, where ¾ of the population and 80% of Gross Domestic Product (GDP) are located (Santos et al., 2017), and where the risk of sea-level rise (SLR) is high (Veloso-Gomes et al., 2004; Antunes & Taborda, 2009; Rocha et al., 2020). All the projections of the Intergovernmental Panel on Climate Change (IPCC, 2014) predict the continuous rise of the mean sea level and the worldwide increase in storminess. This fact puts several coastal zones at high risk of flooding, including densely populated and economically vital areas. However, coastal areas face not only an increase in SLR but also a range of non-climate-change related factors that contribute to the increase of their fragility. Among them, natural and artificial factors stand out, such as: storm surges, sediment deficit, shoreline retreat; increase in anthropogenic pressure translated by the construction of different types of artificial infrastructure such as harbours, defence structures, urbanization and touristic areas and land use change. The factors mentioned above, combined with the growth of migration to coastal areas, industrialization, urbanization processes, and with the increase in the intensity and frequency of extreme events related to climate change (Bertin et al., 2013), enhances the exposure * Corresponding author. Col´ egio de S. Jer´ onimo, Apartado 3087 3000-995, Coimbra, Portugal. E-mail addresses: [email protected] (A.O. Tavares), [email protected] (J.L. Barros), [email protected] (P. Freire), [email protected] (P.P. Santos), [email protected] (L. Perdiz), [email protected] (A.B. Fortunato). Contents lists available at ScienceDirect Applied Geography journal homepage: www.elsevier.com/locate/apgeog https://doi.org/10.1016/j.apgeog.2021.102534 Received 22 September 2020; Received in revised form 13 May 2021; Accepted 5 August 2021 Applied Geography 135 (2021) 102534 2 and susceptibility to coastal flooding (IPCC, 2014; Neumann et al., 2015; Rilo et al., 2017). Several authors have collected historical information related to different natural hazards from newspapers, technical reports and scientific articles, in order to analyse and evaluate past occurrences and impacts associated with natural hazards, and assess future dynamics (Barriendos & Rodrigo. 2006; Raska & Emmer, 2014; Ruocco et al., 2011; Santos et al., 2014). This historical information is often subsequently organized in databases that allow its systematization, treatment and analysis. At a global level, there is a wide range of disaster databases that are distinguished by their temporal and spatial factors, as well as by the types of disasters and insertion criteria considered. Examples of global databases include the EM-DAT from Centre of Research on Epidemiology of Disasters (EM-DAT, 2013), the DesInventar database (La Red, 2009) and the database from Munich Re named NatCatSERVICE (Munich Re, 2011). At national level, some examples can also be mentioned, such as the Spanish Catalonia flood damage database (Barnolas & Llasat, 2007), and the Italian AVI project (Guzzetti & Tonelli, 2004). Specific examples for coastal flooding include a UK database from 1915 to 2016 generated within the project SurgeWatch (Haigh et al., 2017), the US SURGEDAT (Needham et al., 2013) and the Australian Database (Callaghan & Power, 2014). In Portugal, the DISASTER hydro-geomorphologic database, with data between 1865 and 2010, stands out (Zˆ ezere et al., 2014). The Portuguese coastline presents a great diversity of morphosedimentary systems such as estuaries, lagoons, barrier islands, beaches, dunes, and cliffs (Ferreira & Matias, 2013; Ponte Lira et al., 2016). This area is also characterized by a multiplicity of land uses, occupations and activities that make it an area of strategic importance at the economic, social and environmental levels. The growth of human settlements in the coastal zone over the last 7 decades, combined with oceanographic and atmospheric forcers and distinct geological and morphological contexts contributes decisively to understand the evolution and current configuration of the coastal zone. The Continental Portuguese coastal zone is also characterized by an asymmetry in terms of wave direction and energy (Andrade & Freitas, 2002). The Atlantic western coast is characterized by high-energetic waves with dominant northwest swell, contributing to high sedimentary transport values. The south coast, although also Atlantic, is protected from the northwest swell and presents a more moderate wave energy (Ferreira & Matias, 2013). The coastal zone is threatened by several hazards, with emphasis on overtopping and coastal flooding, cliff instability, and coastal erosion (Santos et al., 2017; Veloso-Gomes, 2007). Motivated by the absence of a consolidated national database of coastal floodings and their impacts, a historical database for the continental Portuguese coastal zone between 1980 and 2018 was compiled. This database is part of an effort to develop an innovative methodology to support flood risk management, supported by a better ability to forecast overtopping and flooding occurrences in different coastal typologies. The development of a loss and damage database related to natural disasters is important for risk assessment and regional and local management (Freire et al., 2016; Santos et al., 2014). According to Devoli et al. (2007), the development of this kind of databases is crucial for risk management because they allow the identification and analysis of the relationship between occurrences of disasters, the vulnerable elements, and the respective human and material losses. The main objectives of this study are: a) to present the methodological aspects related to coastal flooding data collection and the construction of the MOSAIC database; b) to present and discuss the spatial and temporal distribution of the coastal flooding occurrences and associated impacts; c) Identify the main oceanographic conditions along the coast, namely water level and wave conditions. 2. MOSAIC database methods The absence of a consolidated coastal flooding database for the Continental Portuguese coast, combined with the need for the identification and selection of the coastal typologies most affected by floods, led to the creation and development of the database here presented. The hemerographic analysis covers the period 1980–2018 in national and regional newspapers. The general methodology followed to generate the MOSAIC database and its analysis is presented in Fig. 1. 2.1. Key concepts Before presenting the main methodological aspects underlying the construction of the MOSAIC database, the main concepts that will be present throughout this study should be clarified. The hazardous processes in this database include coastal flooding and overtopping. The concept of “occurrence” followed the definition of Zˆ ezere et al. (2014) with the necessary adaptations. Occurrence is considered as a specific case related to overtopping or coastal flooding related to a unique spatial location and a specific time period. The concepts of loss (also referred to as human impacts) and damage must also be distinguished. In this study, “loss” or “human impacts” represents the direct hazardous effects on humans, including casualties, injuries, missing, evacuations, and permanent displacement (Santos et al., 2014); “damage” represents the material consequences for any type of facility or property, e.g. road network, buildings, farmland (Santos et al., 2014). 2.2. Database data collection The main source of information for the MOSAIC database were national and regional newspapers. A database of this type can be constructed using a wide range of sources. However, hemerographic analysis presents a set of favourable factors such as the broader coverage of occurrences on a local scale and the same occurrence is, in most cases, reported in different newspapers, allowing a more accurate analysis. The ease of access to newspaper archives, and the broader temporal analysis due to its greater coverage over time are other favourable and preferential factors of the hemerographic analysis over other data sources (La Red, 2013). The information taken from the hemerographic analysis provides a varied set of information. From the start, it allows obtaining quantitative information as to human losses and the number of affected buildings, but also provides qualitative information such as the type of damage observed (Rilo et al., 2017). The starting point was the selection of the different hemerographic sources to be analysed and that form the basis of the data collection, in order to construct a coastal flooding database for the Portuguese continental coast. The selected newspapers took into account two criteria related to their spatial and temporal coverage: a) the newspaper must have been published continuously for at least 20 years and b) the selected newspaper should guarantee a good national and regional distribution of the news. A total of eight newspapers were analysed according to systematic and punctual analysis, as described below (Table 1). First, daily national newspapers were selected. The first newspaper looked over in a systematic way – i.e., all editions were consulted, either of dates where coastal flooding conditions existed, or of any other dates – was Público, which presents two editions with a different regional scope: one focused in the northern Portugal (Porto’s edition) and other focused in the central and southern Portugal (Lisbon’s edition). The two editions differ in that fact each one includes a local segment with specific information regarding its territorial focus. However, because the Público newspaper was only founded in the early 1990’s, the period between 1980 and 1990 was systematically explored using the Jornal de Notícias. In the punctual analysis (only dates from the hindcast and dates with occurrences resulting from systematic analysis are analysed), one national and five regional newspapers from different areas were considered (Table 1), published in different regions of the country, in order to ensure a wide regional coverage. The dates to consider in the punctual analysis were based on the dates taken from the systematic analysis of the newspapers Público and Jornal de Notícias, as A.O. Tavares et al. Applied Geography 135 (2021) 102534 3 well as on dates in which ocean conditions were favourable to the occurrence of storms (hindcast). Regarding this last criterion, the conditions were determined through hindcast simulations described in section 2.4. The data resulting from the hindcast allowed the identification of 67 new dates and 165 occurrences, representing 25% of total occurrences. During the punctual analysis, every six years a systematic analysis (all daily editions of the newspaper were consulted) of the newspaper was carried out in order to validate the applied methodology. From the newspapers with regional coverage, one covers the northern coastal area, two the centre coastal area, and the other two the area located south of Lisbon. In order to validate the hemerographic analysis, a search for occurrences of coastal floodings in technical reports and scientific articles was also performed. The results led to the identification of new occurrences, providing accurate information for its georeferencing. All occurrences were validated by crossing the information obtained in more than one newspaper, as well as in academic works and technical reports. To ensure the consistency of the database, the process of hemerographic searching, insertion of information in the database and occurrences georeferencing was carried out by a single person. 2.3. Database structure and temporal and spatial incidence The MOSAIC database consists of six topics subdivided into different fields that identify and characterize each occurrence in its multiple strands (Fig. 2). Based on the sources identified above, a coastal flooding database presently containing 650 occurrences (see definition of occurrence in section 2.1) was built. The MOSAIC is an adaptable and georeferenced historical database with alphanumeric open and closed fields of the numeric and textual types. Each occurrence inserted in the database is subsequently georeferenced using a point shapefile, based on the information from the analysed source, using the ArcGIS 10.5 software. The PT-TM06/ETRS89 projected coordinate system was adopted. Georeferencing is based on satellite images provided by the Esri® World Imagery service. The location accuracy considered was divided into four classes (Zˆ ezere et al., 2014): 1) location with exact coordinates (scale 1:1000); 2) location based on local toponymy (scale 1:10,000); 3) location in the centroid of the parish (only used when no other location besides the parish is mentioned); 4) location in the centroid of the municipality (idem regarding the mention to the municipality). The temporal incidence of the MOSAIC database ranges from 1980 to 2018. The spatial incidence is the entire Continental Portuguese coastal zone, extending from the mouth of the Minho River, in the northwest, to the mouth of the Guadiana River in the southeast, totalling about 987 km (Fig. 3). Throughout the work, the density of occurrences per km 2 will be Fig. 1. Scheme of the methodology used in the study. Table 1 Newspapers used for data collection. Title Periodicity of publication Type of analysis performed Publication coverage Coverage period Público Daily Systematic National 1990–2018 Jornal de Notícias Daily Systematic National 1980–1989 Di´ ario de Notícias Daily Punctual National 1980–2018 Di´ ario do Minho Daily Punctual Regional 1980–2018 Di´ ario de Aveiro Daily Punctual Regional 1985–2018 Di´ ario de Leiria Daily Punctual Regional 1987–2018 Di´ ario do Sul Daily Punctual Regional 1980–2018 Setubalense Tri-weekly Punctual Regional 1980–2018 A.O. Tavares et al. Applied Geography 135 (2021) 102534 4 Fig. 2. Topics and main fields of MOSAIC database. Fig. 3. Location of the coastal area covered by the MOSAIC database: a) Continental Portugal; b) Population density of Continental Portugal by municipality. (color should be used for any figures in print). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) A.O. Tavares et al. Applied Geography 135 (2021) 102534 5 calculated. The calculation is based on an area of 500 m defined from the highest astronomical tide line up to 500 m inland, according to the definition present in the National Strategy for the Integrated Management of the Coastal Zone (ENGIZC, 2009). This delimitation has also taken into account the current historic occurrence as well as the definition of the safeguard strips for overtopping and coastal flooding present in the approved POC. For the Land Use Land Cover (LULC), two exploratory analysis were carried out considering different territorial areas. The first considers an area of 500 m defined from the highest astronomical tide line up to 500 m inland (referred to in the previous paragraph), called Coastal border. The second is called the Coastal Zone, according to ENGIZC (2009) and extends for 2 km inland, including the Coastal border area. Both analyses are based on the Land Use and Occupation Cartography of 1995, 2007, 2010, 2015, and 2018 provided by the Territory’s General Directorate (DGT, nd), with no cartography for previous periods. 2.4. Hindcast of oceanographic variables The information on occurrences was complemented by multidecadal time series of oceanographic conditions along the coast. Time series were generated for both the water level (due to tides and storm surges) and wave conditions. The time series of water levels were obtained by adding tidal and surge signals. Tides were computed through harmonic synthesis based on the results of a regional tidal model (Fortunato et al., 2016, 2019). This model simulates the generation and propagation of tides in the North-East Atlantic Ocean. Along the Portuguese coast, root mean square errors are of the order of 5 cm. Results were extracted at 7 representative points along the western Portuguese coast. Surges were estimated based on the inverse barometer effect, i.e., assuming that only atmospheric pressure contributes to the surge. This approximation is justifiable in the Portuguese coast (Fanjul et al., 1998) because the wind-generated surge is limited by the small width of the continental shelf. Time series of atmospheric pressure were extracted from the ERA-Interim reanalysis (Dee et al., 2011). The wave characteristics along the western Portuguese coast were determined using the wave model WaveWatch III (Tolman et al., 2009, p. 194) applied to the North Atlantic Ocean, with a nested grid on the Portuguese shelf. The model was run from 1979 to 2018, forced by winds from the ERA-Interin reanalysis. Further details on the application of the model are provided in Fortunato et al. (2017). Significant wave heights (Hs) were stored at hourly intervals. Subsequently, an analysis of the time series of oceanographic conditions was carried out, considering the daily maxima Hs and sea level (Sl) to perform the punctual analysis related to hemerographic research and to analyse the relationship between overtopping and flooding forcings and the number of occurrences. The parameters used in the present study as representative of the overtopping or coastal flooding forcing conditions are: the daily sea level maxima (Slmax) and the daily maximum significant wave height (Hsmax). 3. Results 3.1. Spatial and temporal distribution The hemerographic analysis allowed the identification of 650 occurrences of coastal flooding and overtopping between 1980 and 2018, with high temporal and spatial variability. However, some geographic hotspots with a high number and density of occurrences can be identified (Fig. 4). Two distinct sectors can be recognized: the first encompasses all coastal municipalities to the north of the Tagus River, which concentrates 77% of the occurrences; the second encompasses the remaining coastal municipalities and represents 23% of the occurrences. Considering the planning and management instruments (Spatial Coastal Zone Plans - POC), that divide the coastal area in six different sectors, results show that 85% of the occurrences are located in the three northernmost sectors (Fig. 4b). The analysis also shows that only 9 out of 54 coastal municipalities Fig. 4. Spatial distribution of the occurrences: a) Continental Portugal; b) Occurrences in the different POC. (color should be used for any figures in print). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) A.O. Tavares et al. Applied Geography 135 (2021) 102534 6 do not exhibit any occurrence (Fig. 4a). However, the absence of occurrences in a given area does not mean that it is not affected by flooding or overtopping. This absence can be explained either by incomplete documentary coverage of certain areas or by the characteristics of the affected areas, namely areas that are predominantly natural and with little anthropic occupation, that receive less interest in the news coverage. Fig. 5a shows the five municipalities with the highest number of occurrences (see location in Fig. 4a), both in the entire period, and on each decade. For the period between 1980 and 2018, except for Almada (belonging to the POC Alcobaça-Cabo Espichel), all the municipalities are located north of the Mondego River, belonging to the POC CaminhaEspinho (Porto and Vila Nova de Gaia) and Ovar-Marinha-Grande (Ovar and ´ Ilhavo). This spatial pattern is practically unchanged when the different decades are analysed separately. The ‘80s stands out for presenting a greater spatial variability related to the occurrence of flooding and overtopping. Another fact to highlight is related to the municipality of Porto, which between 1980 and 2018 is the third municipality with the highest number of occurrences. However, these occurrences are essentially concentrated in the ‘80s and ‘90s. Since the end of the ‘90s, the number of occurrences in the municipality of Porto decreased significantly, being this fact related to a set of coastal defence works carried out during the ‘90s. In contrast, in the municipalities belonging to the metropolitan area of Lisbon, in particular Almada, Cascais, and Oeiras, the number of occurrences increased significantly since 2000. With regard to the density of occurrences, with the exception of the municipality of Almada, all other municipalities are located north of the Tagus River (Fig. 5b). The municipality of Porto stands out as having the highest density of occurrences (24.32 occurrences per km 2 ). Fig. 6 shows the monthly and annual distribution of occurrences. Regarding the monthly distribution, the vast majority (93%) occurs during the so-called maritime winter (from October to March) with a clear maximum for January, with 51% of the total occurrences (Fig. 6a). Storm Hercules, which affected mainland Portugal between 3 and 7 January 2014, was responsible for 56% of the total occurrences in January between 1980 and 2018. As for the annual distribution of occurrences (Fig. 6b), the years 2014, 1996, and 2010 stand out, concentrating 56% of the total between 1980 and 2018. As previously mentioned, in 2014 the storm Hercules was responsible for 75% of the total occurrences identified in 2014 and for 37% in the period 1980–2018. The occurrences resulting from the storm Hercules had a great spatial dispersion with a special focus on the west coast. In 1996, most of the occurrences refer to a period of storms between the 1st and the 14th of January dispersed through the western and southern coast. As for 2010, most of the occurrences are due to the storm Xynthia (27th – 28th February) as well as to a stormy period that occurred on the 8th of October. The first event caused occurrences mainly on the south coast, while the second gave rise to a greater concentration of occurrences between the municipalities of Porto and Figueira da Foz. Fig. 6b shows the temporal evolution of the occurrences, where it is possible to observe an increase in the number of occurrences between 1980 and 2018, with different rhythms. The ‘80s were marked by growth and subsequent stabilization in the number of occurrences, representing 8% of the total. The ‘90s represent 27%, being a decade marked by an initial increase in the number of occurrences followed by a stabilization period that is interrupted by a slight increase in 1994 followed by a more pronounced growth caused by 1996 storms. The first decade of the 21st century represents 13% of the occurrences and is characterized by a period with a slight but constant increase, with 2003 standing out as the one with the highest number of occurrences. The last Fig. 5. Municipalities with higher number (a) and density of occurrences (b) in the analysed decades. A.O. Tavares et al. Applied Geography 135 (2021) 102534 7 period analysed (2010–2018) less than ¼ of the database’s time frame – concentrates 53% of the total number of occurrences and is clearly the period characterized by a huge growth in the number of occurrences, especially in the years 2010 and 2014. This increase is related to a period in which the Portuguese continental coast was hit by a set of storms, of which Xynthia (February 2010), H´ ercules (January 2014) and Emma (March 2018) stand out, as well as, a set of stormy periods such as the beginning of October 2010 and February 2011. This could be related to the increase in storminess associated with climate change. According to Santos et al. (2017), the main impacts of climate change that will affect the coastal zone of mainland Portugal will be related to: a) the increase in the global average sea level which will cause a greater frequency of extreme values of sea level; b) the rotation of the direction of the waves on the west coast; c) the change in the storm system with a possible increase in frequency and intensity of extreme weather and climate events. 3.2. Losses and damages The 650 occurrences of coastal flooding and overtopping between 1980 and 2018 resulted in a total of 1708 losses and damages spread over 7 categories and 56 typologies. Fig. 7 presents the categories identified, as well as the proportion that each typology represents in each of them in terms of number of occurrences. As some typologies have a residual value, it is important to highlight in each category of losses and damages the main types of resulting impacts. Fig. 8 shows all the loss types, with the displaced and evacuated people representing 85% of the human impacts. Regarding the damages related to the natural system and environmental degradation, they are fundamentally related to the sand/dune system (Fig. 8), representing 93% and 67% of the total of each category, respectively. Damage to urban streets and beach walkways accounts for 57% of the damage in public areas. Regarding the infrastructure category, coastal infrastructure damages stand out (73% of the total category). With regard to buildings, the commercial type stands out, followed by residential and beach support buildings (cabinets, sanitary facilities, bars, etc.), which together account for 86% of the affected buildings. Fig. 9 presents all losses and damages identified and their temporal distribution. There is a predominance of damages in public areas, followed by human impacts, environmental degradation, and natural system damages. The temporal distribution of the losses and damages show that the 1980s were marked by damages associated with public areas, as well as related to environmental degradation and the infrastructures category. The following decade was characterized by a general increase Fig. 6. Temporal distribution of occurrences: a) Monthly distribution; b) annual distribution. A.O. Tavares et al. Applied Geography 135 (2021) 102534 8 in all categories of impacts. It is a period marked by human impacts, that emerge as the category with the highest number of occurrences, highlighting the 73 evacuated people identified in the municipality of Ovar. Also, the damage to buildings grows significantly, particularly in commercial (23) and residential (18) buildings. Between 2000 and 2009 the impacts decrease relative to the previous period, as well as the number of occurrences. This decade is marked by a balance between the different categories, highlighting the damages to the natural system, mainly related to changes in the sand and dunes system. The period between 2010 and 2018 is marked by a sharp growth in losses and damages, which translates into a growth of 367% compared to the previous decade. Damage to public areas is the category with the highest number of observed impacts, with particular emphasis on damage to walkways, urban streets, and urban furniture. The human impacts category should be highlighted, where 121 displaced people were registered, particularly in the southern municipality of Olh˜ ao. These human impacts were mostly verified during January and February 2010, when a stormy period occurred, highlighting the storm Xynthia (27th – 28th February). It is also important to highlight the damage related to the environmental degradation and natural system, with emphasis on the sand and dunes system, causing in many cases important decreases in the sand and retreat in the dune cord. Finally, damage to buildings is highlighted, with emphasis on commercial, residential, and beach support buildings, as well as damage to coastal protection infrastructures. Fig. 10 presents the spatial distribution of the occurrences, Fig. 7. The categories and typologies of losses and damages differentiated by colors. The size of the polygons is proportional to the number of occurrences in each category. (color should be used for any figures in print). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) A.O. Tavares et al. Applied Geography 135 (2021) 102534 9 considering the different POC. Results show that the areas north of Cabo Espichel (see Fig. 4b) suffer more losses and damages than the other areas to the south. The exception is the POC Vilamoura-Vila Real Santo Ant´ onio with the highest number of human impacts (115 displaced), for which the municipality of Olh˜ ao contributes the most. In all areas, the period between 2010 and 2018 is the one with the highest number of losses and damages. Damages in public areas stand out as the category with the highest number of occurrences in almost all areas. In EspichelOdeceixe and Vilamoura-Vila Real Santo Ant´ onio damages in buildings and human impacts are, respectively, the most frequent categories. In the areas north of Cabo Espichel, damages related to the natural system and environmental degradation are the most important. Globally, with the exception of the 2000–2009 decade, results show an increase of the losses and damages and diversity across all areas between 1980 and 2018. If we take into account the municipal scale, namely the five municipalities with the highest number of occurrences, there are also different trends (Fig. 11). In the municipality of Porto, most of the impacts (65%) were observed between 1980 and 1999, with emphasis on the damage in public areas and related to environmental degradation. However, during the Hercules storm in 2014, the human impacts, such as evacuations and injuries, are the most relevant ones. Vila Nova de Gaia and ´ Ilhavo show a temporal distribution of impacts different from the other municipalities, as 78% and 70%, of losses and damages, respectively, arose between Fig. 8. Principal losses and damages in each category. Human impacts are expressed by the number of affected people. The other categories are expressed by the number of occurrences. (color should be used for any figures in print). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Fig. 9. Number of occurrences and affected people for every decade (colour scale) by categories of losses and damages. Human impacts are expressed by the number of affected people. The other categories are expressed by the number of occurrences. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) A.O. Tavares et al. Applied Geography 135 (2021) 102534 16 typologies identified allows the definition of territorially differentiated alerts. This matter will later be subject to further analysis and study. The construction and development of a database, such as the one presented in this article, is an important contribution for the assessment of territorial vulnerability and coastal risk management. From the start, the results contribute to satisfy some of the requirements that emanate from the EU Floods Directive (European Parliament and Council of the European Communities, 2007). Among them, it contributes to the collection of historical data on coastal flooding occurrences and their associated impacts, even as, the spatial and temporal identification of the most affected areas. The results can also provide a wide and improved temporal and spatial knowledge about coastal historical flood events that can contribute to the validation of flood predictive models and the design of the risk framework. The information from historical data could also be used for the assessment of territorial vulnerability and the definition of tools for the management of coastal flooding risk, assisting decision-makers for the establishment of adaptation and resilience measures based on knowledge of past events. These data allow the understanding of the dynamic relationships existing between coastal floods and the different coastal communities, allowing the communication and mediation between the local population and the different authorities. 5. Conclusions The present study demonstrates the importance of historical databases of natural hazardous processes from hemerographic sources. The analysis of the database made it possible to identify, explore and evaluate the different occurrences of flooding and coastal overtopping, allowing the identification of its impacts. Consequently, it also allowed locating the most vulnerable areas, the spatial and temporal characterization of occurrences and impacts, as well as the respective forcers. The present database differs from others carried out in other countries, due to its range of topics and primary fields that allow the detailed and comprehensive characterization not only of the identified occurrences, but also of the associated impacts and damages. The database presented here is still characterized by its interdisciplinary nature, consistency, and user-friendliness. The construction of a national database of disasters plays an important role in risk management, namely through the prevention, reduction and mitigation of its consequences. The results from this database, its crossing with real-time forecasting and monitoring models of coastal flooding forcers, and the consideration of the different dimensions of the territorial vulnerability, will allow the development of an innovative reference framework to support risk management. It is intended that the data and results obtained become known, so they can be discussed and considered by the different stakeholders and official entities in order to create more resilient communities, in line with the national and regional key strategies and towards the Sendai Framework for Disaster Risk Reduction (UNISDR, 2015). The results and analysis of the present database provide useful information on the past, present and future of coastal areas and consequently of their exposure to coastal flooding. The information resulting from the present database, its integration in sea-level rise models, can work as an important tool for the process of adaptation of coastal zone to climate change, namely through processes of relocation, accommodation or artificial defence of these areas. The present study intends to promote and contribute to a risk and emergency planning and management more focused on the differentiating characteristics of the different coastal typologies that exist in continental Portugal. Credit author statement Alexandre Oliveira Tavares: Conceptualization, Formal analysis, Writing - Original Draft, Review & Editing, Supervision, Project administration, Funding acquisition, Supervision. Jos´ e Leandro Barros: Conceptualization, Methodology, Formal analysis; Investigation, Resources, Data acquisition, Writing - Original Draft, Writing - Review & Editing, Visualization. Paula Freire: Conceptualization, Methodology, Investigation, Writing - Review & Editing, Supervision, Project administration, Funding acquisition. Pedro Pinto Santos: Formal analysis, Resources, Writing - Review & Editing. Luis Perdiz: Investigation, Data acquisition. Andr´ e Bustorff Fortunato: Investigation, Writing - Review & Editing. Declaration of competing interest None. Acknowledgments This work was financed by Portuguese funds through FCT - Fundaç˜ ao para a Ciˆ encia e a Tecnologia, Project MOSAIC.pt (PTDC/CTA-AMB/ 28909/2017). Pedro Pinto Santos was financed through FCT I.P., under the programme of ‘Stimulus of Scientific Employment – Individual Support’ within the contract CEECIND/00268/2017. References Andrade, C., & Freitas, M. C. (2002). Coastal zones. Gradiva, Liboa. In F. D. Santos, K. Forbes, & R. Moita (Eds.), Climate change in Portugal. Portugal: Scenarios, Impacts and Adaptation Measures – SIAM project, 173–2019. Antunes, C., & Taborda, R. (2009). sea level at Cascais tide gauge: Data, analysis and results (Proceedings of the 10th International Coastal Symposium) Journal of Coastal Research, SI, 56, 218–222 (Lisbon, Portugal). Barnolas, M., & Llasat, M. C. (2007). A flood geodatabase and its climatological applications: The case of Catalonia for the last century. Natural Hazards and Earth System Sciences, 7, 271–281. Barriendos, M., & Rodrigo, F. S. (2006). Study of historical flood events on Spanish rivers using documentary data. Hydrological Sciences Journal, 51(5), 765–783. Bertin, X., Prouteau, E., & Letetrel, C. (2013). A significant increase in wave height in the North Atlantic Ocean over the 20th century. Global and Planetary Change, 106, 77–83. Callaghan, J., & Power, S. B. (2014). Major coastal flooding in southeastern Australia 1860–2012, associated deaths and weather systems. Australian Meteorological and Oceanographic Journal, 64(3), 183–213. Coelho, C., Silva, R., Veloso-Gomes, F., & Taveira-Pinto, F. (2009). Potential effects of climate change on northwest Portuguese coastal zones. ICES Journal Of Marine Science, 66, 1497–1507. Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., … Vitart, F. (2011). The ERA-interim reanalysis: Configuration and performance of the data assimilation system. Quarterly Journal of the Royal Meteorological Society, 137, 553–597. https://doi.org/10.1002/qj.828 Devoli, G., Strauch, W., Ch´ avez, G., & Høeg, K. (2007). A landslide database for Nicaragua: A tool for landslide-hazard management. Landslides, 4(2), 163–176. DGT. COS – land use land cover for Portugal. Territory’s general directorate. Avaliable at: http://www.dgterritorio.pt/dados_abertos/cos/. EEA. The changing faces of Europe’s coastal areas. EEA Report, European Environment Agency, Copenhagen. Available online: https://www.eea.europa.eu/publications/ee a_report_2006_6. (Accessed 28 January 2020). EM-DAT. (2013). The OFDA/CRED international disaster database. Brussels, Belgium: Universite Catholique de Louvain. www.emdat.be. ENGIZC. (2009). Estrat´ egia Nacional para a Gest˜ ao Integrada da Zona Costeira (ENGIZC). Resoluç˜ ao do Conselho de Ministros, (82), 6056–6088. European Parliament and Council of the European Communities. (2007). Directive on the assessment and management of flood risks (2007/60/EC). Official J L288/27–34, Strasbourg, 23 October 2007. Fanjul, E. A., Gomez, B. P., Carretero, J. C., & Arevalo, I. R. S. (1998). Tide and surge dynamics along the iberian Atlantic coast. Oceanologica Acta, 21/2, 131–143. Ferreira, O., & Matias, A. (2013). Portugal. In A. Williams, & E. Pranzini (Eds.), Coastal erosion and protection in Europe (pp. 278–293). Routledge. https://doi.org/10.4324/ 9780203128558. IPCC. (2014). Climate change 2014: Impacts, adaptation, and vulnerability. In C. B. Field, V. R. Barros, D. J. Dokken, K. J. Mach, M. D. Mastrandrea, T. E. Bilir, M. Chatterjee, K. L. Ebi, Y. O. Estrada, R. C. Genova, et al. (Eds.), Part A: Global and sectoral aspects. Contribution of working group II to the fifth assessment report of the intergovernmental Panel on climate change (p. 1132). Cambridge, UK and New York, NY, USA: Cambridge University Press. Fortunato, A. B., Li, K., Bertin, X., Rodrigues, M., & Miguez, B. M. (2016). Determination of extreme sea levels along the Iberian Atlantic coast. Ocean Engineering, 111(1), 471–482. https://doi.org/10.1016/j.oceaneng.2015.11.031 A.O. Tavares et al. Applied Geography 135 (2021) 102534 17 Fortunato, A. B., Meredith, E. P., Rodrigues, M., Freire, P., & Feldmann, H. (2019). Nearfuture changes in storm surges along the Atlantic Iberian coast. Natural Hazards, 98/ 3, 1003–1020. https://doi.org/10.1007/s11069-018-3375-z Fortunato, A. B., Oliveira, A., Rogeiro, J., Tavares da Costa, R., Gomes, J. L., Li, K., & Rodrigues, M. (2017). Operational forecast framework applied to extreme sea levels at regional and local scales. Journal of Operational Oceanography, 10(1), 1–15. Francis, J. A., & Vavrus, S. J. (2012). Evidence linking Arctic amplification to extreme weather in mid-latitudes. Geophysical Research Letters, 39(6), 6. L06801. Freire, S., Santos, T., & Tened´ orio, J. A. (2009). Recent urbanization and land use/land cover change in Portugal—the influence of coastline and coastal urban centers. Journal of Coastal Research, 1499–1503. Freire, P., Tavares, A. O., S´ a, L., Oliveira, A., Fortunato, A. B., dos Santos, P. P., Rilo, A., Gomes, J. L., Rogeiro, J., Pablo, R., & Pinto, P. J. (2016). A local-scale approach to estuarine flood risk management. Natural Hazards, 84(3), 1705–1739. Guzzetti, F., & Tonelli, G. (2004). Information system on hydrological and geomorphological catastrophes in Italy (SICI): A tool for managing landslide and flood hazards. Natural Hazards and Earth System Sciences, 4, 213–232. Haigh, I. D., Ozsoy, O., Wadey, M. P., Nicholls, R. J., Gallop, S. L., Wahl, T., & Brown, J. M. (2017). An improved database of coastal flooding in the United Kingdom from 1915 to 2016. Scientific data, 4, 170100. IOC/UNESCO, IMO, FAO, & UNDP. A blueprint for ocean and coastal sustainability. An inter-agency report towards the preparation of the UN conference on sustainable development (Rio+20). Available online: https://digital.library.unt. edu/ark:/67531/metadc226586/m2/1/high_res_d/UNDP 20-%20Rio%20+20.pdf. (Accessed 28 January 2020). Kron, W. (2013). Coasts: The high-risk areas of the world. Natural Hazards, 66(3), 1363–1382. La Red. (2009). The challenge of information sources. Project DesInventar http://www. desinventar.net/data_sources.html. Munich, R. (2011). Topics geo-natural catastrophes 2010: Analyses, assessments, positions. Munich: Munich Reinsurance Company. Needham, H. F., Keim, B. D., Sathiaraj, D., & Shafer, M. (2013). A global database of tropical storm surges. Eos, Transactions American Geophysical Union, 94(24), 213–214. Neumann, B., Vafeidis, A. T., Zimmermann, J., & Nicholls, R. J. (2015). Future coastal population growth and exposure to sea-level rise and coastal flooding—a global assessment. PloS One, 10, Article e0118571. Nicholls, R. J., & Cazenave, A. (2010). Sea-level rise and its impact on coastal zones. Science, 328, 1517–1520. Pereira, C., & Coelho, C. (2013). Mapas de risco das zonas costeiras por efeito da aç˜ ao energ´ etica do mar. Revista de Gest˜ ao Costeira Integrada. Journal of Integrated Coastal Zone Management, 13(1), 27–43. Pinto, P., Cabral, P., Caetano, M., & Alves, M. F. (2009). Urban growth on coastal erosion vulnerable stretches. Journal of Coastal Research, 1567–1571. Ponte Lira, C., Nobre Silva, A., Taborda, R., & Freire de Andrade, C. (2016). Coastline evolution of Portuguese low-lying sandy coast in the last 50 years: An integrated approach. Earth System Science Data, 8(1), 265–278. Raska, P., & Emmer, A. (2014). The 1916 catastrophic flood following the bíl´ a desn´ a damfailure: The role of historical data sources in the reconstruction of its geomorphologic and landscape effects. Geomorphology, 226, 135–147. https://doi. org/10.1016/j.geomorph.2014.08.002 Raˇ ska, P., Z´ abranský, V., Dubiˇ sar, J., Kadlec, A., Hrb´ aˇ cov´ a, A., & Strnad, T. (2014). Documentary proxies and interdisciplinary research on historic geomorphologic hazards: A discussion of the current state from a central European perspective. Natural Hazards, 70, 705–732. https://doi.org/10.1007/s11069-013-0839-z Rilo, A., Tavares, A., Freire, P., Santos, P. P., & Zˆ ezere, J. L. (2017). The contribution of historical information to flood risk management in the Tagus estuary. International journal of disaster risk reduction, 25, 22–35. Rocha, C., Antunes, C., & Catita, C. (2020). Coastal vulnerability assessment due to sea level rise: The case study of the Atlantic coast of mainland Portugal. Water, 12(2), 360. Ruocco, A. C., Nicholls, R. J., Haigh, I. D., & Wadey, M. P. (2011). Reconstructing coastal flood occurrence combining sea level and media sources: A case study of the solent, UK since 1935. Natural Hazards, 59(3), 1773. Santos, F. D., Forbes, K., & Moita, R. (2002). Climate change in Portugal. Scenarios, impacts and adaptation measures. Gradiva, Lisboa: Projeto SIAM. Santos, F. D., Lopes, A. M., Moniz, G., Ramos, L., & Taborda, R. (2017). Grupo de Trabalho do Litoral: Gest˜ ao da Zona Costeira: O desafio da mudança. Filipe Duarte Santos, Gil Penha-Lopes e Ant´ onio Mota Lopes, ISBN 978-989-99962-1-2. Lisboa. Santos, F. D., & Miranda, P. (Eds.). (2006). Alteraç˜ oes clim´ aticas em Portugal. Cen´ arios, impactos e medidas de adaptaç˜ ao. Gradiva, Lisboa: Projeto SIAM II. Santos, P. P., Tavares, A. O., & Zˆ ezere, J. L. (2014). Risk analysis for local management from hydro-geomorphologic disaster databases. Environmental Science & Policy, 226, 135–147. https://doi.org/10.1016/j.envsci.2013.12.007 Screen, J. A. (2013). Influence of Arctic sea ice on European summer precipitation. Environmental Research Letters, 8(4). https://doi.org/10.1088/1748-9326/8/4/ 044015 Silva, S. F., Martinho, M., Capit˜ ao, R., Reis, T., Fortes, C. J., & Ferreira, J. C. (2017). An index-based method for coastal-flood risk assessment in low-lying areas (Costa de Caparica, Portugal). Ocean & Coastal Management, 144, 90–104. Tang, Q., Zhang, X., & Francis, J. A. (2013). Extreme summer weather in northern midlatitudes linked to a vanishing cryosphere. Nature Climate Change, 4, 45–50. https:// doi.org/10.1038/nclimate2065 Taylor, F. E., Malamud, B. D., Freeborough, K., & Demeritt, D. (2015). Enriching great britain’s national landslide database by searching newspaper archives. Geomorphology, 249, 52–68. https://doi.org/10.1016/j.geomorph.2015.05.019 Tolman, H. L. (2009). User manual and system documentation of WAVEWATCH III. NOAA/ NWS/NCEP/MMAB Technical Note 276 version 3.14. . UNISDR (United Nations International Strategy for Disaster Reduction). (2015). Sendai framework for disaster risk reduction 2015– 2030. Veloso-Gomes, F. (2007). A gest˜ ao da zona costeira portuguesa. Journal of Integrated Coastal Zone Management, 7(2), 83–95. Veloso-Gomes, F., Taveira-Pinto, F., Neves, L., Barbosa, J. P., & Coelho, C. (2004). Erosion risk levels at the NW Portuguese coast: The douro mouth - cape Mondego stretch. Journal of Coastal Conservation, 10(1/2), 43–52. Weisse, R., Bellafiore, D., Menendez, M., Mendez, F., Nicholls, R. J., Umgiesser, G., & Willems, P. (2014). Changing extreme sea levels along European coasts. Coastal Engineering, 87, 4–14. Zˆ ezere, J. L., Pereira, S., Tavares, A. O., Bateira, C., Trigo, R. M., Quaresma, I., Santos, P. P., Santos, M., & Verde, J. (2014). Disaster: A GIS database on hydrogeomorphologic disasters in Portugal. Natural Hazards. https://doi.org/10.1007/ s11069-013-1018-y A.O. Tavares et al.