Methodology for the damage assessment of vehicles exposed to flooding in urban areas

1 Methodology for the damage assessment of vehicles exposed to flooding in urban areas Eduardo Martínez-Gomariza,b*, Manuel Gómeza,b, Beniamino Russoc,d, Pablo Sánchezc, Josep-Anton Montese aCetaqua Water Technology Centre, Barcelona, Spain bFLUMEN Research Institute, Technical University of Catalonia, Barcelona, Spain cAQUATEC(SUEZ Advanced Solutions), Barcelona, Spain dGroup of Hydraulic and Environmental Engineering, Technical College of La Almunia (EUPLA), University of Zaragoza, Spain eUrban ecology department, Badalona prosper and sustainable area, Badalona City Council, Spain. *Corresponding author. Email: eduardo.mart[email protected]om Abstract:Within urban areas humans carry out a great diversity of activities, and some of them require the use of vehicles. Floods, especially in urban areas, can generate significant tangible direct damages to vehicles themselves and to the urban elements in case of loss of stability and collision, which cannot be dismissed. In this paper, after a state-of-the-art review on damage curves for vehicles, a methodology to assess the direct economic impact for vehicles exposed to flooding has been described, and applied within a study carried out in the framework of the BINGO H2020 EU Project. Only three different studies focused on damages to vehicles in contact with floodwater have been found. Contrasting damage curves for vehicles are found when comparing the three approaches, however theones proposed by the USACE offer a high level of completeness and accuracy. Moreover, USACE’s development is the most current research and all the steps for the development of the damage curves are comprehensively described. Finally, after the description of a detailed methodology for flood damage mapping for vehicles, a procedure to evaluate the Expected Annual Damage for vehicles is offered. Keywords: Urban flood risk; vehicles; damage curves; Expected Annual Damage. 2 1. Introduction 1 During anextreme flood event, mainly in urban areas,vehiclescan suffer hydroplaning or instability due to 2 three types of phenomena: floating, sliding or toppling (Shand et al. 2011). However, the most frequent 3 are the first two and for most cases the instability occurs as a combination of both, floating and sliding. 4 The instability of vehicles has been studied by different authors and a comprehensive analysis of the state 5 of the art on this issue can found in Martínez-Gomariz et al. (2018).Moreover, after losing stability the 6 vehicle becomes buoyant and may be washed away colliding with other urban elements, with potential 7 injuries or fatalities (intangible damage). In this context, vehicles might be considered as massive debris 8 washed away by the flood that could generate significant economic damage and even compromise 9 pedestrian safety. 10 Floods can cause different damages: economic damage (national, community or individual) 11 causing a certain number of properties affected, damage to individuals (deaths, injuries, affections) and/or 12 ecologic and environmental damage (occasionally expressed in monetary terms). Flood damages can be 13 tangible or intangible, defining tangible as those impacts measurable in monetary terms and intangible as 14 those impacts, like people injuries or deaths, that generally are assessed in different terms (Landefeld and 15 Seskin 1982). Within urban areas it can be said that, generally speaking, properties (tangible damages) 16 and pedestrians (intangible damages) are the most affected when an extreme flood occurs. Urban 17 elements like services or infrastructures are affected in a tangible manner by flooding, and particularly 18 vehicles. In turn, flood damages can be classified also as either direct or indirect, whether, respectively, it 19 is due to the physical contact with the water or not (Grigg and Helweg 1975; Hammond et al. 2015). 20 Therefore, floods, especially in urban areas, can generate tangible direct damages to vehicles themselves 21 and to urban elements in case of loss of stability and collision. A clear illustration of these damages is the 22 massive flash flood that occurred in Boscastle (UK) on 16 August 2004 caused by an extreme rainfall 23 event up to 200 mm in 5 h, causing millions of pounds of damages and more than 100 vehicles washed 24 away. Furthermore, the washed out vehicles caused the blockage of a bridge, which collapsed, 25 aggravating greatly the damages. 26 Nonetheless, also indirect tangible damages (like loss of production) due to traffic disruption can be 27 produced. Costs associated with traffic disruption are an often overlooked indirect impact of flooding 28 3 (Hammond et al. 2015).Furthermore, vehicles do not need to become unstable to suffer economic 29 damages, because even if they do not reach the instability threshold, will be partially flooded, resulting in 30 economic costs that cannot be neglected.Some US entities such as the National Auto Auction Association 31 (NAAA, n.d.) conduct studies to detect when a vehicle being offered for sale was actually damaged by a 32 flood. 33 In this paper, a methodology to assess the direct economic impact for vehicles exposed to 34 flooding is described. Firstly, an overall description of traditional flood damage assessment in urban areas 35 is presented, together with a review of the existing flood damage curves developed for vehicles. 36 Subsequently, a description of the case study where the assessment has been conducted is presented, as 37 well as the reasoning for the adequateness of the selected damage curves for vehicles. After the 38 description of a detailed methodology to assess the economic damage of vehicles exposed to flooding, a 39 procedure to evaluate the Expected Annual Damage for vehicles is offered. Finally, some conclusions are 40 presented together with a proposal for future developments regarding the undertaken technique. 41 2. Criteria for direct flood damage assessment in urban areas 42 Flood damage assessment is one of the major concerns after a flood occurs, especially when it comes to 43 flash floods striking urban areas. Usually, most studies undertaking this assessment focus on damages to 44 properties, and the basic tool to carry out this task is the so-called flood damage curves or functions. 45 The pioneer in considering damage to properties was Gilbert F. White (1945) in his doctoral 46 thesis entitled “Human Adjustment to Floods: A Geographical Approach to the Flood Problems in the 47 Unites States”. Among other issues White (1945) defined in greater detail the types of losses when a 48 flood occurred, describing, among others,losses to properties and shops in urban areas. Losses in 49 residential areas could be foundations and structure of dwellings, garages and other buildings, such as 50 cars or other vehicles, damages to the terrain and others like loss of the property rental income. However 51 White (1945) did not define direct or indirect damages per se. There was stated also that water depth and 52 velocity were the variables which established the degree of severity for damages to foundation and 53 structures of dwellings. According to White (1945), water depth was the most limiting condition for such 54 losses. 55 In flood risksstudies, particularly in urban areas, usually direct and tangible damage assessments 56 (i.e. in monetary terms) based on hydrodynamic models are carried out by employing as input design 57 4 storms related to different return periods. Direct and tangible damages to properties vary depending on 58 the type of property, its value, and the restitution costs to initial state (Grigg and Helweg 1975). 59 Apart from either building types or land use, depending on the scale level of the analysis 60 approach (Messner and Meyer 2006), the potential damage degree to a structure or a land use can be 61 measured through damage curves, which are functions of water depth. These curves can measure damage 62 percentages, according to the water depth around, for different flood event intensities, or may be 63 represented based on monetary units over surface unit, format more useful for mapping purposes. 64 Several proposals of flood damage curves can be found worldwide, such as in Australia with the 65 Queensland Government guidelines (EMA, 2002), in USA with the guidelines proposed by the Federal 66 Emergency Management Agency (FEMA, 2015), in Spain (Valencia) with the criterion proposed within 67 the EU Project CRUE (Francés et al., 2008), in Italy with a particular development proposed by the 68 Research Institute for Geo-hydrological Protectionof Turin (Italy) for river Boesio flooding (Luino et al. 69 2009), in the UK with the criterion proposed in the Multi-Coloured Manual (Penning-Rowsell et al. 70 2005)More recently new damage curves, focused on urban pluvial flooding, wereperformed for the 71 district of El Raval (Barcelona) in the framework of the EU Project CORFU (Russo et al. 2013; Velasco 72 et al. 2013). 73 Therefore, in case of a flood damage assessment for an urban area and if available the geo-74 referenced type buildings and the related damage curves, by employing GIS tools, the direct damage 75 assessment for buildings can be conducted.On the other hand, a great diversity of vehicles can be found in 76 urban areas, being cars the most predominant. Accordingly, the economic impact to vehicles when an 77 urban flood occurs cannot be neglected. 78 3. State-of-the-art review: flood damage curves for vehicles 79 Damage curves associated with vehicles can be presented too, following the same criteria as for 80 buildings, in order to find the damages to the vehicle as a function of the water depth. This approach is 81 found in the literature although it is a less mature research than damage curves for buildings and has been 82 carried out by a small number of authors. Only three developments have been found within the 83 bibliography. The approaches, that are summarized in Table 1, are those proposed in the HAZUS-MH 84 model developed by the US Federal Emergency Management Agency (2015), the criteria proposed in the 85 CRUE project (Francés et al., 2008) and, finally, the one proposed by the U.S. Army Corps of Engineers 86 5 (USACE, 2009). 87 3.1. Federal Emergency Management Agency Criterion 88 The U.S. Federal Emergency Management Agency (FEMA) developed in 2003 the project HAZUS-89 MH(FEMA, 2015), which name comes from “Hazard United States-Multiple Hazards”. Through 90 HAZUS-MH, science, engineering and mathematics are combined with GIS technology to estimate loss 91 of lives and properties through mapping representation. HAZUS-MH estimates the physical, social and 92 economic impact that a community might suffer because of an earthquake, flood or hurricane. This tool 93 development aims to help, prepare, mitigate, response and recovery from a hazardous event such as 94 earthquakes, floods or hurricanes. Specifically, the HAZUS-MH module aimed at floods is able to assess 95 damages due to impacts of coastal or riverine flooding. It estimates potential damages caused to 96 buildings, essential facilities, routes and agriculture areas, shelters and fatalities. Direct losses are 97 estimated based on physical damage to the structure, contents and inside buildings. Flood damage to 98 vehicles can also be estimated through HAZUS-MH(Scawthorn et al. 2006; FEMA 2015). The procedure 99 to carry out this assessment is divided in four parts: 100 1. Calculation of vehicles inventory within the studyarea 101 2. Vehicles localisation according to any time of the day in different positions 102 3. Estimation of vehicles monetary value 103 4. Application of a damage curve according to the type of vehicle 104 The proposed method for the vehicles localisation aims to obtain the number of vehicles according to the 105 car park type, vehicle age and type, and time of the day. 106 Moreover, the program associates a parked vehicles ratio per square meter of building structure 107 which are categorized into 33 occupancy groups (11 are residential, 10 are commercial, and 6 are 108 industrial). Several illustrative examples of the vehicles assignment procedure to any building category 109 can be found in the HAZUS-MH manual. In this manner, the number of potentially vehicles at risk are 110 determined, being this figure distributed among the different types of available car parks within the study 111 area. On the other hand, vehicles were classified within three typologies: cars, light truck and heavy truck. 112 Vehicles age distribution and trucks ratio versus cars was carried out by the National Automobile Dealers 113 Association (NADA). 114 6 A price for each type of vehicle is proposed: 22618.47$ (cars), 20969.21$ (light trucks) and 115 76087.67$ (heavy trucks), as a result of averaging different values obtained by consulting car dealerships 116 and websites. A used vehicle is allowed to be valued as 50% of the original price. In order to compute the 117 total value of all vehicles located within the study area, the number of total vehicles will be multiplied by 118 the percentage of cars/light trucks/heavy trucks, percentage of new/used vehicles and the average value of 119 vehicles that match both categories. 120 The implemented flood damage curves for vehicles were developed by considering a gradual damage due 121 to the water level. That is to say, whereas for 30 to 60 cm (1 or 2 feet) of water depth damages to the 122 vehicle are unlikely, a total damage will be caused when the engine compartment will become totally 123 submerged. Moreover, an early warning system is considered to avoid the damage completely because the 124 vehicles’ owners have time enough to move them to safety areas. The development of the flood damage 125 curves was conducted with the support of an expert in this field. Their development was based on the 126 delimitation of three levels of water depths: under the vehicle’s carpeting, between the carpeting and 127 dashboard, and beyond the dashboard. These flood damage curves are shown in Figure 1a. 128 3.2. CRUE EU Project Criterion 129 ERA-Net CRUE is a network of European government departments who directly fund flood risk 130 management programmes and related research actions. The creation and implementation of a European 131 research area in flood risk management – as intended by the CRUE ERA-Net - is an important 132 contribution to an improved trans-national perspective for flood-related research in Europe. 133 Each of the seven successful joint projects within CRUE’s 1st funding initiative for flood risk 134 management research was designed to understand different national approaches to the use and appraisal 135 of non-structural measures, explore what is successful, and what can be improved in terms of efficiency 136 and effectiveness of such measures themselves. 137 One of the sections of this project focuses on the risk assessment to the RambladelPoyo basin 138 (Valencia, Spain), developing firstly hazard maps, secondly, studying the vulnerability within the flooded 139 area, and finally estimating the risk. In the second step (i.e. vulnerability analysis), it is intended to assess 140 damages to movable assets (residential constructions and elements of public areas) and buildings 141 (residential constructions). Moreover, damages to furniture and structures for industrial, commercial and 142 7 storing usesare estimated. The ultimate purpose is to obtain the vulnerability curve (i.e. damage curve) 143 associated with each use. This curve reflects the damage expressed in Euro to each affected use as a 144 function of water level. 145 The RambladelPoyo covers the West area of Valencia. It mainly and specifically affects to the 146 municipalities of Chiva, Cheste, Loriguilla, Quart de Poblet, Torrent, Picanya, Paiporta and Catarroja. 147 The analyzed municipalities to obtain the vulnerability curve are these: Alaquás, Aldaia, Alfafar, 148 Catarroja, Massanassa, Mislata, Paiporta, Picanya, Quart de Poblet, Riba-rojadelTúria, Torrente, Valencia 149 and Xirivella. 150 In the same manner that damage curves were proposed to different typologies of buildings, also 151 damage curves for vehicles were presented. Those curves were developed only for two types of vehicles: 152 diesel and gasoline. In Figure 1b both curves are showed together with the averaged one. The scale of 153 analysis in this case study, in order to assess the flooding risk, considers the number of cars per square 154 meter of flooded street. In this damage assessment for vehicles, also costs related to street cleaning and 155 drainage were taken into account. 156 3.3. U.S. Army Corps of Engineers Criterion 157 The US Army Corps of Engineers (USACE) conducted a specific study for the development of damage 158 curves for vehicles exposed to flooding. This study is reported as a Memorandum: Economic Guidance 159 Memorandum, 09-04, Generic Depth-Damage Relationships for Vehicles (USACE, 2009). The USACE's 160 Water Resources Institute's "Flood Damage Data Collection Program" collects information from past 161 floods in order to produce reliable estimations on economic damages due to flood events. As part of the 162 surveys carried out to determine the effects of flooding on residential properties, data were also collected 163 on the damage caused to vehicles parked in such dwellings for the ten communities that suffered the 164 greatest floods. The baseline information for the development of such curves (Figure 1c) was therefore 165 the data provided by the affected owners in relation to the estimation of the vehicle, the damage suffered 166 and the depth of water that affected the vehicle.Damage curves were developed for five vehicle types 167 (sedans, pickups, SUVs, sports, mini vans)(Table 2) from a sample of 640 vehicles. Such data were 168 processed statistically to construct such curves by regression analysis. 169 Ultimately, the purpose of the memorandum is to provide guidelines for the generic use of damage curves 170 8 developed for flood risk management studies requested by the USACE. These curves will be normally 171 used in studies for urban flooding since in rural areas the density of vehicles is not considerable. 172 There are two methods to apply these curves, the first focuses on vehicles parked in residential locations 173 and the other focuses on non-residential locations. The first requires different data: the height of the 174 vehicle, which is assumed to be the rise of the affected residential property; an average of vehicles per 175 property in the study area; the classification of these in the different types proposed; and finally, the 176 percentage of vehicles that will actually be parked on the property when the flood affects that area. The 177 memorandum offers different sources of information in the United States to obtain the data required to 178 conduct the assessment of damage to vehicles. 179 The application to vehicles that are parked in non-residential locations is analogous but more specific data 180 should be collected. In this case, sources are not provided in the memorandum. Obtaining the number of 181 vehicles parked in shops cannot be carried out using the proposed residential method. The distribution of 182 vehicle numbers and typology should be grouped by individual shops to accurately assess damage. 183 However, the same generic damage curves can be used for both parked vehicles in residential and 184 commercial areas. 185 4. Badalona case study (EU project BINGO) 186 The case study to undertake the flood damage assessment to vehicles is located in the Spanish city of 187 Badalona. This city is one of the six areas across Europe in which the EU project BINGO (Bringing 188 INnovation to onGOing water management - a better future under climate change (2015-2019)) is focused 189 on: Portugal, Spain, Cyprus, Germany, the Netherlands and Norway. BINGO aims at providing practical 190 knowledge and tools to end users, water managers, decision and policy makers affected by climate 191 change to enable them to better deal with all climate projections, including droughts and floods. 192 Badalona, with more than 215,000 inhabitants within its administrative limits on a land area of 193 more than 21.2 Km2, is located in eastern Catalonia (Spain) and is part of the Barcelona metropolitan area 194 (in Spanish AMB). It is located on the left bank of the Besòs River facing on the Mediterranean Sea, and 195 backed by the Marinamountain range (Figure 2).Badalona is the third most-populated municipality in 196 Catalonia after Barcelona and L'Hospitalet de Llobregat. The average population density of the city is 197 9 around 10.000 inhab./Km2, although this value can increase significantly in some districts of the old town 198 and the city centre. 199 The morphology of Badalona presents areas close to the Marina mountain range with high 200 gradients (4% on average) and other flat areas close to the Mediterranean Sea, around the historical 201 centre. Quick hydrological response of these areas can generate runoff with high peak flow and 202 significant velocities. On 14 September 1999 (Figure 3), as a result of precipitations between 50-100 mm 203 in two rainfall events of around one hourof duration each, some ephemeral urban water courses 204 overflowed. More than 1 million Eurowas claimed to insurance companies for direct damages (e.g. 205 underground car parks and metro stations flooded).Focusing only on vehicles damages, 136,000 Euro 206 were claimed which was almost 14% of the total damages. 207 An integrated 1D sewer model (MOUSE Software) was developed in 2012 in the framework of 208 the elaboration of the Badalona Drainage Master Plan (DMP) with a specific planning purpose, so it is not 209 currently operative. In the framework of BINGO project, the sewer model was updated and exported to 210 Infoworks Integrated Catchment Modelling (Infoworks ICM) provided by Innovyze. A 2D model for the 211 overland flow simulation was coupled with the 1D sewer model. Finally, the integrated model 212 wasvalidated using the existent field data, 3 rain gauges covering uniformly the administrative land of the 213 city and 17 flow depth level sensors located in the sewer network (van Alphen et al., 2016). 214 Four design storms have been developed for 1, 10, 100 and 500 years of return period, based on 215 the existing rainfall IDF curves (Casas et al. 2010) developed through 5 minutes resolution time series 216 measured in The Fabra Observatory, located in Barcelona next to Badalona. Thus, these design storms 217 have been employed as inputs for the updated and coupled 1D/2D model and thereby obtaining, as 218 outputs, four delimitations (i.e. one per return period considered) for the flooded area within the Badalona 219 municipality limits. These outputs are the triangular non-structured grid cells considered for the 2D 220 hydrodynamic calculation, and the resulting variable water depth, among others, is stored in each 221 triangular cell. That is the starting data before the application of the methodology for the vehicular 222 damage assessment due to flooding for each return period considered. 223 6. Methodology to assess the economic damageof vehicles exposed to flooding 224 In this section, a methodology to assess damages to vehicles exposed to flooding is proposed and 225 implemented to the case study of Badalona. To do this, GIS tools, damage curves for vehicles and water 226 16 Actual damage data from the Consorcio de Compensación de Seguros (CCS), the reassurance that 390 covers the catastrophic and extreme situations in Spain, was obtained for the previously-described 14 391 September 1999 event and 31July 2002 event. Among this data only tangible directdamages for vehicles 392 were considered, adding up to a total of 136,000 € and 142,000 € respectively. The CCS damage data was 393 compared to the estimated values. 394 By plotting damages to vehicles and return periods considered, a function can be fitted as 395 indicated in Figure 12. In that manner, an estimation of damages for both considered events, which are 396 related to a 3.3 (1999) and 1.7 (2002) years return period, can be carried out. The estimated damages, 397 according to the fitted function (Figure 12), were of 199,240.14 € for the 14 September 1999 event and of 398 145,864.718 € for the 31July 2002 event. 399 In order to conduct the damage assessment for all return periods it was considered a minimum 400 limit of 500 € to be claimed for each average plan area of a vehicle (i.e. 7.51 m2 according to Table 3). 401 This assumption means that a vehicle’s owner will not claim an amount lower than 500 €. It is consistent 402 with the analysed real claims for the 1999 flood event, where only 5% of claims were lower than 500 €. 403 Even considering this assumption, as observed, the damages assessed are higher than the 404 reported ones. Although this might seem inaccurate, especially for the 14 September 1999 event, some 405 reasons can explain such differences: 406  Vehicle distributions at the time of these actual events were different to the distribution used 407 from the single aerial photo. 408  Some of the flooded vehicles may have not reported their damages to the CCS (Velasco et al., 409 2015) because they were low (even higher than 500 €), or they were not aware that they could be 410 compensated. Similar situations are observed elsewhere: in a study undertaken in the US 411 (Cummins et al., 2010), it was found that for several natural hazards, the ratio between insured 412 and total losses was on average smaller than 50%. 413  The spatial distribution of the rainfall is essential when assessing damages for a larger area such 414 as the municipality of Badalona. The design storms for each return period have been considered 415 as uniform for the whole catchment. Therefore, only a part of the municipality might has been 416 affected by the real 1999 flood event. No rainfall data in this regard is available for this event. 417 17  Vehicles are movable elements within urban areas, thus, when owners are aware of the 418 possibility of a hazardous event such as flooding, those could be moved to more secure areas. In 419 the present simulations vehicles are considered to be placed at the same spot and damaged may 420 be higher than in a real situation. 421  Actual damages occur also in underground car parks, which have not been considered within this 422 study but will be in future studies. The great amount of private underground car parks precludes 423 to the municipality to have a database of their locations and characteristics. Moreover, the CCS 424 does not know if vehicles were either damaged on surface streets or in underground car parks. 425 Further research must be conducted in this regard. 426 8. Conclusions and future developments 427 Within cities there is a great diversity of vehicles and its number seems to increase worldwide. Although, 428 particularly in Badalona its number has decreased (mainly due to economic crisis) over the last 429 years.Traditionally, most studies undertaking flood damage assessment in urban areas focus on buildings, 430 and the basic tool to carry out this task are the so-called flood damage curves or functions. However, 431 floods, especially in urban areas, can generate significant tangible direct damages to vehicles themselves 432 and to urban elements in case of loss of stability and collision, which cannot be neglected. 433 In this study, after a state-of-the-art review for damage curves for vehicles, a methodology to 434 assess the direct economic impact for vehicles exposed to flooding has been described and applied within 435 the BINGO H2020 EU Project framework. Only three different studies focused on damages to vehicles in 436 contact with floodwater have been found. Contrasting damage curves for vehicles are found when 437 comparing the three approaches, however the ones proposed by the USACE offer a high level of 438 completeness and accuracy. Moreover, USACE’s development is the most current research and all the 439 steps for the development of the damage curves are comprehensively described. For those reasons, 440 together with the availability of damage curves for five types of vehicles, these curves were considered 441 the most adequate to be adopted in this case study. On the other hand, the fact of the damage is expressed 442 as a percentage makes that these curves can be transferred and applicable to other countries, taking into 443 account local prices of vehicles. Finally, after the description of a detailed methodology for flood damage 444 18 mapping for vehicles, a procedure to evaluate the Expected Annual Damage for vehicles is offered. The 445 obtained value of the EAD for the Municipality of Badalona was320,983.31 €. 446 Regarding future research, there is a clear necessity of development of photogrammetry 447 techniques to facilitate the delineation of all vehicles within the study area. Moreover, in that manner, the 448 weighted damage curve could be applied directly on each vehicle, instead of employing the proposed 449 distributed approach. This would allow a spatial assessment of the damage in the city.A frequent 450 availability of satellite images, together with an appropriate photogrammetry technique, would allow to 451 analyse the vehicular occupation for different time instants along the day. 452 In addition, the development of damage curves specifically for European typical vehicles would 453 be a significant contribution. Moreover, only water depth has been considered as unique variable that 454 affects to damages to vehicles, however some others such as velocity, flooding duration, the polluting or 455 sediments load conveyed by water can affect also and should be studied. Finally, the applicability of 456 damage curves is only when the vehicle remains stable, thus once the vehicle becomes unstable the 457 damages should be assessed in a different manner. Note that when applying damage curves the 458 assumption of vehicle stability is considered regardless the watertight level of the vehicle and the velocity 459 of water, critical aspects when considering vehicles stability. Therefore, a more challenging research may 460 be conducted by relating the vehicle’s stability functions with the damage curve, thereby putting aside the 461 assumption that vehicles remain stable when applying damage curves. 462 ACKNOWLEDMENTS 463 The authors thank the BINGO project, which is funded by the EU H2020 (Grant Agreement No. 641739), 464 whose support is gratefully acknowledged. 465 References 466 Casas M. C., Rodríguez R., Redaño Á. (2010). Analysis of extreme rainfall in Barcelona using a 467 microscale rain gauge network.Meteorological Applications17(1): 117-123. 468 Cummins, J.D., Suher, M., and Zanjani, G., 2010. Federal financial exposure to natural catastrophe. In: D. 469 Lucas, ed. Risk Measuring and managing Federal financial risk. Chicago, IL: University of 470 Chicago Press, 61–92. 471 España (2016) Orden HFP/1895/2016, de 14 de diciembre. Boletín Oficial del Estado, 17 de diciembre de 472 2016, 304, pp. 87816-88485. [online: http://www.boe.es/buscar/doc.php?id=BOE-A-2016-473 11948]. 474 19 Federal Emergency Management Agency (FEMA). Department of Homeland Security. Mitigation 475 Division. (2015). Multi-hazard Loss Estimation Methodology. Flood Model. Hazus-MH MR5 476 Technical Manual. Washingtong, D.C., USA. 449p. 477 Francés F., García-Bartual R., Ortiz E., Salazar S., Miralles J. L., Blöschl G., Komma J., Habereder C., 478 Bronstert A., Blume T. (2008). Efficiency of non-structural flood mitigation measures: “room 479 for the river” and “retaining water in the landscape.”CRUE Research Report No I-6. 242p. 480 Grigg N. S., Helweg O. J. (1975) State-of-the-art of estimating flood damage in urban areas. J Am Water 481 Resour Assoc, 11, 379–390. 482 Hammond M. J., Chen A. S., Djordjević S., et al (2015) Urban flood impact assessment: A state-of-the-483 art review. Urban Water Journal, 12, 14–29. 484 Landefeld J.S., Seskin E.P. (1982) The economic value of life: linking theory to practice. Am J Public 485 Health, 72, 555–566. 486 Luino F., Cirio C.G., Biddoccu M., et al (2009) Application of a model to the evaluation of flood damage. 487 Geoinformatica, 13, 339–353. 488 Martínez-GomarizE. (2016). Inundaciones Urbanas: Criterios de Peligrosidad y Evaluación del Riesgo 489 para Peatones y Vehículos (Urban floods: Hazard Criteria and Risk assessment for pedestrians 490 and vehicles). PhD Thesis, Technical University of Catalonia. Barcelona, Spain. 491 Martínez-Gomariz E, Gómez M, Russo B, Djordjević S. (2017) A new experiments-based methodology 492 to define the stability threshold for any vehicle exposed to flooding. Urban Water Journal, 14(9), 493 930-939. 494 Martínez-Gomariz E, Gómez M, Russo B, Djordjević S. (2018) Stability criteria for flooded vehicles: a 495 state-of-the-art review. Journal of Flood Risk Management, 11, S817-S826. 496 Messner F., Meyer V. (2006) Flood damage, vulnerability and risk perception–challenges for flood 497 damage research. UFZ-Diskussionspapiere, No. 13/2005. Leibniz Information Centre for 498 Economics, Germany. 19p. 499 Meyer V., Haase D., Scheuer S. (2009) Flood Risk Assessment in European River Basins—Concept, 500 Methods, and Challenges Exemplified at the Mulde River. Integr Environ Assess Manag 5:17. 501 Meyer V., Priest S., Kuhlicke C. (2011) Economic evaluation of structural and non-structural flood risk 502 management measures: examples from the Mulde River. Nat Hazards, 62, 301–324. 503 NAAA (n.d.). Flood Damage. Supplemental Information. Retrieved May 2, 2018, from 504 http://www.naaa.com/http://www.naaa.com/pdfs/NAAA_Flood_Damage_Supplement.pdf. 505 Penning-Rowsell E., Johnson C., Tunstall S., et al. (2005) The benefits of flood and coastal risk 506 management: A Handbook of Assessment Techniques. Flood Hazard Research Centre, 507 Middlesex University Press, North London Business Park, Oakleigh Road South, London. 508 20 Russo B., Velasco M., Suñer D. (2013) Flood hazard assessment considering climate change impacts-509 Application to Barcelona case study using a 1D/2D detailed coupled model. In: International 510 Conference on Flood Resilience: Experiences in Asia and Europe. Exeter, United Kingdom, 10p. 511 Scawthorn C., Flores P., Blais N., Seligson H., Tate E., Chang S., Miffling E., Thomas W., Murphy J., 512 Jones C., Lawrence M. (2006). HAZUS-MH Flood Loss Estimation Methodology. II. Damage 513 and Loss Assessment. Natural Hazards Review, 7(2), 72–81. 514 Shand T.D., Cox R.J., Blacka M.J., Smith G.P. (2011) Australian Rainfall and Runoff (AR&R). Revision 515 Project 10: Appropriate Safety Criteria for Vehicles (Report Number: P10/S2/020). Sydney, 516 Australia. 7p. 517 U.S. Army Corps of Engineers (USACE)(2009). Economic Guidance Memorandum, 09-04, Generic 518 Depth-Damage Relationships for Vehicles. Washington, D.C, USA. 9p. 519 Van Alphen H., Alves E., aus der Beek T., et al. (2016) Characterization of the catchments and the water 520 systems. Deliverable D3-1. BINGO H2020 EU Project. 98p. [online: 521 http://www.projectbingo.eu/downloads/BINGO_Deliverable3.1_EC.pdf]. 522 Velasco, M., Cabello, À., Russo, B. 2015. Flood damage assessment in urban areas. Application to the 523 Raval district of Barcelona using synthetic depth damage curves. Urban Water Journal, 13(4), 524 426-440. 525 Velasco M., Russo B., Kersting T., Cabello À. (2013) Flood vulnerability assessment considering climate 526 changes impacts. Application to Barcelona case study using relative depth damage curves. In: 527 Internacional Conference on Flood Resilience: Experiences in Asia and Europe. Exeter, United 528 Kingdom.11p. 529 White G. F. (1945) Human Adjustment to Floods: A Geographical Approach to the Flood Problem in the 530 United States. PhD Thesis, University of Chicago, USA. 531 21 532 Table 1. Summary of the depth damages curves identified in the state of the art review (Martínez-533 Gomariz, 2016) 534 Ref. Model Country Development Damage Types of vehicles Initial Cost Analysis Approach (FEMA, 2015; Scawthorn et al., 2006) HAZUSMH (FEMA) EEUU Synthetic Relative (%) Car Light Truck Heavy Truck New or used applying 50% of new one price Individual Objects (Francés et al. 2008) CRUE Spain Synthetic Absolute (€) Gasoline Diesel Averaged No specified Individual objects every 100 m2 affected (USACE, 2009) USACE EEUU EmpiricalSynthetic Relative (%) Sedan Pickup Truck SUV Sports Car Mini Van Market value Individual Objects Table 2. Percentage of damage related to water depth per each vehicle type (adapted from USACE 535 (2009)) 536 Sedan Pickup Truck SUV Sports Car Minivan Depth (m) Damage (%) Depth (m) Damage (%) Depth (m) Damage (%) Depth (m) Damage (%) Depth (m) Damage (%) 0 0 0 0 0 0 0 0 0 0 0.15 7.6 0.15 5.2 0.15 0 0.15 1.4 0.15 0 0.30 28.0 0.30 20.3 0.30 13.8 0.30 29.2 0.30 17.8 0.61 46.2 0.61 34.4 0.61 30.6 0.61 52.8 0.61 38.3 0.91 62.2 0.91 47.5 0.91 45.8 0.91 72.2 0.91 56.8 1.22 76.0 1.22 59.6 1.22 59.4 1.22 87.4 1.22 73.3 1.52 87.6 1.52 70.7 1.52 71.4 1.52 98.4 1.52 87.8 1.83 97.0 1.83 80.8 1.83 81.8 1.83 100 1.83 100 2.13 100 2.13 89.9 2.13 90.6 2.13 100 2.13 100 2.44 100 2.44 98.0 2.44 97.8 2.44 100 2.44 100 2.74 100 2.74 100 2.74 100 2.74 100 2.74 100 3.05 100 3.05 100 3.05 100 3.05 100 3.05 100 Table 3. Averaged characteristics for the different types of vehicles considered 537 Type of Vehicle Average plan area (m2) Average price (€) Average Age (years) Depreciation (%) Depreciated average price (€) % in Badalona Sedan 7.31 18,253 € 12 87 2,373 83 Pickup Truck 7.54 17,599 € 13 90 1,760 7.5 SUV 8.80 58,337 € 13 90 5,834 1.5 Sports Car 8.80 103,949 € 12 87 13,513 0.5 Minivan 9.35 29,961 € 13 90 2,996 7.5 538 539 22 540 541 Table4. Characteristicsof analysed zonesand occupancy percentage of vehicles within each one 542 Zone 1 2 3 4 5 6 7 8 Area (m2) 114,566.99 114,566.99 114,566.99 114,566.99 114,566.99 114,566.99 114,566.99 114,566.99 Usable area (m2) 30,487.83 35,522.64 27,172.37 23,439.16 37,327.20 27,892.59 49,376.99 37,601.37 % area of the total 0.54% 0.54% 0.54% 0.54% 0.54% 0.54% 0.54% 0.54% Number vehicles 449 418 355 202 515 102 196 295 % of the total 0.46% 0.42% 0.36% 0.21% 0.52% 0.10% 0.20% 0.30% m2vehicle 2994.21 2777.02 2288.77 1289.96 3122.94 639.16 1191 1837.11 m2vehicle/m2 usable area 9.8% 7.8% 8.4% 5.5% 8.4% 2.3% 2.4% 4.9% Table5 Comparativeof economic assessment of distributed method against the complete one 543 500 years Zone Method 1 2 3 4 5 6 7 8 Complete 30,311.87 € 32,811.38 € 8,686.23 € 2,482.04 € 8,872.70 € 261.79 € 41,424.83 € 32,621.32 € Distributed 24,914.40 € 26,963.44 € 13,938.91 € 1,575.10 € 15,231.68 € 1,381.00 € 58,927.77 € 25,622.86 € Table6 Probabilities and damages for floods resulted from the different design storms considered 544 Return Period (years) 1 10 100 500 Probability 1 0.1 0.01 0.002 Damage (€) 93,604.34 € 450,767.72 € 1,000,823.20 € 1,672,750.33 € 545 a) b) c) Figure 1. Depth damage curves per type of vehicle proposed in a) HAZUS-MH, b) CRUE project, and c) 546 USACE 547 23 548 Figure 2: Map including Badalona location and AMB administrative limits (in red) 549 Figure 3: Jornet river flooding in Badalona during the event of 14/09/1999 (Source: Gregori Muñoz-550 Ramos Trayter, biologist and head of service of the environmental and sustainability of the Badalona city 551 council) 552 553 Figure 4: Weighted damage curve according percentage of types of vehicles in Badalona 554 24 555 Figure5. Badalona detail area. Aerial photograph of 25 cm resolution 556 557 Figure 6. Location of the eight analysis areas and patterns of vehicular occupation 558 559 25 Zone1 Zone 2 Zone 3 Zone 4 Figure 7. Detail of first four study zones in Badalona and delimitation of existing vehicles 560 561 Figure 8. Triangular cells, output of the two-dimensional hydrodynamic calculations. Each cell stores 562 hydraulic information, specifically water depth 563 564 Vehicular area associated to each cell. It will be a % of the cell according to the pattern where it is located on. Cell contour