Environmental Modelling and Software 163 (2023) 105670

Contents lists available at ScienceDirect

Environmental Modelling and Software

journal homepage: www.elsevier.com/locate/envsoft

A web-based analytical urban flood damage and loss estimation framework

Yazeed Alabbad a, b, c, *, Enes Yildirim a, Ibrahim Demir a, b, d
a
IIHR Hydroscience and Engineering, University of Iowa, Iowa City, USA
b
Civil and Environmental Engineering, University of Iowa, Iowa City, USA
c
Civil Engineering, King Saud University, Riyadh, Saudi Arabia
d
Electrical and Computer Engineering, University of Iowa, Iowa City, USA

A R T I C L E I N F O A B S T R A C T

Handling Editor: Daniel P Ames Information and communication technology serves a crucial role in communicating flood risk to various
stakeholders and facilitating mitigation decisions. While studies extensively utilize flood inundation maps for
Keywords: communicating flood risks, there is a need to integrate a broad spectrum of physical vulnerability parameters
Flood damage into risk estimates. This research aims to build a publicly accessible web platform to analyze and estimate
Flood vulnerability
riverine flood-related damages using HAZUS and HEC-FIA damage functions. This framework will provide loss
Flood risk
estimation for properties, business interruption, vehicles, bridges, and lives. The analysis is available for two
Decision support
HAZUS scopes, including community and property. The community extent enables the user to explore socioeconomic
flood information for several communities in the State of Iowa. In the property scope, the user can generate
outcomes for the impacts of “what if” flood scenarios using user-provided data. The framework offers a guidance
tool to help decision-makers with flood management, such as investigating mitigation interventions.

1. Introduction vulnerability of people and the built environment resulting from a flood
event (Alabbad and Demir, 2022). Flood control projects and floodplain
Flooding is one of the most destructive and recurrent natural threats management contribute to reducing the risk of flooding (Hyde, 2010).
to communities worldwide, causing severe social, environmental, and Although flood risk cannot be avoided altogether in most cases, it is a
economic consequences. Flood risk is anticipated to grow due to climate challenge due to environmental and land use changes and limited re­
change, increasing population, land use development, and infrastructure sources, including data and computational limitations (Li et al., 2022)
decay (Quintero et al., 2018; Yang et al., 2011). Most regions in U.S. and modeling challenges (Salman and Li, 2018; Yildirim et al., 2022).
have been impacted by a flood event, with average annual flood losses of While most studies focus on mapping flood hazards, it is necessary to
$32.1 billion and projected to increase to $40.6 billion by 2050 (Wing disseminate flood-related information, including exposure and vulner­
et al., 2022). Among the U.S. states, Iowa is considered a flood-prone ability, to increase awareness of the impact of flooding and evaluate
state, and throughout the last two decades, many individuals, struc­ alternative solutions (Mohanty et al., 2019) to support mitigation de­
tures, and infrastructure (Alabbad et al., 2021) have been impacted due cisions (Teague et al., 2021).
to flooding (Holmes et al., 2010). Developing innovative visualization Most decisions on flood mitigation need a multitude of data, infor­
and communication methods (Sermet and Demir, 2020) that can quickly mation, and modeling efforts, which might be time-consuming and
disseminate flood-risk information to various stakeholders, including resource-intensive (Ewing and Demir, 2021). An interactive decision
the general public, enhances flood risk knowledge, resulting in early support system can increase the efficiency of common flood mitigation
planning and mitigation, and a more effective response to a flood event and response decisions, for instance, by decreasing the time the decision
(Albano et al., 2015). maker spends evaluating “what-if” scenarios (Wang, 2007; Yildirim,
Flood vulnerability assessment is critical to understanding how 2022). Flood event repositories and information and communication
floods affect infrastructure and communities and what mitigation stra­ technologies have facilitated addressing challenges in flood risk man­
tegies and measures are appropriate to take place in order to minimize agement (Almoradie et al., 2013; Haltas et al., 2021). In the U.S., HAZUS
the negative impacts. It involves analyzing the exposure and and HEC-FIA are common tools used widely by decision makers to

* Corresponding author. IIHR Hydroscience and Engineering, University of Iowa, Iowa City, USA.
E-mail address: [email protected] (Y. Alabbad).

https://doi.org/10.1016/j.envsoft.2023.105670
Received 19 January 2023; Received in revised form 5 March 2023; Accepted 6 March 2023
Available online 7 March 2023
1364-8152/© 2023 Elsevier Ltd. All rights reserved.
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

estimate flood-related damage and play an essential role in planning literature on the communication of flood vulnerability information at
purposes. The flood risk products can be used to prepare urban and the property and community geographical scales using a web platform
agricultural flood response plans (e.g., shelter needs, inundated area, with the goal of increasing awareness of flood impacts, supporting CRS
projected losses) (Yildirim and Demir, 2022). However, these systems activities for a community, and facilitating individual and community
often require a specialized Geographic Information System (GIS) level access to flood-related information. We have demonstrated the
workstation and a skilled operator to access and run these tools and platform’s capabilities to generate detailed flood impact information on
interpret the results (Hearn et al., 2013; Yildirim and Demir, 2021). demand for several Iowa communities during 100- and 500-year flood
Web-based systems are actively used in hydrology and water re­ events. Also, we looked into how updating the default HAZUS building
sources for informing the public on watershed management (Demir and inventory data would change the way flood damage was estimated.
Beck, 2009), water quality (Sit et al., 2021), and sedimentation (Xu The following sections are devoted to explaining the development
et al., 2019). In the event of a destructive flood event, web-based sys­ and components of the web application in the methodology section, the
tems are aimed at increasing the public’s awareness, warning the public framework capabilities and functions, the pilot use case and findings,
for an upcoming flood event, preparing citizens and crisis managers for and investigating and discussing outcomes in the results section before
efficient protection and rescue measures, and improving the quality of delivering the final remarks of the research in conclusion.
information made available to all flood crisis stakeholders before, dur­
ing, and after the event (Holz et al., 2006; Kochilakis et al., 2016). 2. Methodology
Transferring such a complex system into a friendly web interface elim­
inates many barriers (e.g., technical skills) for different stakeholders, The objective of this research is to develop an easily accessible web
including the public. The analysis helps identify vulnerable areas to framework for decision-makers and the public to be aware of direct and
flooding and enhances the implementation of mitigation measures at indirect flood impacts, including damage to buildings, bridges, and
community (e.g., levees) and property (e.g., elevating a house) levels utility systems; business interruption; loss of life; vehicle damage; and
(Cikmaz et al., 2022). Also, it contributes efficiently to allocating re­ debris amount. Running a similar analysis using HAZUS or HEC-FIA
sources to better respond to and recover from flooding. requires significant effort on data preparation, computation, presenta­
Communicating information about flood risk is essential to inform tion, and interpretation and could take hours to days based on the spatial
the general public and decision makers (e.g., emergency managers) of scale. The proposed framework automates the entire process, prepares
flood impacts and help them plan well to reduce future flood damage the results, and presents them in a matter of seconds using the latest web
(Jonoski and Popescu, 2011; Maidl and Buchecker, 2015). A web-based technologies and standards. The following sections focus on the details
system is regarded as a crucial tool for providing information on flood of back-end and front-end development for the web system (section 2.1),
risk (Mohanty and Karmakar, 2021). It promotes an interdisciplinary flood damage and loss functionalities integrated into the system (section
approach to exploring flood scenarios and potential flood damage and 2.2), and a case study and data preparation (section 2.3).
losses, contributing to taking actions to minimize the impacts. Research
on developing web-based systems for flood risk calculations is rare (Holz 2.1. Framework development
et al., 2006) and limited to monitoring (Muste et al., 2017), early
warning, forecasting, inundation mapping (Hu and Demir, 2021), and The Geographic Information System (GIS) is a common tool that has
flood zoning information. Recently, researchers have developed the ability to manage, visualize, and analyze spatial and non-spatial
web-based systems for flood risk management, including flood inunda­ data. It is utilized by HAZUS and HEC-FIA software to visualize flood
tion map generation (Li and Demir, 2022), communicating flood maps analysis. However, it requires GIS specialists to deal with complex tools
and real-time water level forecasts (Khalid and Ferreira, 2020), flood and translate outcomes into meaningful maps. Therefore, creating a
damage estimation (Alabbad and Demir, 2022), and a property-level web-based flood vulnerability system facilitates accessibility, offers
flood mitigation decision support system (Alabbad et al., 2022). None­ understandable results, and avoids software complexity for the end-
theless, there are flaws in the available web-based flood analysis tools at users.
the property level, which may lead to individuals making incorrect de­
cisions when purchasing a home (Mostafiz et al., 2022). Also, the 2.1.1. Server-side components
acquisition of precise and comprehensive property-level data is a Developing the web-based application starts with preparing the data,
notable challenge for researchers and planners (Hecht et al., 2015), then data processing, and ends with outcome visualization and analysis.
impeding the advancement of research and presenting a hurdle to the Our web interface was developed (Fig. 1) using scripting languages:
creation of satisfactory flood-related policies and projects (Yildirim HyperText Markup Language (HTML), Cascading Style Sheets (CSS), and
et al., 2022). Flood vulnerability and management need additional JavaScript. We utilized Leaflet, an open-source JavaScript library for
research and resources in the scope of disseminating flood knowledge interactive web maps, to build a graphical user interface (GUI). It con­
through web applications to provide a comprehensive picture of flood verts data into map layers and uses mouse interactions to show and
risk and management at various spatial scales. interact with map data.
FEMA has founded the National Flood Insurance Program (NFIP) to First, the map object is initialized with OpenStreetMap and Mapbox
offer flood insurance to property owners, renters, and businesses in tile layers using HTML, CSS, and Leaflet libraries. Besides the tile layers,
communities that achieve its minimum standards to regulate flood- zoom, scale, area measurement, print function, and attribution controls
prone areas. In fact, the NFIP is still limited to minimizing the flood are added to the map. Also, the control geocoder is attached to the map,
impacts and promoting climate change preparedness, which led FEMA enabling the end-user to look up a location. The leaflet library offers the
to create a Community Rating System (CRS) and give insurance dis­ capability to interact with map vectors (points, lines, and polygons)
counts of up to 45% (FEMA, 2021a). To enroll in the CRS, communities created from GeoJSON objects.
should implement activities and programs that exceed the minimum For creating the interactive map of flood vulnerability, spatial and
requirements of the NFIP. Indeed, less than 10% of US communities non-spatial data are required, including flood depth and extent, build­
participated in the CRS, raising concerns regarding whether the ad­ ings, essential facilities, vehicles, demographics, and damage functions.
vantages of CRS enrollment are fairly spread across demographic and They are organized into CSV tables and converted to JavaScript Object
socioeconomic groups (Gourevitch and Pinter, 2022). One of the Notation (JSON) files as a database. JavaScript functions, including
CRS-creditable activities is establishing a local public library or website Leaflet and jQuery, are used to read, integrate, and map JSON data.
that provides flood-related information to a community. JSON files may need to be reduced in size by deleting extraneous at­
This research aims to bridge the scientific knowledge gap in the tributes. In addition, QGIS was used to extract the flood depth at the

2
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

Fig. 1. The overall architecture and components of web-based flood vulnerability application.

centroid of each structure and census block within a community to for estimating vehicle damage for various vehicle types (i.e., passenger
investigate flood damage and loss at community and property levels. cars, light trucks, sedans, and pickups). Fig. 1 shows the flow chart of the
Section 2.2 illustrates the flood analysis functions utilized in the system. flood impact platform developed in this study.

2.1.2. Client-side components

On the client side, community and property level analysis functions 2.2. Flood vulnerability analysis
are developed to enable the user to interact with the map data and
features and explore structure and content damage, business interrup­ Building Damage: Structure and content damage are calculated once
tion losses, the amount of debris, vehicle damage, and potential life a building is within the floodplain. At the property level, we extract
losses resulting from flooding. At the community level, the analysis is flood depth at the centroid of each building to determine the damage
performed for the 100- and 500-year flood scenarios for several com­ percentage. After that, the damage estimation is determined by multi­
munities in Iowa. The user can select a community and access each plying the damage percentage by the structure and content values. We
impacted building to see the analysis. The summary statistics informa­ consider the default HAZUS flood depth damage curves (Fig. 2). The
tion panel, map legend, and layer control are displayed on the map once percentage of structure and content damage associated with flood
the community scenario is selected. Finally, the utility system’s (i.e., depths is extracted from curves developed by the Federal Insurance
wastewater treatment plant) damage can be estimated in the system as Administration and the US Army Corps of Engineers for 33 specific oc­
well. In addition, we used a leaflet plugin that will allow the system to cupancy types (i.e., residential, commercial, industrial, educational, and
upload and read user-provided building, vehicle, and bridge files locally government).
and visualize the outcomes of flood analysis on the map. Business Interruption Losses: Business interruption losses are
In the property scope, the user can explore “what if” scenarios ac­ composed of income, wages, relocation expenses, and rental income.
cording to various flood levels. The system allows individuals to un­ Loss functions embed multiple parameters (i.e., flood depth, occupancy
derstand and visualize the impact of flooding on their property by using type, and area) and utilize the equations for estimating the losses as
flood depth and property information (i.e., area, structure value, num­ listed in Table 1.
ber of people) as inputs. The system covers 33 occupancy types (i.e., Debris Amount: Debris weight estimation due to flooding is limited
residential, commercial, and industrial) and also provides information to the finishes, structure, and foundation of buildings. The debris esti­
mation relies on the occupancy type, area, and foundation. The debris

Fig. 2. An example of structure and content depth-damage curves for a single-family dwelling, 2 stories with basement (RES1-2SWB) and a retail trade (COM1)
(FEMA, 2020).

3
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

Table 1 select the flood depth, vehicle type, and vehicle value, and then the
Business interruption terminology and equations. system can estimate the damage.
Loss Type Equation Abbreviation Bridge and Utility System Damage: HAZUS bridge damage function
is based on the scour index, flood scenario, and bridge value and type.
Income ($) (1-IRFi) * Ai* IDi*LOFi i: represents different
flood depths, The utility system damage estimation involves flood depth, utility
IRF: income recapture replacement value, and the percentage of damage extracted from depth
factor, damage curves.
ID: income per day per
sq.ft ($),
Wage ($) (1-WRFi) * Ai* WDi* LOFi LOF: loss of function, 2.3. Case study
A: area,
WR: wage recapture
The State of Iowa is bounded by the Mississippi River on the east and
factor,
Relocation expenses ($) Ai * [ (1-OOi) * (DCi) + OOi WD: wage per day per the Missouri River on the west, with major rivers (e.g., Cedar, Iowa)
(for damages > 10%) * (DCi+ RDi* LOFi)] sq.ft ($), running into its communities. The state has been experiencing extreme
OO: percent owner flood events over the last 50 years. In 2008, a major flood event inun­
occupied, dated most of the Iowa communities and caused devastating damage.
DC: disruption cost ($),
Cedar Rapids, for example, recorded flood extent beyond the mapped
Rental income ($) (for (1-OOi) * Ai* RDi* LOFi RD: rental per day per
damages > 10%) sq.ft ($) 500-year floodplain for several weeks, resulting in the displacement of
18,000 residents and causing $5.4 billion in losses (USACE, 2018). As a
comprehensive case study, several Iowa communities named Cedar
generation per occupancy type is the area of a building multiplied by the Rapids, Iowa City, Waterloo, Davenport, Cedar Falls, Bettendorf, and
weight of debris associated with a flood depth. Waverly are integrated into the web application to investigate and
Loss of Life: We use the simplified LifeSim methodology, which is visualize the impacts of the 1% (100-yr) and 0.2% (500-yr) flood
applied with HEC-FIA software (HEC-FIA, 2018), to estimate the po­ probabilities at the property and community levels (Fig. 4). Our research
tential number of fatalities during a flooding event. The LifeSim focuses on communities with available tax assessor data for property
approach (McClelland and Bowles, 2002; Lehman and Light, 2016) re­ information. Also, each of the selected communities lies on a riverbank,
quires user input data, including inundation maps, building and popu­ making them at risk of riverine flooding.
lation information, warning and evacuation information, and the arrival
time of the flood. Due to the absence of flood arrival data in our study,
2.4. Data summary
our analysis assumes that people remain at home during flooding and
seek safety inside. Three fatality zones (safe, compromised, and chance)
Flood Maps: 1 m resolution flood maps are obtained from the Iowa
are assigned to the impacted buildings based on the occupancy type,
Flood Center (IFC). IFC produced statewide flood inundation mapping
flood depth, and the age of the population (under 65, over 65) (Table 2).
(1D) and community-level flood scenarios (2D) associated with 50-, 20-,
Each zone is limited by a flood level and has an average fatality rate
10-, 4-, 2-, 1-, 0.5-, and 0.2-percent annual-chance flows using the HEC-
(safe = 0.00002, compromised = 0.12, chance = 0.90). The expected life
RAS models and a coupled hydrodynamic modeling software package
loss per occupancy type is the number of people multiplied by the
(MIKE FLOOD) (Gilles et al., 2012).
average fatality rate.
Building Inventory: The census block is the geographic scale used in
Vehicle Damage: Estimating vehicle damage during flooding is
the HAZUS database to map building inventory compiled from the 2010
dependent on flood depth and vehicle type (car, light truck, heavy
US Census Bureau. Buildings are distributed evenly within each census
truck). The percentage of vehicle damage is extracted from curves
block, leading to overestimating or underestimating flood damage. Our
developed by FEMA (Fig. 3). To estimate the damage, the percent of
research utilizes building inventory collected from county tax offices,
vehicle damage is multiplied by the vehicle value (Equation (1)).
which provides a more complete dataset for building information rep­
Vehicle damage ($) = %D * V * %A (Eq.1) resented spatially as a point.
Demographic: HEC-FIA software (HEC-FIA, 2015) can import the
where %D is the damage percentage for a specific vehicle type, V is the population information per occupancy type from the HAZUS database
vehicle value, and %A is the impacted area percentage per a census using the 2010 US Census Bureau, taking into account population age (i.
block. e., over 65 years) and time (day and night).
Due to the absence of vehicle information per occupancy type within Business Interruption: Variables associated with business interrup­
the study area, our day and night vehicle inventory is based on a census tion (i.e., $ income per square foot per day, $ wage per square foot per
block aggregation level. In the community scenario, we consider the day, $ disruption cost per square foot, $ rental income per square foot
average value of flood depth through a census block to reflect the flood per day, recapture factor, percent owner occupied, recovery time) are
level variation within a census block. During the flooding, some census obtained from the HAZUS technical document for each occupancy type.
blocks were partially impacted, so we weighed the damage in the floo­ The last update to economic values was in 2017.
ded area. We assume all vehicles are on the ground and present in the Debris Weight: Damaged buildings generate three types of debris
flood event. At the property level, besides the HAZUS vehicle categories, (finishes, structure, and foundation). The debris weights per occupancy
we integrate the depth-damage relationships available (USACE, 2009) type associated with flood depths are collected from the HAZUS data­
for different types of vehicles (e.g., sedans). Therefore, the user can base (2020). In our analysis, we assume footing foundations for all

Table 2
An example of default values for flood lethality zone by occupancy type and age group adopted from (USACE, 2018).
Occupancy type Zone height (ft)

Over 65 Under 65

Safe Compromised Chance Safe Compromised Chance

Residential, 1 Story, No basement 2 ft or foundation height 4 6 2 ft or foundation height 4 6

Educational Facility 2 ft or foundation height 13 15 2 ft or foundation height 22 24

4
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

Fig. 3. The vehicle damage functions adapted from (FEMA, 2020; USACE, 2009).

3. Results and discussions

3.1. Flood Impact Data Analytics System (FiDAS)

The web-based flood vulnerability application is composed of two

spatial frames (community and property), designated to help the end-
user explore flood damage to buildings, highway bridges, utility sys­
tems, and vehicles, as well as the amount of debris and loss of life
(Fig. 5). The framework is named FiDAS, which stands for Flood Impact
Data Analytics System, and is accessible publicly at https://hydroin
formatics.uiowa.edu/lab/fidas.
On the main web interface, two spatial levels (community and
property) are available to the user for analysis. Through the community
level, three flood impact analysis options (Fig. 6) are available for
properties, bridges, and utilities. First, a community scenario option
allows the user to see the impacts of 100-year (1% chance) and 500-year
(0.2% chance) flood scenarios on seven communities in Iowa named
Cedar Rapids, Iowa City, Waterloo, Cedar Falls, Davenport, Bettendorf,
Fig. 4. Iowa communities integrated into the web system.
and Waverly. Once a community is selected, an impacted geographical
location of buildings, bridges, and census blocks appears on the inter­
buildings at the community level due to the tax assessor’s lack of
face. Besides, a summary panel for the selected community provides the
foundation type in the dataset. We provide options at the property level
total building, vehicle, and bridge damage under the selected flood
to modify the foundation choices by user.
scenario.
Vehicle Inventory: Per census-block spatial level, the HAZUS data­
Within a specified community border, the user can access damage
base includes day and night vehicle counts from the 2000 Census
analysis for each building. The analysis results include building ID, oc­
building areas and 2006 Dun & Bradstreet data (Dun & Bradstreet,
cupancy type (i.e., residential or commercial), flood depth associated
2006), and vehicle values and ages from the National Automobile
with the flood scenario, structure, content, income, wage, rental income
Dealers Association (NADA, 2011) updated in 2014. Vehicles are
losses, relocation expenses, expected life losses during the day and night
grouped into three types: passenger cars, light trucks (i.e., SUVs), and
timeframes, and finally, the debris amount. Since the available vehicle
heavy trucks (industrial and commercial trucks up to 18-wheelers).
inventory data comes at the census block level, we use an average flood
Bridge and Utility Inventory: Bridges data are imported from the
depth value for each impacted census block to calculate the vehicle
National Bridge Inventory (National Bridge Inventory, 2018) and rep­
damage. The user can access the analysis summary at each census block,
resented as a point location. Each bridge is classified based on its
including the census block ID, vehicle count during day and night,
structural characteristics (i.e., number of spans). Continuous and
average flood depth, and damage during day and night for cars, light
single-span bridges are two types of bridges that are used for flood
trucks, and heavy trucks. Also, highway bridge damage information can
analysis. Besides a bridge’s classification and geographical location,
be explored by locating a bridge point on the map. Then, the bridge ID,
parameters such as the replacement value and scour index for each
type, scour index, damage percentage, damage cost, and functional
bridge are provided by the Federal Highway Administration (last
percentage are shown on the map.
updated: 2019). Utility system data has been compiled from different
An additional option under the community scenario is “upload file,”
sources (The Homeland Infrastructure Foundation-Level Data, 2019).
which allows users to run analysis on their own data. This option will
Data Challenges: We found that some census blocks from the 2010
enable the user to load local files in KML and GeoJSON format, and then
US Census Bureau missed demographic information, limiting the esti­
the flood analysis mentioned above will be revealed spatially on the
mation of the number of people in buildings. Also, vehicle inventory is
framework. In the case of loading a building file, the data for each
missing for census blocks. The County Tax building inventory database
building in the file must include the following information: the build­
is missing some information (for example, foundation type and first-
ing’s geolocation, occupancy type, foundation and area, structure and
floor elevation). We used the default HAZUS values in the absence of
content values, number of people under and over 65 years old, and the
information for some parameters. For example, for locations with
flood depth. To load vehicle inventory files, vehicle types, counts, and
missing values, residential buildings are given a foundation height of 4
costs, as well as the average flood depth for the impacted area and the
ft. The following document provides comprehensive information
impacted area percentage, should be prepared to proceed with the
regarding the HAZUS default building inventory inputs (FEMA, 2021b).

5
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

Fig. 5. The landing page for FiDAS.

communities. Therefore, we created an additional analysis function to

allow the user to explore flood-related losses at the property level
(Fig. 8). First, the user can select the flood depth and provide informa­
tion regarding the property (occupancy type, foundation type, area,
structure and content value, and the number of people). We use 46 oc­
cupancy types in the analysis, considering the basement exists for some
types. Also, two types of foundations (footing and slab on grade) are
available to the user.
Once property-related information is provided, the “property losses”
module allows the system to compute structure, content, and percent
damage, while accessing the “business interruption” function leads to
estimated income, wage and rental income losses, and relocation ex­
penses. The weight of debris resulting from the specified inputs can be
computed after executing the “amount of debris” function. Also, the
“loss of life” function will estimate the expected life losses for people
caught in a building during a flood event. For vehicle damage analysis,
the system provides different types of vehicle classes based on HAZUS
and USACE functions for the user to choose, and then the user provides
the vehicle count and value. As a result, vehicle damage information will
be shown on the result panel, including damage percentage and damage
Fig. 6. An example of flood impact visualization at the community and prop­ dollar amount.
erty levels.
3.2. Community-level flood impact analysis
analysis.
For loading the bridge file, the analysis requires data that has bridge Buildings, vehicles, bridges, and utility systems datasets for several
type (continuous or single span), scour index, flood scenario (i.e., 25-yr, Iowa communities have been integrated into the FiDAS to investigate
50-yr, 100-yr), and replacement value. Also, utility system datasets can flood scenario-based community impacts. Table 3 shows a summary of
be imported into FiDAS to analyze and visualize the damage on the map. the total property damage (structure + content + business interruption)
We provided a data-structure template for each analysis to simplify the per occupancy class for the analyzed communities. During the 100- and
process and guide the user. The user can print the analysis outcomes, 500-year flood events, residential buildings were the most affected class
including map components (Fig. 7), as a PDF file. The upload file option among the studied communities. However, the total damage for other
provides easily accessible flood-related analysis information without the classes with lower building counts (e.g., commercial and government)
need to deal with complex software requirements or time related can exceed the residential losses due to business interruption losses, as
challenges. seen in Waterloo and Davenport. Also, we found that losses generated by
In the case of missing data regarding bridge and utility systems education and agricultural classes had the lowest impacts. When it
within a community, we provide a user input panel under community- comes to total loss, Waterloo, Davenport, Iowa City, and Cedar Rapids
level analysis to discover bridge and utility system damage. For bridge will experience significant damage and losses, especially during the 500-
damage estimation, the user can select the bridge type, scour index, and year flood, while Bettendorf and Waverly would be the least vulnerable
flood scenario and enter the replacement value. Then, the damage communities during the two flood scenarios.
percentage and amount can be determined and visualized on the inter­ Table 4 shows the potential life losses for the studied communities
face. Also, utility damage can be calculated once the user selects the during the 100- and 500-year flood events with the assumption that no
flood depth, utility system type, and replacement value. evacuation happened. It has been found that Cedar Rapids, Waterloo,
The second level of analysis is property-oriented. In many cases, Iowa City, Cedar Falls, and Bettendorf are resilient to life loss during the
building and vehicle information is not available or limited for most 100-year flood, while Davenport and Waverly appear vulnerable. In the
500-year flood, some communities (i.e., Davenport) would be

6
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

Fig. 7. An example of flood impact analysis report.

threatened with up to 10 deaths. In contrast, there would be no life loss

in Iowa City, Cedar Falls, or Bettendorf based on this research meth­
odology. Identifying homes that are vulnerable to life losses helps sup­
port mitigation decisions for flood losses (i.e., give them priority for
evacuation).
Vehicle damage during the 100- and 500-year flood scenarios for the
study area is presented in Fig. 9. In the event of a 100-year flood, vehicle
damage for some communities could reach up to $70 million, while
under a 500-year flood, the damage could reach $160 million, as in
Cedar Rapids and Waterloo. The floodplain areas for the studied

Table 4
Total loss of life during day and nighttime.
Community Loss of life

100-yr 500-yr

Day Night Day Night

Davenport 3 6 9 10
Waverly 1 1 1 2
Waterloo – – 1 2
Cedar Rapids – – 1 –
Iowa City – – – –
Cedar Falls – – – –
Fig. 8. An example of the business interruption losses at the property scope. Bettendorf – – – –

Table 3
Total damage ($Million) and number of impacted buildings per occupancy class.
Community Flood Residential Commercial Industrial Government Education Agricultural

$ # $ # $ # $ # $ # $ #

Waterloo 100-yr 0.6 10 0.2 1 – – 26.6 6 – – – –

500-yr 218.7 5038 261.7 590 10.6 36 933.4 144 37.3 5.0 0.3 2
Davenport 100-yr 5.8 84 95.0 60 7.4 6 118.4 19 – – 0.3 3
500-yr 11.3 161 192.0 102 40.6 17 399.5 28 – – 0.3 3
Iowa City 100-yr 261.7 540 38.9 105 – – – – 0.2 1 – –
500-yr 462.6 566 72.7 175 0.1 2 – – 0.2 1 – –
Cedar Rapids 100-yr 43.9 415 59.4 145 8.6 7 31.7 1 – – 3.5 13
500-yr 175.8 2049 223.2 345 23.7 23 42.8 1 5.1 1 4.8 22
Cedar Falls 100-yr 9.1 114 4.9 21 1.2 5 25.2 10 – – 0.1 1
500-yr 18.2 222 33.2 123 2.1 7 142.4 22 – – 0.2 1
Bettendorf 100-yr 2.9 38 4.8 8 0.8 3 – – – – – –
500-yr 5.5 60 6.9 10 1.7 6 4.0 1 – – – –
Waverly 100-yr 3.1 40 1.2 3 – – – – – – 0.2 2
500-yr 5.7 81 1.2 3 – – – – 1.7 1.0 0.3 2

7
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

Fig. 9. Day and nighttime vehicle damage sorted by floodplain area.

communities vary in size and property distribution. We found that some 3.3. Comparative damage analysis
communities have similar floodplain areas but different vehicle damage,
like Cedar Falls and Davenport. That might be due to the differences in This section explores how the quality of the building inventory data
flood levels or the vehicle counts, values, and distributions. impacts flood damage estimation. The HAZUS default data is based on
the 2010 US Census and is aggregated at the census block level, with the
assumption that building stock is evenly distributed across each census

Fig. 10. Comparison between damage derived from the default HAZUS and tax assessor data.

8
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

block. Also, area-weighted damage estimation is utilized in the HAZUS needs and facilitate preparedness and mitigation efforts. This research
software. As a result, flood damage analysis can be overestimated or emphasizes updating default building inventory data in HAZUS to gain
underestimated, but it can give good insights into flood risk (Ghimire more accurate damage estimation.
and Sharma, 2021). A study was conducted for several Iowa commu­ Data accuracy and non-monetary loss quantification are the main
nities to analyze the flood damage using the default HAZUS building challenges of this research. The tax assessor’s dataset often ignores
inventory data from the national building inventory database (Alabbad public and non-profit buildings due to their exempt status for tax pur­
and Demir, 2022). In this study, we compare property damage resulting poses. Particularly in larger communities, enabling public infrastructure
from using the default HAZUS data and the tax assessor data. The tax can allow us to estimate more accurate results. On the other hand, loss of
assessor data is spatially represented as a point location, which enhances life and vehicle losses are quantified based on certain assumptions, such
building location accuracy over the HAZUS approach. However, tax data as population age and vehicle intensity in a census block. Also, the role
regarding building characteristics (e.g., foundation type, number of of early warning systems is ignored in these methodologies. However,
stories) is not complete in the tax assessor’s dataset. In this analysis, we the research employed the best available datasets and methodologies to
assume residential buildings collected from tax assessor data have one conduct damage and loss assessments. Because the study presented a
story and no basement. Also, HAZUS provides two values (full and scalable web-based model, new datasets can be easily enabled, and more
depreciated) as a replacement value for a building to estimate structure accurate analysis can be provided to the public.
and content damage. We use depreciated replacement values to estimate Further research can be conducted to provide an easily accessible
damage and compare it to county data-based damage. data analytics framework during flooding, including evacuation routes
Fig. 10 represents the comparative analysis of damage and losses and location allocation. A more complete assessment of flood impacts
derived from the tax assessor data and HAZUS. A positive percentage can be derived by investigating scenario-based climate projections and
indicates that damage is overestimated by HAZUS compared to tax other flood variables such as velocity and duration. Also, flood vulner­
assessor data results, while a negative percentage points to under­ ability can be linked with river forecasts (e.g., for the next five days),
estimated damage by HAZUS. It is obvious that HAZUS default data giving a community enough time to prepare for and react to flooding (e.
overestimates the impacted buildings during the two flood scenarios g., moving building content and vehicles out of hazard areas). This study
(1% and 0.2% chance) for the analyzed communities. That can be presents an opportunity for researchers to convert other natural disaster
attributed to the HAZUS aggregation method, which considers that all analyses (e.g., earthquake, tsunami) from package software (i.e.,
buildings within a census block are affected even if they are inundated HAZUS) into interactive maps and make them available and more
partially or not at all. accessible to the public, non-technical people, and communities with
Within the analyzed communities, our results indicated that the limited resources.
HAZUS default data methodology could affect damage and losses
differently. We found that HAZUS overestimated the damage for Cedar Software and data availability
Rapids, Waterloo, and Cedar Falls, while Iowa City appeared under­
estimated. The HAZUS building data information (i.e., counts, occu­ Application name: Flood Impact Data Analytics System (FiDAS).
pancy types), along with the area-weighted approach and building Developer: Yazeed Alabbad, Enes Yildirim, Ibrahim Demir.
replacement valuation, may lead to differences in the structure and First year available: 2022.
content damage estimation. In addition, building characteristics (i.e., Program languages: HTML, CSS, JavaScript.
occupancy type, square footage) play an important factor in estimating Cost: Free.
income, wages, relocation, and rental income losses. In HAZUS default Application availability: https://hydroinformatics.uiowa.edu/lab/
data, square footage by occupancy is assumed by census regions. Our fidas.
results indicate that HAZUS can overestimate and underestimate the
business interruption losses within the studied communities. For
example, during the 500-year flood, Davenport’s wage losses exceeded Declaration of competing interest
HAZUS’s estimation by nearly 250%, while HAZUS overestimated Cedar
Rapids’ wage losses during the two flood scenarios by 80%. Providing The authors declare that they have no known competing financial
accurate geographic locations of buildings along with their detailed interests or personal relationships that could have appeared to influence
characteristics is a major driver for robust building damage estimation. the work reported in this paper.
Still, it is a challenge to obtain complete building inventory data for
most communities in the US. Data availability

4. Conclusion I have shared the data sources and the developed web application
link in the attached manuscript
Flood vulnerability assessment is an important step in informing
decision-makers about the consequences of flooding and performing References
benefit-cost analyses for mitigation measures. We integrated an exten­
sive flood damage assessment methodology based on HAZUS and HEC- Alabbad, Y., Demir, I., 2022. Comprehensive flood vulnerability analysis in urban
communities: Iowa case study. Int. J. Disaster Risk Reduc. 74, 102955.
FIA for the built environment into a web-based system (https://hydroin Alabbad, Y., Mount, J., Campbell, A.M., Demir, I., 2021. Assessment of transportation
formatics.uiowa.edu/lab/fidas) to investigate flood impacts at commu­ system disruption and accessibility to critical amenities during flooding: Iowa case
nity and property levels. Our system covers damage to buildings, vehi­ study. Sci. Total Environ. 793, 148476.
Alabbad, Y., Yildirim, E., Demir, I., 2022. Flood Mitigation Data Analytics and Decision
cles, infrastructure, and expected loss of life. Promoting such flood Support Framework: Iowa Middle Cedar Watershed Case Study. Science of The Total
information through a friendly, accessible web tool and making it un­ Environment, 152768.
derstandable by different stakeholders (i.e., the public, emergency Albano, R., Sole, A., Adamowski, J., 2015. READY: a web-based geographical
information system for enhanced flood resilience through raising awareness in
managers) helps increase awareness of flood impacts, take actions to
citizens. Nat. Hazards Earth Syst. Sci. 15 (7), 1645–1658.
protect communities and properties, obtain credits in the CRS, and avoid Almoradie, A., Jonoski, A., Stoica, F., Solomatine, D., Popescu, I., 2013. Web-based flood
flood losses. The introduced web applications enable non-expert users to information system: case study of Somesul Mare, Romania. Environmental
quickly find information while avoiding complex models that require Engineering & Management Journal (EEMJ) 12 (5).
Cikmaz, B.A., Yildirim, E., Demir, I., 2022. Flood Susceptibility Mapping Using Fuzzy
technical skills and resources regarding data manipulation and analysis. Analytical Hierarchy Process for Cedar Rapids. Iowa. EarthArxiv, p. 3416. https://
The system can also be utilized to support existing decision-making doi.org/10.31223/X57344.

9
Y. Alabbad et al. Environmental Modelling and Software 163 (2023) 105670

Demir, I., Beck, M.B., 2009. GWIS: a prototype information system for Georgia McClelland, D.M., Bowles, D.S., 2002. Estimating Life Loss for Dam Safety Risk
watersheds. In: Georgia Water Resources Conference: Regional Water Management Assessment: A Review and New Approach. Institute for Water Resources.
Opportunities. UGA, Athens, GA, US. Mohanty, M.P., Karmakar, S., 2021. WebFRIS: an efficient web-based decision support
Dun & Bradstreet (D&B), 2006. Business Population Report Aggregated by Standard tool to disseminate end-to-end risk information for flood management. J. Environ.
Industrial Classification (SIC) and Census Block. May 2006. Manag. 288, 112456.
Ewing, G., Demir, I., 2021. An ethical decision-making framework with serious gaming: a Mohanty, M.P., Karmakar, S., Ghosh, S., 2019. WEBFRIS: a web based flood information
smart water case study on flooding. J. Hydroinf. 23 (3), 466–482. system for enhanced resilience through risk awareness. June. In: 11th World
FEMA. Multi-Hazard Loss Estimation Methodology, Flood Model. Retrieved from. Congress on Water Resources and Environment: Managing Water Resources for a
https://www.fema.gov/sites/default/files/2020-09/fema_hazus_flood-model_techni Sustainable Future-EWRA 2019. Proceedings.
cal-manual_2.1.pdf. Mostafiz, R.B., Rohli, R.V., Friedland, C.J., Lee, Y.C., 2022. Actionable information in
FEMA, 2021a. Community rating system. Retrieved from. https://www.fema.gov/floodp flood risk communications and the potential for new web-based tools for long-term
lain-management/community-rating-system. planning for individuals and community. Front. Earth Sci. 242.
FEMA, 2021b. Hazus inventory technical manual. Retrieved from. https://www.fema.go Muste, M., Lyn, D.A., Admiraal, D., Ettema, R., Nikora, V., García, M.H. (Eds.), 2017.
v/sites/default/files/documents/fema_hazus-inventory-technical-manual-4.2.3.pdf. Experimental Hydraulics: Methods, Instrumentation, Data Processing and
Ghimire, E., Sharma, S., 2021. Flood damage assessment in HAZUS using various Management: Volume I: Fundamentals and Methods. CRC Press.
resolution of data and one-dimensional and two-dimensional HEC-RAS depth grids. National Automobile Dealers Association, 2011. NADA data: state of the industry report
Nat. Hazards Rev. 22 (1), 04020054. 2011. http://www.nada.org/NR/rdonlyres/0798BE2A-9291-44BFA1260D372FC89
Gilles, D., Young, N., Schroeder, H., Piotrowski, J., Chang, Y.J., 2012. Inundation B8A/0/NADA_DATA_08222011.pdf.
mapping initiatives of the Iowa Flood Center: statewide coverage and detailed urban National Bridge Inventory. Bridges and Structures. https://www.fhwa.dot.gov/bridg
flooding analysis. Water 4 (1), 85–106. e/nbi.cfm.
Gourevitch, J., Pinter, N., 2022. FEMA’s community rating system: worth the Effort? Int. Quintero, F., Mantilla, R., Anderson, C., Claman, D., Krajewski, W., 2018. Assessment of
J. Disaster Risk Reduc. 21, 396–404. changes in flood frequency due to the effects of climate change: implications for
Haltas, I., Yildirim, E., Oztas, F., Demir, I., 2021. A comprehensive flood event engineering design. Hydrology 5 (1), 19.
specification and inventory: 1930–2020 Turkey case study. Int. J. Disaster Risk Salman, A.M., Li, Y., 2018. Flood risk assessment, future trend modeling, and risk
Reduc. 56, 102086. communication: a review of ongoing research. Nat. Hazards Rev. 19 (3), 04018011.
Hearn, P.P., Longenecker III, H.E., Aguinaldo, J.J., Rahav, A.N., 2013. Delivering Sermet, Y., Demir, I., 2020. Virtual and augmented reality applications for
integrated HAZUS-MH flood loss analyses and flood inundation maps over the Web. environmental science education and training. In: New Perspectives on Virtual and
J. Emergency Manage. 11, 293–302. Augmented Reality. Routledge, pp. 261–275.
HEC-FIA, 2015. Flood impact analysis. User’s manual. Retrieved from. https://www.hec. Sit, M., Langel, R.J., Thompson, D., Cwiertny, D.M., Demir, I., 2021. Web-based data
usace.army.mil/software/hec-fia/documentation/HEC-FIA_30_Users_Manual.pdf. analytics framework for well forecasting and groundwater quality. Sci. Total
HEC-FIA, 2018. Flood impact analysis. Technical reference manual. Retrieved from. Environ. 761, 144121.
https://www.hec.usace.army.mil/software/hec-fia/documentation/CPD-81c,%20 Teague, A., Sermet, Y., Demir, I., Muste, M., 2021. A collaborative serious game for water
HEC-FIA%20Technical%20Reference%20Manual.pdf. resources planning and hazard mitigation. Int. J. Disaster Risk Reduc. 53, 101977.
Hecht, R., Meinel, G., Buchroithner, M., 2015. Automatic identification of building types The Homeland Infrastructure Foundation-Level Data (HIFLD), 2019. Open data.
based on topographic databases–a comparison of different data sources. Retrieved from. https://hifld-geoplatform.opendata.arcgis.com/.
International Journal of Cartography 1 (1), 18–31. USACE, 2009. Economic guidance memorandum, 09-04, generic depth-damage
Holmes, R.R., Koenig, T.A., Karstensen, K.A., 2010. Flooding in the United States relationships for vehicles. Retrieved on November 15, 2022 from. https://planning.
Midwest, 2008. US Geological Survey. erdc.dren.mil/toolbox/library/EGMs/egm09-04.pdf.
Holz, K.P., Hildebrandt, G., Weber, L., 2006. Concept for a web-based information USACE, 2018. Cedar Rapids flood risk management: a case study in disaster recovery.
system for flood management. Nat. Hazards 38 (1), 121–140. Retrieved from. https://usace.contentdm.oclc.org/digital/api/collection/p1
Hu, A., Demir, I., 2021. Real-time flood mapping on client-side web systems using hand 6021coll2/id/2961/download.
model. Hydrology 8 (2), 65. Wang, J., 2007. Development of a Decision Support System for Flood Forecasting and
Hyde, J., 2010. Flood control and the growth machine in Des Moines. Cities in the 21st Warning–A Case Study on the Maribyrnong River (Doctoral Dissertation. Victoria
Century 2 (1), 5. University.
Jonoski, A., Popescu, I., 2011. Distance learning in support of water resources Wing, O.E., Lehman, W., Bates, P.D., Sampson, C.C., Quinn, N., Smith, A.M., Kousky, C.,
management: an online course on decision support systems in river basin 2022. Inequitable patterns of US flood risk in the Anthropocene. Nat. Clim. Change
management. Water Resour. Manag. 26, 1287–1305. 12 (2), 156–162.
Khalid, A., Ferreira, C.M., 2020. Advancing real-time flood prediction in large estuaries: Xu, H., Demir, I., Koylu, C., Muste, M., 2019. A web-based geovisual analytics platform
iFLOOD a fully coupled surge-wave automated web-based guidance system. Environ. for identifying potential contributors to culvert sedimentation. Sci. Total Environ.
Model. Software 131, 104748. 692, 806–817.
Kochilakis, G., Poursanidis, D., Chrysoulakis, N., Varella, V., Kotroni, V., Eftychidis, G., Yang, M., Qian, X., Zhang, Y., Sheng, J., Shen, D., Ge, Y., 2011. Spatial multicriteria
et al., 2016. A web based DSS for the management of floods and wildfires (FLIRE) in decision analysis of flood risks in aging-dam management in China: a framework and
urban and periurban areas. Environ. Model. Software 86, 111–115. case study. Int. J. Environ. Res. Publ. Health 8 (5), 1368–1387.
Lehman, W., Light, M., 2016. Using hec-fia to identify indirect economic losses. In: E3S Yildirim, E., 2022. Assessment of Flood Risk for Urban and Rural Communities in Iowa: A
Web of Conferences, vol. 7. EDP Sciences, 05008. Vulnerability and Mitigation Analysis Framework. Doctoral dissertation, The
Li, Z., Demir, I., 2022. A comprehensive web-based system for flood inundation map University of Iowa.
generation and comparative analysis based on height above nearest drainage. Sci. Yildirim, E., Demir, I., 2021. An integrated flood risk assessment and mitigation
Total Environ. 828, 154420. framework: a case study for middle Cedar River Basin, Iowa, US. Int. J. Disaster Risk
Li, Z., Mount, J., Demir, I., 2022. Accounting for uncertainty in real-time flood Reduc. 56, 102113.
inundation mapping using HAND model: Iowa case study. Nat. Hazards 112 (1), Yildirim, E., Demir, I., 2022. Agricultural flood vulnerability assessment and risk
977–1004. quantification in Iowa. Sci. Total Environ. 826, 154165.
Maidl, E., Buchecker, M., 2015. Raising risk preparedness by flood risk communication. Yildirim, E., Just, C., Demir, I., 2022. Flood risk assessment and quantification at the
Nat. Hazards Earth Syst. Sci. 15 (7), 1577–1595. community and property level in the State of Iowa. Int. J. Disaster Risk Reduc. 77,
103106.

10