Annotated Bibliography

International Journal of Disaster Risk Reduction 56 (2021) 102113

Available online 17 February 2021 2212-4209/© 2021 Elsevier Ltd. All rights reserved.

An Integrated Flood Risk Assessment and Mitigation Framework: A Case Study for Middle Cedar River Basin, Iowa, US

Enes Yildirim *, Ibrahim Demir Civil and Environmental Engineering, University of Iowa, Iowa City, IA, 52242, USA

A R T I C L E I N F O

Keywords: Flood mitigation Benefit-cost analysis Property buyout Data analytics Web systems

A B S T R A C T

Property buyout is one of the most frequently preferred flood mitigation applications by decision-makers for long-term risk reduction. Due to its high-level funding requirements as a mitigation solution, it requires extensive benefits and costs analysis for the selected region. Many communities in the State of Iowa experienced extreme flood events (i.e., 1993, 2008, 2014, 2019), which resulted in a heavy economic impact over the last couple of decades. Nearly 3000 property acquisitions have been made between 2007 and 2017 using federal programs. This study presents a web-based Flood Risk Assessment and Mitigation Environment (FRAME), which provides visual data analytics capabilities to analyze property and community level benefit-cost analysis for property acquisitions. The FRAME allows users to explore and visualize historical mitigation projects and buyouts, and evaluate avoided damages for their communities. As a case study, a detailed benefit-cost analysis of historical property buyouts and direct losses of existing properties in the Middle Cedar watershed in Iowa is studied using stream gauge data from the United States Geological Survey (USGS). Projected stream gauge datasets, which are outputs of two climate scenarios (A1FI-fossil intensive and A2-low emission), are also utilized to assess future avoided losses for acquisitions and possible direct economic losses for existing properties. Case study results indicate that the average benefit-cost ratio (BCR) for buyouts in the studied region is around 0.86. Nearly half of the buyouts reached 4.72 BCR in low emission and 6.3 BCR in fossil intensive climate projections if future floods are considered.

1. Introduction

In the United States, flooding had devastating impacts on commu- nities in terms of social and economic aspects over the last couple of decades. Insured flood losses alone reached nearly $11 billion between 1999 and 2009 in the US [1]. Every year, state and local authorities apply to federal disaster aid programs for recovery and relief efforts after presidential disaster declarations. Flooding has the greatest proportion of the presidential disaster declarations [2], accounting for over 45%. Nearly $80 billion have been allocated over the past 15 years for disaster-related programs, most of which are mainly caused by flooding in the United States [3]. Based on the climate projections, several studies reveal that mean annual streamflow values are projected to increase for many watersheds in the United States [4–6]. Therefore, flood damage mitigation activities will remain critical for many watersheds to avoid future losses [7].

Flood damage mitigation applications can be classified into two main groups: namely, structural and non-structural measures [8]. Structural

efforts focus on reducing the impact of flooding on communities by building levees, floodwalls, and improving drainage systems [9]. On the other hand, non-structural measures like land-use control, acquisition, relocation, and early flood warning systems are preventive actions [8, 10,11]. While prioritizing levee repairs can be a cost-effective solution to mitigate flood loss [12] as a structural measure for flood mitigation, large-scale buyouts could be a feasible solution as a non-structural flood mitigation strategy for private properties [58].

Benefit transfer is another common approach for an area to designate best practices for flood mitigation. Benefit transfer is defined as scaling existing benefit estimations from an old study to a new study site [13, 14]. Even though some studies support benefit transfer as a way of assessing flood mitigation practices, they also highlight issues such as overestimation or underestimation of benefits, which may lead to the failure of mitigation efforts [59,64]. Benefit transfer may not be suc- cessful in the context of assessing flood mitigation on properties due to the unique features of sites, including land use [15], geography [16], and building type [17]. Therefore, site-specific and detailed

* Corresponding author. E-mail address: [email protected] (E. Yildirim).

Contents lists available at ScienceDirect

International Journal of Disaster Risk Reduction

journal homepage: http://www.elsevier.com/locate/ijdrr

https://doi.org/10.1016/j.ijdrr.2021.102113 Received 19 August 2020; Received in revised form 21 January 2021; Accepted 3 February 2021

International Journal of Disaster Risk Reduction 56 (2021) 102113

2

investigations are essential for assessing flood mitigation on properties for selecting best practices.

Urbanization on floodplain results in a rise in property values, therefore, increases potential flood damage [18]. Because flood damage is inevitable in floodplains, settling and investing in the floodplain contributes to a higher risk of flood damage [56]. Thus, a long-term flood mitigation strategy may be possible by relocating structures in the higher-risk flood zones. Although people who live near major wa- terways can be persuaded to participate in property buyout programs, the buyout cost is another challenge for many communities [19]. Property buyout is one of the most common practices to mitigate flooding impact in communities. Social and biophysical indicators [20] can be used to support adaptive strategies [21] in the buyout process and benefit-cost analysis. One of the advantages of property buyout is creating permanent flood mitigation by removing a structure located in flood-prone areas. Existing property buyouts should be investigated to promote or demote the housing recovery policy. The housing recovery has not been examined in detail and is a relatively new subject for policy domains [65]. In the United States, voluntary property buyout is sup- ported by several federal programs such as the Hazard Mitigation Grant Program and Flood Mitigation Assistance Program. The main goal of these programs is to mitigate future hazards in communities by imple- menting long-term disaster mitigation measures [22].

In the buyout process, the property is sold by the owner to the government through the grant application process, which involves local, state, and federal participation [23]. Depending on the grant program, the property must meet specific criteria to become eligible for the grant. For instance, the location of the property must be in a 100-year flood zone or a benefit-cost ratio (BCR) for the property to be cost-effective to proceed with the buyout [24]. Therefore, benefit-cost analysis (BCA) must be as accurate and comprehensive as possible. An extensive flooding impact assessment should aim to cover the direct and indirect economic consequences of the flooding [25]. However, quantifying in- direct flood loss is a great challenge due to uncertainties of the phe- nomena such as its long-term effects [26], data confidentiality [55], and its impact on the outer flood-prone area [27]. Alternatively, compre- hensive direct damage estimations can be considered for mapping the communities’ vulnerability [28]. At this point, detailed direct flood vulnerability analysis becomes crucial input for the decision-making process to understand flooding impact for communities.

1.1. Benefit-Cost Analysis of Buyouts

The majority of the studies are using annualized flood loss to esti- mate BCR [29–31]. Although the statistical approach is one way to es- timate possible losses, flooding may occur more or less frequently due to climate change. Therefore, annualized flood loss estimations may mislead the results for BCA in study sites. To close this gap, historical stream gauges widely deployed in many regions in the United States can be utilized to assess avoided or existing losses for the properties. His- torical gauge records can be processed to reveal peak flows over time in the study site. Then, avoided or existing flood losses are estimated using flood maps that are corresponding to the peak flows. This approach is also applicable to the output of short-term flood forecasting [32,33] and long-term climate projections studies that allow estimations for poten- tial future damages [34]. Damage functions [35], fragility curves [36, 37], and relevant analytical methods are critical to quantify direct los- ses. Based on climate projection studies, different precipitation scenarios and streamflow estimates can be generated for mitigation analysis. Thus, BCA for a property can be investigated by evaluating multiple climate projections such as extreme, moderate, and optimistic precipitation scenarios.

Delivery of the BCA may be improved by using web-based frame- works that can allow decision-makers to access and evaluate their area of responsibilities. Web frameworks allow assessing property losses [35] and evaluating areas with different geographic scales and reduce effort,

time, and resource requirements for the decision-makers. Decision support systems [38] can be enabled via web applications [39] with data analytics capabilities to evaluate what-if scenarios and analysis in one environment [40]. Moreover, public participation [41] and under- standing of BCA and mitigation decisions can be improved with easy-to-use web interfaces to encourage voluntary property buyouts.

1.2. Proposed Framework

Web-based systems are becoming increasingly popular for research and operational applications in water resources and hydrology [42]. Management and analysis of large-scale datasets on the web require optimized data structures [43,44], crowdsourced data collection efforts [61], and distributed computing frameworks [45]. The latest web standards (i.e., WebXR, Speech Recognition API) can augment decision support systems with standardized vocabularies [46] and virtual reality [47] to communicate flood information and model results. On the other hand, existing decision support systems for mitigation applications require extensive resources and skills for database management, geographic information systems, mitigation methodologies, and hy- drology. Because timely and accessible communication of the benefit-cost analysis is critical for decision-makers and the public, minimizing knowledge requirements is critical for a better understand- ing of a mitigation application and its results.

In this study, a generalized web-based Flood Risk Assessment and Mitigation Environment (FRAME) is developed to provide visual data analytics capabilities and simplify property and community level benefit-cost analysis. The FRAME is populated with historical mitigation projects and property acquisitions in Iowa completed by Iowa Homeland Security Emergency Management Department (IHSEMD) after the 2008 and 2014 flood events as a large-scale case study. One of the main ob- jectives of this study is to provide a real-time data analytics and decision support framework to help mitigation decisions. The case study is designed to evaluate the effectiveness of federal mitigation programs in Iowa for property buyouts. It also delivers insights for the region on historical buyouts and benefit-cost analysis for mitigation efforts sup- ported by federal programs.

The framework streamlines the risk and mitigation analysis and improves the accessibility of analysis results to support daily operations of emergency agencies and organizations. Similar analysis often requires knowledge and expertise on geographic information systems, database systems, data analytics, time series analysis and takes significant time to process large-scale data. Besides, analysis of large-scale mitigation datasets (e.g., high-resolution inundation maps, parcel information, climate projections) can generate insights for the historical mitigation applications. The outcomes of the study include a data analytics framework for historical mitigation projects and property acquisition in multiple geospatial scales, providing historical benefit and damage estimation for property acquisitions by using historical gauge data and flood event identification, and analyzing future BCA based on climate forecasts (2 scenarios) for next 30 years.

The remaining sections of the paper are followed by methodology and procedures for the BCA, damage analysis for historical and future scenarios with a case study in Iowa, and development of a web-based framework. Following, results are shared along with the detailed case study for the Middle Cedar watershed. The data analytics framework, mitigation efforts in the state, and statewide analysis are discussed. In the end, the conclusion of the study and prospective studies are provided.

2. Methodology

2.1. Flood Risk Assessment and Mitigation Framework

The web-based framework is designed to enable users to query and evaluate historical mitigation projects and property acquisitions, and

E. Yildirim and I. Demir

International Journal of Disaster Risk Reduction 56 (2021) 102113

3

populated with datasets for the State of Iowa as a case study. The generalized structure of the framework allows easy adaption of the system in other regions. There are 3 main layers within the framework: data management layer, data analytics layer, and mapping and visual- ization layer. Each layer is described in the following subsections. The structure for the framework is illustrated in Fig. 1. Briefly, datasets are stored, processed, and prepared in the data management layer and in- tegrated into the data analytics layer. The data analytics layer processes the datasets and provides the outputs of the analysis to the mapping and visualization layer. At last, the visualization and mapping layer delivers the information to the client-side user interface.

2.2. Data Management Layer

The framework utilizes various datasets such as historical mitigation projects and acquisitions, United States Geological Survey (USGS) stream gauge data, streamflow estimations based on climate scenarios, county tax assessor property dataset, flood inundation maps, damage curves (damage functions), and satellite imagery data. The datasets are stored in PostgreSQL server for querying and processing through Quantum Geographic Information System (QGIS). We collaborated with government agencies and academic institutes (e.g., IHSEMD, IFC, county tax assessor) to collect datasets (i.e., inundation maps, buyout data, parcel information). Historical mitigation projects contain infor- mation about federal grants that are distributed at the county and city level. Property acquisition data stores acquisition cost, date, and eligible grant type in county, city, and property level. The county tax assessor property dataset is integrated into the system for supporting future property buyout decisions. USGS stream gauge data is collected in each community to classify historical flood events. The classification is also

applied to streamflow projection datasets to list projected flood events. The details of datasets are provided below:

Damage Curves: Damage curves which are developed by the United States Army Corps of Engineers (USACE) are employed in the frame- work. In Figs. 2 and 6 out of the 36 widely used damage curves are shared for residential and commercial properties. Briefly, damage curves provide the relation between flood depth and damage percentage for certain occupancy types [48]. Each specific occupancy type has separate structural and content damage curves.

Climate Scenarios and Gauge Records: Historical USGS stream gauge datasets from 2009 to 2019 and streamflow projections based on climate models are analyzed to classify flooding events. The gauge records are collected for 7 different locations for the selected communities (Fig. 3). Streamflow projections are generated for these gauge sites based on 19 climate projections [49]. In this study, the projected streamflow values between 2020 and 2050 from the CCSM3 (Community Climate System Model) model and A1FI and A2 climate scenarios are used. While the A1FI scenario is generated based on fossil intensive activities and high greenhouse gases conditions [50], the A2 scenario is based on low emissions [51] and reduced growth in the economy for different regions [52]. The details of the climate model and streamflow projections are available in Ref. [49]. Although multiple peak flows are possible to observe in a short period of time from a hydrological point of view, the highest peak flow is considered to estimate flood loss due to the long flood recovery process which allows us to prevent overestimation of the loss. Therefore, the algorithm filters the peak flows in the time series that fall in the 6 months range and classifies the highest observed peak as the main flood event. Following, rating curves which are created by National Weather Service (NWS) are used to select the flood map.

Parcel Information: Property information is collected from 2 different sources including county tax assessor parcel data and IHSEMD acquired property data. Geolocation of the existing parcels and property buyouts is crossed-checked by using Google Maps satellite imagery. Following, foundation height and damage curve ids are connected using occupancy types to estimate flood damage. The methodology from the HAZUS (Hazard United States) that assign foundation height based on occu- pancy is employed in the system. Flood depths for individual buildings are estimated using flood inundation maps by extracting foundation height for each building. Then, damage percentage, structural loss, and content losses are calculated for each building.

Fig. 1. Cyberinfrastructure system for the web-based framework.

Fig. 2. Flood inundation depth–damage (structural) functions [48].

E. Yildirim and I. Demir

International Journal of Disaster Risk Reduction 56 (2021) 102113

4

Flood Maps: The flood maps for the Middle Cedar watershed are ac- quired from Iowa Flood Center. The flood maps correspond to several flood scenarios including 50-, 20-, 10-, 4-, 2-, 1-, 0.5-, and 0.2-percent- annual-chance flows (2-, 5-, 10-, 25-, 50-, 100-, 200-, and 500-year re- turn period flows, respectively) by using high-resolution LiDAR digital elevation model data [53]. USGS gauge locations are used as reference points to generate flood inundation maps. Therefore, inundation maps that correspond to the gauge records are selected for the analysis.

Mitigation Project Records: Historical mitigation project datasets are acquired from IHSEMD and stored in a relational database (i.e. Post- greSQL) in the framework within Projects, Project Applicants, and Allo- cated Grants data tables. A relational data schema is created based on unique grant ids, applicant ids, and project ids. Following, advanced queries to filter and extract information from the project records are created in the database and integrated into the data analytics framework.

2.3. Data Analytics Layer

The data analytics layer primarily utilizes historical mitigation project records, property acquisitions, and estimated historical and avoided losses for the case study. The layer is developed using HTML (HyperText Markup Language), JavaScript (JS), and various JS libraries. Datasets are handled in the PostgreSQL server which has the capability for storing, managing, and querying geospatial data. A data web service API (Application Programming Interface) is created using a server-side scripting language (i.e. PHP: Hypertext Preprocessor) to access and share data from database. The data service is integrated to the client-side user interface by creating multiple scopes into query datasets such as location (state-wide, county, city, property level), time parameter, and program type. Damage assessment and benefit-cost analysis for existing and acquired properties are handled in the data analytics layer.

Damage Assessment: The methodology for the damage assessment is employed from the software called HAZUS-MH (Hazard United States) [60] HAZUS tool is one of the most commonly used loss assessment software by decision-makers. However, it requires users to acquire and install various software such as ESRI ArcGIS, Spatial Analyst, and

Microsoft SQL Server. The damage assessment methodology from HAZUS is utilized in the real-time framework by taking advantage of web technologies to reduce the limitations and provide more capabil- ities to create an enhanced and integrated data analytics framework. Therefore, software limitations such as GIS (Geographic Information System), requirements for software licenses, GIS and database expertise, and desktop software maintenance needs are eliminated. Besides, the accessibility of the analysis is enhanced and made available for not only decision-makers but also for the general public. The following loss equation is used in the framework:

Loss = Property Value × Damage Ratio

100 (1)

The property value stands for structural or content value. The dam- age ratio is acquired using damage functions (see Fig. 2) based on building type for both structural and content loss estimation.

Benefit-Cost Analysis: BCA for buyouts requires two main inputs namely the cost of acquisition and total avoided losses. For each prop- erty buyout, the final acquisition cost and date of acquisition are recorded. Total avoided losses are estimated for both structural and content damage considering the results of multiple flood events. It is important to emphasize that the total avoided losses only cover direct structural and content losses that occurred between the final acquisition date and 2019. The ratio of total loss and acquisition cost gives the benefit-cost value for the property. If the ratio is equal or higher than 1, the acquisition can be accepted as successful. To summarize city level BCA, property level acquisition costs and total losses are aggregated and presented at the city level. Climate scenario-based losses are estimated separately to reflect future avoided losses and BCA. Historical and future losses are estimated for both acquired and existing properties.

Limitations: In this study, avoided and future losses are investigated only from the direct flood loss perspective due to data limitations. Although estimating direct loss is significant, other items can be used as cost to contribute to the BCA such as avoided loss of life and injuries, emergency expenses, avoided displacement costs, avoided rental income loss, recreational facilities as benefits and park construction and main- tenance, and loss of property tax [54].

Fig. 3. USGS gauges locations in the Middle Cedar watershed.

E. Yildirim and I. Demir

International Journal of Disaster Risk Reduction 56 (2021) 102113

5

2.4. Mapping and Visualization Layer

The mapping and visualization layer receive query outputs from the data analytics layer and conveys results to the user interface and visu- alizes results using Google Maps JavaScript API. At the state-wide and county level, the results are visualized using a common color schema (from red to green) to illustrate variations. The interface allows results to be visualized by program type, data source, and time. At the city and property level, summary tables are generated to deliver information about the applicant of the project, cost of the project, avoided losses, and historical losses. The information panel provides a summary of the analysis for the selected city or property.

3. Results and Discussions

This section includes a summary of results generated by the data analytics framework for the State of Iowa and a detailed analysis of the Middle Cedar watershed case study.

3.1. Data Analytics Framework

The data analytics system is developed to visualize historical miti- gation projects and property acquisitions for the State of Iowa between 2007 and 2017. The framework provides a data analytics panel to query by data source, program type, location, and time parameter. Table 1 shows the scope of queries in data analytics framework, which allows filtering with the data sources (e.g., historical mitigation projects, project applicants, and allocated grants), program types, location scopes, and date parameters.

Figs. 4 and 5 demonstrate that property acquisition is a common application in Iowa to mitigate flooding impact. Nearly 3000 property acquisitions have been made between 2007 and 2017. The majority of the property acquisitions were completed after the 2008 flooding event. The second largest property acquisitions were made after the 2014 flooding. Public assistance has the largest proportion of grants compared

Table 1 List of filter and query parameters in data analytics framework.

Data Source Program Type Location Scope Date Parameter

Projects Hazard Mitigation State-wide Project Start Date Property

Acquisitions Individual Assistance

County Jurisdiction

Project Closeout Date

Project Applicants Public Assistance City Boundary Final Acquisition Date

Allocated Grants Pre-disaster Mitigation

Property Level Historical Loss

Fig. 4. Distribution of property buyouts in the state of Iowa.

Fig. 5. Distribution of post and pre-hazard mitigation projects.

E. Yildirim and I. Demir

International Journal of Disaster Risk Reduction 56 (2021) 102113

6

to Hazard Mitigation Assistance and Individual Assistance Grant pro- grams. Most of the pre- and post-hazard mitigation projects are completed in eastern Iowa (Fig. 6). The main reason is that eastern Iowa is highly urbanized and hosts large cities in Iowa such as Cedar Rapids, Waterloo, Iowa City, Cedar Falls, and Vinton. Overall, Linn County received the highest number of projects which is 1285 out of the 21,712 projects between 2007 and 2017 in the State of Iowa. Linn County is followed by Johnson, Black Hawk, Pottawattamie, and Polk counties in total number of projects. North-western Iowa has received the least number of projects compared to the other regions due to rare flood events.

The data analytics framework provides mitigation analysis in mul- tiple scales such as state-wide, county, city, and property level. In Fig. 6, the city level summary panel is visualized for the City of Vinton. The city-level analysis summarizes the number of records, accounts, and applicants for mitigation grants, total eligible obligated, and federal obligated dollars for the city. The system also allows us to examine the total amount of acquisition cost, number of classified flooding events

after the acquisition date, number of affected buildings, and total damage as an avoided loss. In addition, damage estimates for the existing buildings are also given to support decisions for future property acquisitions or other mitigation activities. Overall, a general idea about mitigation grants, existing mitigation efforts, and future mitigation possibilities are provided in the interface to decision-makers.

Fig. 7 illustrates 20 counties in Iowa that had the most disaster declarations over the last 65 years due to flooding events compared to their populations and the number of property acquisitions. Pre- and post-hazard mitigation activities are generally funded by federal aid so that a federal disaster declaration is required to receive the aid. On the other hand, population density should be investigated to understand the distribution of the property buyouts in the state to understand the main driving factor behind the decisions for property acquisitions. We found that the number of disaster declarations and the population of the community does not show a strong correlation to property acquisitions. Because property acquisitions are made voluntarily, the public may be willing to participate based on the amount of damage on their property

Fig. 6. Mitigation projects and property acquisition summary for cities.

Fig. 7. Project and acquisition analysis summary for cities.

E. Yildirim and I. Demir

International Journal of Disaster Risk Reduction 56 (2021) 102113

7

or the participation from their community. Willingness to participate in property buyout programs can be investigated further with additional research and surveys.

3.2. Middle Cedar Watershed Case Study

Middle Cedar watershed has experienced severe flooding events in history and has several large communities in the State of Iowa (Fig. 8). The majority of the buyouts in the watershed are made in Cedar Falls, Cedar Rapids, and Waterloo. In this study, benefit-cost analysis is carried out for all properties in the Middle Cedar watershed for historical and future climate scenarios.

As shown in Figs. 4 and 5, communities in the Middle Cedar water- shed received relatively larger amounts of flood mitigation grants compared to other regions in Iowa. Therefore, Middle Cedar is selected

to investigate benefit-cost analysis for property acquisitions based on historical and projected flooding events. 282 property buyouts are analyzed in the Middle Cedar watershed between 2009 and 2017. His- torical gauge records and streamflow projections from 2020 to 2050 are processed to identify flood events and their magnitude. Following, corresponding flood maps are used based on classified flood events to estimate historical and future flood losses in the communities. In Fig. 9, direct flood losses for historical and future flood events are generated for existing properties in three major cities of the study region. Red columns represent the direct flood losses between 2009 and 2017. Blue and green columns represent future damage estimates as a combination of struc- tural and content losses between 2020 and 2050.

In Fig. 10, direct loss per structure is presented for selected Middle Cedar communities. Waterloo shows the highest loss per structure. This is a strong indicator that the community has vulnerable industrial or commercial structures within the flood-prone areas. In the property dataset, commercial and industrial buildings are relatively more valu- able compared to residential, governmental, and public buildings which are reflected in structural and content damage.

Fig. 8. Middle Cedar watershed study area for benefit-cost analysis.

Fig. 9. Avoided and projected total direct losses for major cities in Middle Cedar watershed.

Fig. 10. Direct losses per structure in Middle Cedar communities (2009–2017).

E. Yildirim and I. Demir

International Journal of Disaster Risk Reduction 56 (2021) 102113

8

In Table 2, historical and projected avoided losses, number of properties, number of events, and community level BCAs are shared for the property buyouts in the Middle Cedar watershed. The city of Cedar Falls has the highest portion of analyzed property buyouts in this study. Between the acquisition date which is 2009 for most of the buyouts and 2019, nearly half of the analyzed buyouts are found to be successful. The community BCR for Cedar Falls is estimated as 0.86 by only considering direct losses. Considering projected streamflow values that are outputs of climate scenarios (a2 and a1fi), overall buyouts in the Middle Cedar watershed are estimated to be successful in both scenarios.

In Table 3, estimated direct losses for existing properties in the Middle Cedar watershed is shared for historical events and future sce- narios. Cedar Falls, Waterloo, and Vinton are found to be the most vulnerable between 2009 and 2019. Also, projected streamflow values are most likely to cause heavy losses in these communities. It can be

concluded that future mitigation efforts may be needed for these communities.

One of the capabilities of the data analytics framework is generating analysis at the property level. In Fig. 11, results and visualization of property level analysis from the framework are demonstrated. The sys- tem generates avoided and historical losses for acquired and existing properties respectively. Building value or acquisition cost, property- specific id, the applicant for the property acquisitions, and other related information are also provided in the information panel. This allows investigating the benefit-cost ratio for the historical buyout, and potential future buyout decisions for an existing property in the com- munities. Acquired properties are illustrated with red markers and existing properties are shown with yellow markers. When the property is selected, the data analytics tool reveals historical floods that affected the property. Date of the events based on USGS gauges, return periods for

Table 2 Estimated avoided direct flood losses for major cities in Middle Cedar.

City Total Acquisitions

Historical (2009–2017) Low Emission – a2 (2020–2050) Fossil Intensive – a1fi (2020–2050)

Avoided Loss

Properties Flood Events

BCR Projected Avoided Loss

Properties Flood Events

BCR Projected Avoided Loss

Properties Flood Events

BCR

Cedar Falls

$5.67 M $4.91 M 121 4 0.86 $26.94 M 121 18 4.72 $36.36 M 121 19 6.35

Cedar Rapids

$2.21 M $0.20 M 2 2 0.09 $2.71 M 52 7 1.25 $6.72 M 52 15 3.05

Palo $0.77 M $0.11 M 6 5 0.14 $0.90 M 15 19 1.17 $1.37 M 15 19 1.75 Vinton $1.24 M $0.44 M 16 8 0.37 $1.92 M 26 26 1.63 $2.81 M 26 26 2.35 Waterloo $3.94 M $1.13 M 44 5 0.29 $14.53 M 50 14 3.72 $20.34 M 50 19 5.22

Table 3 Estimated potential direct flood losses for major cities in Middle Cedar.

County Cost of Total Acquisitions

Historical (2009–2017) Low Emission – a2 (2020–2050) Fossil Intensive – a1fi (2020–2050)

Total Avoided Loss

Properties Events Projected Avoided Loss

Properties Events Projected Avoided Loss

Properties Events

Cedar Falls

$487.38 M $69.31 M 414 6 $417.61 M 554 28 $605.81 M 555 26

Hudson $30.15 M $1.92 M 12 5 $5.99 M 12 19 $8.47 M 52 20 Palo $70.04 M $0.69 M 17 5 $0.13 M 136 14 $23.03 M 136 19 Vinton $84.97 M $18.33 M 103 7 $91.44 M 140 26 $132.09 M 140 26 Waterloo $942.54 M $51.59 M 204 6 $416.26 M 817 23 $586.82 M 817 25

Fig. 11. Avoided and historical flood losses for acquired (left) and existing (right) property.

E. Yildirim and I. Demir

International Journal of Disaster Risk Reduction 56 (2021) 102113

9

flooding events, the estimated total of direct content and structural flood losses, and total loss over the years are given.

Summary of property acquisitions in the Middle Cedar watershed compared to disaster declarations and population are provided in Fig. 12. Similar to Iowa level analysis, property buyouts are also not strongly correlated with the number of disaster declarations. However, the population is an indicator of the number of property buyouts in the Middle Cedar watershed.

4. Conclusion

This study presents a web-based data analytics framework for his- torical mitigation projects and property acquisitions with a case study for the State of Iowa. The case study is carried out to investigate the benefit-cost analysis of historical buyouts and potential property ac- quisitions for existing properties based on climate scenarios in the future. The framework provides results in multiple scopes such as data source, program type, location, time, and multiple occupancies. Unlike GIS and desktop level applications, web systems allow non-technical users to explore the comprehensive mitigation analysis results and generate customized reports for their property or community with limited technical knowledge. This allows improving daily workflow for decision-makers and state agencies to analyze the vast amount of data at various geospatial and temporal scales through a user friendly web interface. Due to a lack of standardized parcel information, data collection, spatial correction, and other data handling tasks require laborious work to utilize in the study for other regions.

The case study reveals that the majority of the property buyouts are successful between the year of the acquisitions and today. The projected streamflow data is also a strong indicator of the success of property buyouts by avoiding significant damage in the future. It is important to emphasize that the case study is considered direct flood losses for structural and content damages. On the other hand, many indirect losses are hard to quantify but cannot be ignored. These indirect losses would increase the current BCAs, therefore the success of the buyouts. As new methodologies are introduced to quantify indirect losses, they can be integrated into existing BCA to enhance estimations in the framework. In the future, the framework can be expanded by adding other mitigation options by researchers to carry out custom and more comprehensive mitigation analyses. However, some mitigation applications (e.g., le- vees, drainage reservoirs) will require custom flood maps. Thus, future frameworks must consider using dynamic flood maps rather than static flood maps due to the changes that come along with these mitigations.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This project was supported partially by the Iowa Department of Homeland Security and Emergency Management Department (H547600-CG, H547500-CG).

References

[1] W.E. Highfield, S.D. Brody, Evaluating the effectiveness of local mitigation activities in reducing flood losses, Nat. Hazards Rev. 14 (4) (2012) 229–236.

[2] M.W. Downton, R.A. Pielke Jr., Discretion without accountability: politics, flood damage, and climate, Nat. Hazards Rev. 2 (4) (2001) 157–166.

[3] Federal Emergency Management Agency (FEMA) Website, Retrieved online Available online: https://www.fema.gov/data-visualization-summary-disas ter-declarations-and-grants. (Accessed 15 February 2020).

[4] D. Wang, M. Hejazi, Quantifying the relative contribution of the climate and direct human impacts on mean annual streamflow in the contiguous United States, Water Resour. Res. 47 (10) (2011).

[5] D.J. Wuebbles, K. Hayhoe, Climate change projections for the United States Midwest, Mitig. Adapt. Strategies Glob. Change 9 (4) (2004) 335–363.

[6] I. Mallakpour, G. Villarini, The changing nature of flooding across the central United States, Nat. Clim. Change 5 (3) (2015) 250.

[7] L.J. Weber, M. Muste, A.A. Bradley, A.A. Amado, I. Demir, C.W. Drake, W. F. Krajewski, T.J. Loeser, M.S. Politano, B.R. Shea, N.W. Thomas, The Iowa Watersheds Project: Iowa’s prototype for engaging communities and professionals in watershed hazard mitigation, Int. J. River Basin Manag. 16 (3) (2018) 315–328.

[8] D.J. Thampapillai, W.F. Musgrave, Flood damage mitigation: a review of structural and nonstructural measures and alternative decision frameworks, Water Resour. Res. 21 (4) (1985) 411–424.

[9] F. Scionti, M.G. Miguez, G. Barbaro, M.M. De Sousa, G. Foti, C. Canale, Integrated methodology for urban flood risk mitigation in cittanova, Italy, J. Water Resour. Plann. Manag. 144 (10) (2018), 05018013.

[10] I. Andjelkovic, Guidelines on non-structural measures in urban flood management, in: International Hydrological Programme (IHP), United Nations Educational, Scientific and Cultural Organization (UNESCO), 2001.

[11] T. Egli, Non-structural flood plain management: measures and their effectiveness, in: International Commission for the Protection of the Rhine (ICPR), 2002.

[12] B. Meunier, V. Merwade, Prioritizing levee repairs: a case study for the city of Indianapolis, Indiana, Nat. Hazards 72 (2) (2014) 997–1019.

[13] W.H. Desvousges, M.C. Naughton, G.R. Parsons, Benefit transfer: conceptual problems in estimating water quality benefits using existing studies, Water Resour. Res. 28 (3) (1992) 675–683.

[14] K.J. Boyle, J.C. Bergstrom, Benefit transfer studies: myths, pragmatism, and idealism, Water Resour. Res. 28 (3) (1992) 657–663.

[15] S.D. Brody, W.E. Highfield, Open space protection and flood mitigation: a national study, Land Use Pol. 32 (2013) 89–95.

[16] S.D. Brody, J.E. Kang, S. Bernhardt, Identifying factors influencing flood mitigation at the local level in Texas and Florida: the role of organizational capacity, Nat. Hazards 52 (1) (2010) 167–184.

[17] C. Kousky, M. Walls, Floodplain conservation as a flood mitigation strategy: examining costs and benefits, Ecol. Econ. 104 (2014) 119–128.

[18] E. Roy, J. Rousselle, J. Lacroix, Flood damage reduction program (FDRP) in quebec: case study of the chaudiere River, Nat. Hazards 28 (2–3) (2003) 387–405.

[19] D.M. Knobloch, Moving a community in the aftermath of the great 1993 Midwest flood, J. Contemp. Water Res. Educ. 130 (1) (2005) 41–45.

[20] M. Das, A. Das, S. Momin, R. Pandey, Mapping the effect of climate change on community livelihood vulnerability in the riparian region of Gangatic Plain, India, Ecol. Indicat. 119 (2020) 106815.

[21] S. Sekhri, P. Kumar, C. Fuerst, R. Pandey, Mountain specific multi-hazard risk management framework (MSMRMF): assessment and mitigation of multi-hazard and climate change risk in the Indian Himalayan Region, Ecol. Indicat. 118 (2020) 106700.

Fig. 12. Comparison between property acquisitions and population/disaster declarations.

E. Yildirim and I. Demir

International Journal of Disaster Risk Reduction 56 (2021) 102113

10

[22] Federal Emergency Management Agency (FEMA), Hazard mitigation grant program, Available online: http://www.fema.gov/hazard-mitigation-grant-progra m, 2020.

[23] E. Marino, Adaptation privilege and Voluntary Buyouts: perspectives on ethnocentrism in sea level rise relocation and retreat policies in the US, Global Environ. Change 49 (2018) 10–13.

[24] Federal Emergency Management Agency (FEMA), T. Robert, Stafford Disaster Relief and Emergency Assistance Act, as Amended, FEMA, Washington, DC, USA, 2013.

[25] T. Sieg, T. Schinko, K. Vogel, R. Mechler, B. Merz, H. Kreibich, Integrated assessment of short-term direct and indirect economic flood impacts including uncertainty quantification, PloS One 14 (4) (2019), e0212932.

[26] B. Merz, A. Thieken, H. Kreibich, Quantification of socio-economic flood risks, in: Flood Risk Assessment and Management, Springer, Dordrecht, 2011, pp. 229–247.

[27] B. Merz, H. Kreibich, R. Schwarze, A. Thieken, Review article" Assessment of economic flood damage, Nat. Hazards Earth Syst. Sci. 10 (8) (2010) 1697–1724.

[28] H. Moel, B. Jongman, H. Kreibich, B. Merz, E. Penning-Rowsell, P.J. Ward, Flood risk assessments at different spatial scales, Mitig. Adapt. Strategies Glob. Change 20 (6) (2015) 865–890.

[29] C. Arrighi, L. Rossi, E. Trasforini, R. Rudari, L. Ferraris, M. Brugioni, F. Castelli, Quantification of flood risk mitigation benefits: a building-scale damage assessment through the RASOR platform, J. Environ. Manag. 207 (2018) 92–104.

[30] J. Dierauer, N. Pinter, J.W. Remo, Evaluation of levee setbacks for flood-loss reduction, Middle Mississippi River, USA, J. Hydrol. 450 (2012) 1–8.

[31] S. Hallegatte, C. Green, R.J. Nicholls, J. Corfee-Morlot, Future flood losses in major coastal cities, Nat. Clim. Change 3 (9) (2013) 802–806.

[32] M. Sit, I. Demir, Decentralized Flood Forecasting Using Deep Neural Networks, 2019 arXiv preprint arXiv:1902.02308.

[33] Z. Xiang, J. Yan, I. Demir, A rainfall-runoff model with LSTM-based sequence-to- sequence learning, Water Resour. Res. 56 (1) (2020), e2019WR025326.

[34] L.M. Bouwer, P. Bubeck, J.C. Aerts, Changes in future flood risk due to climate and development in a Dutch polder area, Global Environ. Change 20 (3) (2010) 463–471.

[35] E. Yildirim, Hazus-MH Flood Loss Estimation on a Web-Based System, Master’s Thesis, University of Iowa, 2017.

[36] R. De Risi, F. Jalayer, F. De Paola, S. Carozza, N. Yonas, M. Giugni, P. Gasparini, From flood risk mapping toward reducing vulnerability: the case of Addis Ababa, Nat. Hazards 100 (1) (2020) 387–415.

[37] S. Carozza, F. Jalayer, R. De Risi, G. Manfredi, E. Mbuya, Comparing alternative flood mitigation strategies for non-engineered masonry structures using demand and capacity fac-tored design, in: 12th International Conference on Application of Statistics and Probability in Civil Engineering, Vancouver2015, 2015.

[38] H. Xu, M. Windsor, M. Muste, I. Demir, A web-based decision support system for collaborative mitigation of multiple water-related hazards using serious gaming, J. Environ. Manag. 255 (2020) 109887.

[39] A. Carson, M. Windsor, H. Hill, T. Haigh, N. Wall, J. Smith, R. Olsen, D. Bathke, I. Demir, M. Muste, Serious gaming for participatory planning of multi-hazard mitigation, Int. J. River Basin Manag. 16 (3) (2018) 379–391.

[40] I. Demir, E. Yildirim, Y. Sermet, M.A. Sit, FLOODSS: Iowa flood information system as a generalized flood cyberinfrastructure, Int. J. River Basin Manag. 16 (3) (2018) 393–400.

[41] Y. Sermet, I. Demir, M. Muste, A serious gaming framework for decision support on hydrological hazards, Sci. Total Environ. (2020) 138895.

[42] I. Demir, M.B. Beck, April. GWIS: a prototype information system for Georgia watersheds, in: Georgia Water Resources Conference: Regional Water Management Opportunities, UGA, Athens, GA, US, 2009.

[43] I. Demir, R. Szczepanek, Optimization of river network representation data models for web-based systems, Earth Space Sci. 4 (6) (2017) 336–347.

[44] M. Sit, Y. Sermet, I. Demir, Optimized watershed delineation library for server-side and client-side web applications, Open Geospatial Data, Softw. Stand. 4 (1) (2019) 8.

[45] R. Agliamzanov, M. Sit, I. Demir, Hydrology@ Home: a distributed volunteer computing framework for hydrological research and applications, J. Hydroinf. 22 (2) (2020) 235–248.

[46] Y. Sermet, I. Demir, Towards an information centric flood ontology for information management and communication, Earth Sci. India 12 (4) (2019) 541–551.

[47] Y. Sermet, I. Demir, Flood action VR: a virtual reality framework for disaster awareness and emergency response training, in: ACM SIGGRAPH 2019 Posters, 2019, pp. 1–2.

[48] E. Yildirim, I. Demir, An integrated web framework for HAZUS-MH flood loss estimation analysis, Nat. Hazards 99 (1) (2019) 275–286.

[49] F. Quintero, R. Mantilla, C. Anderson, D. Claman, W. Krajewski, Assessment of changes in flood frequency due to the effects of climate change: implications for engineering design, Hydrology 5 (1) (2018) 19.

[50] M.L. Parry, C. Rosenzweig, A. Iglesias, M. Livermore, G. Fischer, Effects of climate change on global food production under SRES emissions and socio-economic scenarios, Global Environ. Change 14 (1) (2004) 53–67.

[51] W. IPCC, Special Report on Emissions Scenarios. Intergovernmental Panel on Climate Change Special Reports on Climate Change, Cambridge University Press, Cambridge, 2000, p. 570.

[52] N.W. Arnell, Climate change and global water resources: SRES emissions and socio- economic scenarios, Global Environ. Change 14 (1) (2004) 31–52.

[53] D. Gilles, N. Young, H. Schroeder, J. Piotrowski, Y.J. Chang, Inundation mapping initiatives of the Iowa Flood Center: statewide coverage and detailed urban flooding analysis, Water 4 (1) (2012) 85–106.

[54] E. White, Establishing Long-Term Cost Effectiveness of Fema Buyouts: a Loss Avoidance Study of the Acquisition/demolition of 22 Properties in Shepherdsville, 2011. Kentucky.

[55] C. André, D. Monfort, M. Bouzit, C. Vinchon, Contribution of insurance data to cost assessment of coastal flood damage to residential buildings: insights gained from Johanna (2008) and Xynthia (2010) storm events, Nat. Hazards Earth Syst. Sci. 13 (8) (2013) 2003–2012.

[56] H. Kreibich, A.H. Thieken, T. Petrow, M. Müller, B. Merz, Flood loss reduction of private households due to building precautionary measures–lessons learned from the Elbe flood in August 2002, Nat. Hazards Earth Syst. Sci. 5 (1) (2005) 117–126.

[58] J.W. Remo, M. Carlson, N. Pinter, Hydraulic and flood-loss modeling of levee, floodplain, and river management strategies, Middle Mississippi River, USA, Nat. Hazards 61 (2) (2012) 551–575.

[59] R. Ready, S. Navrud, International benefit transfer: methods and validity tests, Ecol. Econ. 60 (2) (2006) 429–434.

[60] C. Scawthorn, P. Flores, N. Blais, H. Seligson, E. Tate, S. Chang, M. Lawrence, HAZUS-MH flood loss estimation methodology. II. Damage and loss assessment, Nat. Hazards Rev. 7 (2) (2006) 72–81.

[61] Y. Sermet, P. Villanueva, M.A. Sit, I. Demir, Crowdsourced approaches for stage measurements at ungauged locations using smartphones, Hydrol. Sci. J. 65 (5) (2020) 813–822.

[64] David Brookshire, Dale Whittington, Water Resources Issues in the Developing Countries, Water Resour. Res. (1993), https://doi.org/10.1029/92WR02988.

[65] Alex Greer, Sherri Brokopp Binder, A Historical Assessment of Home Buyout Policy: Are We Learning or Just Failing? Hous. Pol. Debate (2016) https://doi.org/ 10.1080/10511482.2016.1245209.

E. Yildirim and I. Demir