Flooding, particularly in cities and towns, has long been a costly recurring natural hazard both in terms of the number of lives lost as well as the overall cost in damages to property, public works infrastructure, and natural resources. As the world becomes more aware of the implications of the changing climate and what effects it might have on water resources, concern has risen about society’s ability to anticipate and prepare for more extreme climatic events such as urban flooding.
In response to this concern, a team of researchers at the University of Oklahoma designed and carried out a project that provides urban floodplain managers and other interested stakeholders with visual animations and images of how future flood events may impact urban watersheds.
In this website, we provide some general information on the global and national threat of urban flooding and offer the reader a detailed narrative of the urban flooding and climate change project. We discuss what effect a changing climate may have on urban flooding; current flood management strategies; how the research project was designed to address these issues and achieve its goals including visualization; and what the team concluded from the project. The multi-year project was completed in 2014.
Members of the research team included:
Profs. Scott Greene, Yang Hong, Mark Meo*, and Baxter Vieux,
with the research assistance of
Jonathan Looper and Zhanming Wan, and Amy Goodin of OU POLL.
The project was supported by the
Profs. Greene and Meo are faculty members in the Department of Geography and Environmental Sustainability. Profs. Hong and Vieux are faculty and retired faculty members, respectively, in the School of Civil Engineering and Environmental Science. Jonathan Looper and Zhanming Wan are former and current Graduate Research Assistants, respectively, at OU.
Urban Flooding and Climate Change - Visualizing the Impacts
Urban flooding is a chronically recurring natural hazard that causes destruction around the world, and is expected to become even worse as global climate and climate variability change. Recent studies have suggested that adverse impacts from urban flooding can be diminished through effective adaptive and mitigation actions taken by local watershed managers and their governments. In order to achieve these goals, however, climate scientists, hydrologists, and urban flood managers will need to find creative ways to collaborate with each other and interested parties, or stakeholder groups. This project was undertaken to test the effectiveness of watershed flood modeling and visualization as a technique that can inform local planners and affiliated stakeholders about the possible impacts that might accrue under a future climate.
With the support of NOAA’s Sectoral Applications Research Program, a team of faculty and graduate students at the University of Oklahoma tested a watershed modeling and visualization procedure in five urban watersheds located in Oklahoma City and Tulsa, (OK), and Austin, Dallas, and Houston, (TX). In collaboration with the identified urban watershed floodplain manager, the team first developed a watershed model for a site selected by the floodplain manager, convened an educational workshop with the manager to describe the resultant watershed model and its visualization, and then provided an Internet-based survey that included this material for distribution to key stakeholders. The two-year project was completed in 2014.
The watershed survey results indicate that stakeholders prefer the watershed modeling and visualization technique over static maps (71%), and would welcome wider application and use of it (67%). Respondents from the five study sites found the watershed animations to be informative (27% extremely so), useful (29% extremely so), and credible (22% extremely so) for planning and management. These findings support the development of creative techniques that link science and engineering with practical application, suggest that technical outreach by research institutions could advance urban climate adaptation and planning capacity.
Urban flooding is a naturally-occurring hazard that affects cities and regions around the world, and is expected to become even more damaging in the future (Jha et al., 2011). With a world population that is now above 7 billion and rising, pressure on urbanized areas to accommodate more people has altered land uses to facilitate flooding. With over half the world’s population, the expansion of impervious surfaces and the removal of water recharge areas and natural flood barriers such as wetlands have increased the likelihood of flood damages in cities. Moreover, the manifestation of climate changes through more severe storms, rising sea levels and altered precipitation regimes is increasing the probability of damaging flood events in the future. It is not unusual to read about severe flood events in the United States, Canada, Europe, and Asia.
Currently, disaster statistics appear to show that flood events are becoming more frequent (Jha et al. 2011). In the last three decades, flood events have become more frequent than droughts and non-climate related hazards.
Figure 1 Trends in Reported Water-Related Disasters
Among the regions of the world most affected by floods are the Asia-Pacific region, followed by Africa. During the last 30 years, flood events in Asia totaled about 40% of all events with the Americas accounting for 25%. The most affected country is China, followed by Indonesia, Bangladesh, Vietnam, Thailand, and Pakistan. The trend in reported flood events reveals a clear increase over time.
Figure 2 Number of Reported Flood Events
Damages from floods are also increasing as are the number of people who are affected by them. While the 10-year average of flood damages is about $20 billion, some bad years can run as high as $40 billion.
Figure 3 Global Losses Due to Flooding
In the United States, flood losses are high. According to Stanley Changnon, the nation’s total property losses from floods in the period 1972-2006 totaled $94 billion, representing an annual average of $2.67 billion and an average of $176 million per event (Changnon 2008). The average loss for flood-only catastrophes was $80 million, with $88 million for convective storm floods, $70 million for snow-melt floods, and $973 million for hurricane-related floods. The distribution of floods and their losses across the nation varied by type of flood. As expected, hurricane floods were most frequent in the Southeast, whereas snow-melt floods were most common in the Northeast and Central regions (Changnon 2008). The flood-only cases and the convective storm-related floods were both most frequent in the South and second most frequent in the Central region. Texas was the state with the greatest number of flood-related losses, being first for the flood-only and convective storm cases and being the tenth-ranked state for hurricane flood losses. Other high loss ranked states included New York, Florida, Pennsylvania, and Missouri. Losses were greatest in the South ($23.7 billion) and Central ($20.4 billion) regions, and losses were least in the Northwest and West regions. Changnon (2008) states that this pattern is expected to get worse since climate studies have shown upward trends in annual precipitation and heavy rain events since the 1930s; other studies have identified growing societal vulnerability to floods; and the nation’s flood control structures have been degrading with time.
The variability in flood damages over time and location is striking. For example, 2011 was a year of record-breaking prolonged floods along some of the nation’s largest rivers (National Weather Service 2011). Direct flood damages totaled $8.41 billion, which was 108% of the thirty-year average of $7.82 billion (for the period 1981-2010). There were 108 flood-related deaths of which 61 were vehicle related and 71 were attributed to flash flood events. In the following year, total damages were just $.5 billion, 6% of the thirty-year average (National Weather Service 2012). While the onset of drought across most of the nation reduced flooding, the inherent variability associated with flooding must be weighed carefully in decision making. Rapid shifts in variability with extreme precipitation most recently were experienced by the catastrophic urban flooding that struck cities in Texas and Oklahoma in May, 2015.
For over two decades the Intergovernmental Panel on Climate Change (IPCC) has been monitoring, analyzing, and reporting on the impacts generated by the changing climate. Once considered a future-oriented problem caused by the continued emission of trace greenhouse gases, such as carbon dioxide, the impacts associated with a warming climate are beginning to manifest. In their fourth assessment report, published in 2007, the IPCC reported that climate change was having a significant effect on the hydrologic cycle, and that by the middle of the century, annual average river runoff and water availability are projected to increase by 10-40% at higher latitudes and in some wet tropical areas (IPCC 2007). In some regions, such as Northern Europe, changes in the hydrologic cycle are expected to generate increased risk of inland flash flooding as well as coastal flooding, the latter influenced by sea-level rise. In North America, global warming is projected to cause decreased snowpack, more winter flooding, and reduced summer flows, exacerbating competition for over-allocated water resources (IPCC 2007).
According to the National Climate Assessment (NCA) for the United States, variability in water resource and associate impacts such as urban flooding are expected to become more severe in certain locations. In the Great Plains, heaviest daily precipitation is expected to increase above the time period 1971 to 2000 by about seven days each year (Melillo et al. 2014). More generally, the NCA determined that:
FLOOD MANAGEMENT & ADAPTATION STRATEGIES
Cities have long struggled to overcome the risks posed by flood events, but with uneven success. Due to the mixture of different hydrologic and engineered infrastructure systems within variable and changing land uses, both current and projected challenges posed by increased severity of flooding admit to a bundle of principles for effective flood hazard mitigation, known as integrated flood risk management.
Ideally, integrated flood risk management seeks to characterize the relevance of flooding to the attention of all affected parties to the process of risk reduction and management. This approach necessitates:
However, the degree to which cities can attain these goals may be facilitated, or hampered by, national, regional, and local leadership, public concern, useable knowledge, institutional and legal impediments, adequate resources, and community awareness (Jha et al. 2012).
Among those cities that have demonstrated an ability to address these issues in a more integrative manner, success with flood hazard mitigation has helped to improve the overall capacity of the city to address other climate-related hazards. In Oklahoma, for example, effective flood hazard mitigation in Tulsa has served as a basis for addressing related water quality and air quality concerns. City-wide flood mitigation has been addressed through a more coherent process of watershed management, land use restrictions, effective public communication, dedicated financing for infrastructure improvements, and local human resources capacity building (Patton and Chakos 2009) The incorporation of state-of-the-art flood hazard warning and modeling technologies has improved the capability of cities as diverse as Austin, Texas and Cologne, Germany, for example, to prepare for future climate-oriented contingencies (Jha et al. 2012)
With respect to the world, the large difference in local capacities to address flooding coupled with the growing threat of more severe floods led the IPCC to publish a set of recommendations recently for managing the risks of extreme events (IPCC 2012). Among the measures recommended are the following:
Concerned that climate change will stress the nation’s aging infrastructure with more heavy precipitation and runoff events, the NCA report (Melillo et al. 2014) offers a series of recommendations for improved adaption. In addition to the ones that replicate those listed above, the NCA notes that “Building networks, partnerships, and support systems has been identified as a major asset in building adaptive capacity.” The NCA notes that the economic, social, and environmental implications of climate change-induced water cycle changes are very significant, as is the cost of inaction. Adaptation responses will therefor need to address uncertainties, be proactive, integrated, and iterative, and be developed through well-informed stakeholder decision processes (Melillo et al. 2014).
RESEARCH DESIGN & METHODS
With a high degree of variability in both the physical and human environments, the aim of the research design is to develop a process that links current knowledge about climate and weather to a modeling technique that can accurately and realistically portray current and future flood events to stakeholders in an understandable manner. The goal is to determine if local watershed stakeholders prefer visual animations of flood events to the conventional graphics and charts currently used in planning and decision making. Since concerns about local flood events are, in fact, local and vary greatly from place to place, the research team adopted an iterative research design that links the analytic capabilities of the team with the interests of local urban floodplain managers and stakeholders. Stakeholders for a local watershed include the professional analysts and floodplain managers responsible for, or with an in interest in, the watershed’s behavior and sustainability, as well as citizens who have a clear stake in the overall management of the watershed. The team’s approach is shown below in Figure 4.
Figure 4 Iterative Research Design for Urban Flooding Project
With the cooperation of local floodplain managers, the research team selected five cities in Oklahoma and Texas for detailed study. Both states have experienced a history of devastating floods and are concerned about mitigating potential damages that might arise in the future. The case study cities include: Oklahoma City and Tulsa in Oklahoma, and Austin, Dallas, and Houston in Texas.
The case study watersheds in each city are:
|OKLAHOMA||OKLAHOMA CITY||CHISHOLM CREEK|
Figure 5 Map of the Locations of Case Study Watersheds in Oklahoma and Texas
Once a candidate watershed was identified, the team conducted its analysis of the watershed by investigating historic precipitation frequencies and projected precipitation under climate change conditions. This information was then modeled using the Vflo physics-based hydrologic model and visualized for presentation. The steps are illustrated in Figure 6.
Figure 6 Steps in Urban Watershed Analysis
Estimates of precipitation under climate change were conducted in the following manner.
The distributed hydrologic model used in the analysis is based on the actual physical topography and geographic character of the watershed (see [Vieux 2004] for details). Geospatial data were collected and assembled that defined soils, topography, land use/cover, and imperviousness for the five urban basins in Oklahoma and Texas. Model parameters were derived from the geospatial data to simulate infiltration and runoff processes. Model grids were defined for each basin variously from 10 to 100 m resolution. For purposes of sensitivity testing, precipitation depths for 2, 5, 10, 25, 50, 100, and 500 yr return intervals were assembled for each of the basins. For testing current and future climate scenarios, precipitation data at 3-hr intervals were assembled for continuous simulation input. Initial model runs were then made to confirm model parameter choices and validity. A diagram of the parameters used in the distributed hydrologic model is presented in Figure 7 below.
Figure 7 Vflo Distributed Hydrology for Chisholm Creek
Model simulations of each watershed were run to demonstrate the degree of agreement between the modeled watershed and historic precipitation frequencies. Projections of precipitation frequencies under climate change were modeled as well for a variety of plausible scenarios. Both historic and projected flooding events were then visualized as an animation using Google Earth as geographic template. A static image of Chisholm Creek at maximum inundation from a 100-year event is shown in Figure 8 below.
Figure 8 Visualization of Chisholm Creek
Examples of the watershed animations are presented below for Fred Creek in Tulsa and Joe’s Creek in Dallas. The Fred Creek animations portray inundation from a 100-year rainfall event and for the same event with an increment of 10%. The three animations for Joe’s Creek also portray inundation from a 100-year event as well as an increment of 10%. In addition, the research team designed a scenario that was based on an increase of 23% that was derived from the North American Regional Climate Change Assessment Program (see Figure 9 below).
Figure 9 Precipitation – Frequency Climate Forecast
Fred Creek Wastershed in Oklahoma
Inundation from 100-Year Event
Inundation from 100-Year Event plus 10%
Joe's Creek Watershed in Texas
Inundation from 100-Year Event
Inundation from 100-Year Event plus 10%
Inundation from 100-Year Event plus 23%
Upon completion of the watershed analysis, the team made a formal presentation to each floodplain manager to explain the assumptions used in the analysis and the procedures that were followed. The floodplain managers were able to view the visual animation of the flooded watershed and ask questions about the process (see Figure 10 below). This step helped to build technical credibility between the research team and the city floodplain managers.
Figure 10 Prof. Baxter Vieux and Jonathan Looper presenting at the Dallas workshop
Next, the team prepared a watershed survey based on its analysis and presented it to the city floodplain manager for his review. The survey included both the graphical information and the visual animation presented to the floodplain manger. Qualtrics software, which can accommodate animations, was used for the survey. Once the survey was approved, the floodplain manager distributed the survey to a small group of watershed stakeholders. The Qualtrics watershed survey was Internet-based and all participating stakeholders were granted anonymity.
For each watershed, the target audience for the survey was a small group (~10 to 15) of knowledgeable stakeholders who are well-versed in the key issues associated with each of the case study watersheds. This group could be either technically proficient or closely involved in watershed management activities. The survey was designed to be answered in about 10 minutes. The URL for the web-based survey was sent to the stakeholders by the city watershed manager, and a follow-up reminder was distributed after one week.
The survey was comprised of twelve questions, several of which had follow-up queries. The questions used a variety a scale questions that included, for example, the 6-items: Strongly Agree, Agree, Neither Agree nor Disagree, Disagree, Strongly Disagree, and Don’t Know. Other questions asked respondents to indicate their preferences on a 1-10 point interval scale from Not at all Useful to Extremely Useful and Don’t Know. Other questions asked for a Yes, No, or Don’t Know response.
The questions used a variety a scale questions that included, for example, the 6-items: Strongly Agree, Agree, Neither Agree nor Disagree, Disagree, Strongly Disagree, and Don’t Know. Other questions asked respondents to indicate their preferences on a 1-10 point interval scale from Not at all Useful to Extremely Useful and Don’t Know. Other questions asked for a Yes, No, or Don’t Know response.
FINDINGS & IMPLICATIONS
The total number of survey respondents came to 66, which was an average of about 13 for each watershed. While responses were collected for all the questions, only some key results are presented here.
The survey results revealed that watershed stakeholders in all five study sites were very favorably inclined toward the use of animations in floodplain management. For instance:
Question 7 – 61.2% agreed and 22.4% strongly agreed that hydrologic models and images are effective in discussions with stakeholders about capital improvement project needs.
Question 8 – 59.2% agreed and 16.3% strongly agreed that hydrologic models enable people to understand increases in flooding due to increases in precipitation.
Question 10(a) – 71.1% indicated that the watershed animation was more effective than a static map.
Question 10(b) – 66.7% indicated that they would like to see more watershed animations.
Question 11 – 29.3% indicated that the animations were extremely useful; 26.8% indicated that they were extremely informative; and 22% indicated that they were extremely credible. In addition, 50% of respondents indicated that they preferred animations over maps and charts.
In general, watershed managers were receptive to the University of Oklahoma project. The research team believes that the workshops presented to the watershed managers and their staff were key to establishing a shared understanding of the hydrology and policy issues associated with specific floodplain management issues as well as building a working relationship with the city’s personnel. Credibility and professional legitimacy were achieved by the team by presenting watershed models and animation to city technical staff and answering questions.
The research team also learned that watershed managers were more inclined toward relying on the record of historic hydrological events instead of climate change projections of future events. As a result, the team designed animations for presentation that were more compatible with these viewpoints. The modeling and visualization approach was well received by managers and technical staff. Accordingly, more detailed visual imagery of flood events was requested by respondents.
This project was designed to address the growing need to improve the level of communication and understanding between the climate research community, hydrologists, watershed managers, and the stakeholder community about the need to develop adequate and effective adaptation strategies for flooding events brought about through climate variability and change. With the utilization of existing technology, including Geographic Information Systems, hydrologic modeling, animated visual imagery, and the Internet, this project demonstrated that interactive engagement among the key groups can enhance understanding and preparedness for future flooding events.
While formal climate projections may lack some credibility with local officials and stakeholders at the current time, familiarity with local watershed hydrology is important for researchers to help local planners and managers design effective strategies. To this end, expanded use of watershed modeling and visualization techniques can prove to be useful for adaptation planning and management.
The University of Oklahoma research team would like to acknowledge and thank the following individuals for making the watershed case studies possible: Adhir Agrawal (Oklahoma City); Dr. Philip Bedient (Houston); Hector Guerrero (Austin); Stephen Parker (Dallas); and Bill Robison (Tulsa). Their willingness to participate in and contribute to the overall success of the project was critical.
REFERENCES & SUGGESTED READINGS