IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Sensors, Data Aid Rural Areas in Flooding Resilience

While flood mitigation and resilience studies often focus on urban areas, researchers in Michigan are using sensors, machine learning and crowdsourcing to create disaster response tools for rural communities.

Metrolab December 2021.png
MetroLab Network has partnered with Government Technology to bring its readers a segment called the MetroLab Innovation of the Month series, which highlights impactful tech, data and innovation projects underway between cities and universities. In a special series, the Innovation of the Month is currently focusing on the award-winning and innovative projects championed by MetroLab’s member universities and civic partners that advanced to Stage 2 of the NSF Civic Innovation Challenge. If you’d like to learn more or contact the project leads, please contact MetroLab at for more information.
In this month’s installment of the Civic Stage 2 Innovation of the Month series, we highlight a project called “Helping Rural Counties to Enhance Flooding and Coastal Disaster Resilience and Adaptation” from Michigan’s Great Lakes Region. The project is applying remote sensing data resources and citizen engagement (crowdsourcing) to address current data gaps for improved flood hazard modeling and visualization that is transferable to rural communities in the Western Upper Peninsula of Michigan. MetroLab’s Elias Gbadamosi and Josh Schacht spoke with the team’s civic and academic partners about their implementation plan as they advance to Stage 2 of the challenge.

Elias Gbadamosi: Can you tell us about the goals of your project and what the motivation was?

Thomas Oommen: In the United States, flooding is a leading cause of natural disasters, with annual loss estimates of $54 billion. And while both urban and rural areas are vulnerable to flood hazards, most natural disaster resilience studies have focused on urban areas.

Melanie Kueber Watkins: The project’s goals are to develop methods that use remote sensing data resources and citizen engagement to address current data gaps for improved flood and coastal hazard and risk modeling and visualization that is transferable to rural communities, home to 60 million.

Oommen: A key motivation for this project was the 2018 Father’s Day Flood that hit the city of Houghton and the surrounding areas where Michigan Technological University is located and caused $100 million in damages.

Watkins: This event highlighted the need for cost-effective planning tools that rural communities can use for public awareness and to create mitigation strategies or pre-disaster plans to minimize the impact of future flooding or coastal hazards to increase resilience.

Josh Schacht: Who has been involved in this work, and what perspectives do they bring?

Ooommen: The project team includes Michigan Technological University and the University of Washington as the academic partners, as well as the Western Upper Peninsula Planning and Development Region and the Keweenaw Bay Indian Community Natural Resources Department. I bring expertise in the application of remote sensing for geohazard characterization. The other team members from Michigan Tech include Dr. Guy Meadows, who has over 40 years of research experience on coastal hazard-related issues. Dr. Tim Havens’ expertise is in machine learning and he will lead the crowdsourced data collection and analysis. Dr. Melanie Kueber Watkins is a hydraulics engineer and will lead the flood modeling tasks of the project. Ryan Williams will lead the development of the publicly accessible geospatial visualization tools.

Dr. Himanshu Grover is the University of Washington team member. Dr. Grover’s research focus is at the intersection of land use planning, community resilience and climate change. Dr. Grover is leading the social science research questions of the project.

Rachael Pressley is one of the community partners on the project team. Rachael recently updated the regional hazard mitigation plans in collaboration with local emergency management and Keweenaw Bay Indian Community. Her work focuses primarily on climate adaptation and resilience, rural food systems planning, and youth and elder engagement. Rachael will organize and facilitate engagement opportunities and integrate data validation tools into local classrooms, schools and community-centered workshops.
A map of the Great Lakes region.

Schacht: Can you talk about what data you are using and the approach you took to collate the data?

Guy Meadows: In our project we are concerned with both coastal flooding as well as inland flooding. With respect to coastal flooding, we are addressing both direct shoreline inundation and erosion and inland flooding through connections to the coastline through drowned river-mouth harbors subjected to coastal storm surge. Our approach is similar to that proposed by the Federal Emergency Management Agency (FEMA) in the Great Lakes Coastal Analysis and Mapping Study, which is not yet available for our rural setting in central Lake Superior. Our goal is to provide useful and actionable estimates of coastal flooding for hazard planning while the FEMA effort is being completed. Our region has already experienced a devastating flooding event during this interim period. To achieve this interim goal, we are relying on high temporal resolution water elevation data (six-minute intervals) from the National Oceanic and Atmospheric Administration (NOAA) Great Lakes water level gauge network. This data extends from 1980 to the present and captures not only mean water elevations along the shoreline but also surge during storm events. This analysis coupled with estimated wave runup calculations will provide a valuable planning and emergency response database.

Williams: A new generation of data-fused hazard maps will be created using higher-quality remote-sensing elevation data, supplemented with crowdsourced data. Flood models and 3D inundation maps will be generated by computational engineering processes using available 3DEP lidar and Lake Superior static water surface levels for watershed, then regional maps in Houghton and Baraga counties. Machine learning is trained via crowdsourced data to develop data-fused models and maps. A flood risk database will be created in FEMA’s Hazus program for this rural area and then adapted to other regions.

Gbadamosi: What are some of the interesting initial findings from your planning stage or the early execution stage of the project that you can share with us?

Himanshu Grover: The planning grant had three key objectives: identify key actors and build partnerships, build common ground, and co-explore needs. We hosted weekly planning and partnership-building meetings, a community listening event, and an online community needs survey to realize these steps. University and community partners attended these and focused on identifying key actors, building partnerships and building common ground. Weekly meetings helped develop a survey to understand the community needs and data gaps and plan for the community listening event. As per the survey results, most respondents (68 percent) were unaware that their residence was located inside a floodplain, and 80 percent had experienced some degree of flooding in the past five years. During the same period, most of the respondents (92 percent) had experienced some degree of waterlogging on the roads. Over the next two years, most of the residents (95 percent) expect to experience some degree of waterlogging on the road, and 95 percent expect their residences (including property grounds) to experience some degree of flooding. Over the next 10 years, 83 percent of the respondents felt that a flood would likely cause property damage. These results unequivocally highlight the public need for devoting resources to the identification, assessment and mitigation of flood hazards in this region. Among the respondents, 80 percent felt that at least moderate improvements in the availability of flood risk information were necessary for them and their neighbors.

Oommen: Half of the respondents (50 percent) felt that significant enhancements (defined as “more than a moderate amount”) were necessary to improve the availability of flood risk information. These survey results reflect the increasing realization of a flood knowledge deficiency among the residents, especially in increased flooding risk over the next 10 years. Another important outcome of the survey was the stated preferred mode of flood risk information dissemination by the local government. Most of the respondents (56 percent) expressed a preference for the use of Internet-based flood risk information dissemination, followed by a preference for radio (13 percent), newspapers (11 percent), public meetings (8 percent), brochures/printed materials (8 percent) and other informal means of communication (4 percent). The Internet was among the top three preferences for 88 percent of the respondents.

Gbadamosi: How are you measuring community resilience in this project, and how is your work improving those resilience metrics?

Grover: We adopt a broad perspective on community resilience from hazard research literature, wherein it is viewed as a proactive and positive expression of communities’ ability to mitigate, anticipate, cope with and recover from hazard events. Contemporary perspectives of community resilience focus on community as a complex socio-ecological system, and often integrate elements of different research strands. Consequently, we find that the measures of community resilience in the research literature reflect the diversity and complexity of underlying perspectives. However, it is also evident that most of these indicators reflect the availability of national or regional data with none reflecting community capacity to respond to a hazard event. An example would be whether communities with a lower number of hospitals per capita have developed surge capacity support by training community volunteers or investing in clinics. This is more relevant for rural counties, which are likely to have lower scores on the national level indicators but at the same time are likely to have developed stronger community capacity. Thus, any assessment of community resilience in such communities will require active participation of local agencies and residents in identifying the existing levels of community capacities that can offset the lower performance on national indicators. We will conduct an online public survey and conduct participatory workshops to co-develop appropriate indicators of community resilience with the community members. We expect to identify indicators of rural community resilience including an assessment of community capacity. The indicators will be measured using available local data to create an integrated Community Resilience Index. The indices and measurements will be shared with the community through the proposed user flood mapping interface. In addition to helping deliver community resilience assessment methodology tailored to the needs of our case study communities, this analysis can be used by agencies and other communities to support initiatives for all phases of emergency management, mitigation, response and recovery.

Schacht: Based on what you’ve learned from your citizen engagement exercises, what big priorities should other communities with high flood risk keep in mind?

Williams: The cities of Houghton and Hancock, located in Houghton County, line opposite sides of the Portage Lake Keweenaw Waterway. These communities are characterized by their steep topography, with many homes and buildings built directly into the hillsides. This unique setting can also lead to challenging and difficult-to-predict impacts, as experienced during the Father’s Day Flood of 2018. The incorporation of crowdsourced flood observations, improved infrastructure, building mapping, high-resolution elevation models, and machine learning techniques will contribute to a better understanding of the unique vulnerabilities of these communities and enable informed planning for strong resiliency.
Josh Schacht is the director of technology and strategy at MetroLab Network. He works to support MetroLab members and the civic research community as a whole in promoting evidence-based policy and local community engagement. Prior to his role at MetroLab, Josh was a solutions architect on the Master Data Management team at Katerra, working to leverage sustainable building materials to create efficient and affordable housing.

Elias Gbadamosi is civic research communications manager for Metrolab Network, responsible for the organization's communication, outreach and engagement programs. His work and interests converge at the intersection of civic communication, civic engagement and policy research.