Workshop Overview
The University of Washington is hosting a 5-day collaborative workshop from September 9-13, 2024, to advance research products and methods for improving observations, assessments, and forecasts across appropriate temporal and spatial scales to accomplish three goals:
Investigate the human behavior and societal adaptive responses to, and impacts of, severe weather and climate-related events, particularly flooding associated with atmospheric rivers, hurricanes, and severe storms, but also including other extreme events such as heat or fire.
Address the research gaps linking mitigation to adaptation and resilience in relation to severe weather. This will involve exploring co-benefits for human well-being from climate adaptation strategies that will further contribute to resilience to extreme weather events and climate mitigation.
Explore pathways to better understand the dynamics of decisions and population disparities in responses to and impacts of past extreme climate / weather events.
The workshop will create improved data products and methods (data integration, data assimilation, analytic tools, new approaches to analyses) that integrate social and weather and/or climate data across space and time, through interdisciplinary collaborations. Data products should inform decision models that can guide decision making to address the needs of individuals, households, neighborhoods, and communities, with projections of impacts on the scales of minutes to hours, days, weeks, or years. Data and their analyses should be capable of informing impact analysis and risk reduction planning.
The workshop will focus on the data and analytic challenges of linking climate- and weather-related impacts and mitigation efforts to human behavior, health, and well-being by:
- facilitating all participants’ access to data sources and tools in a secure computing environment, when required, and provide open source tools and resources otherwise;
- exploring and innovating new approaches, including the use of machine learning (ML), to assimilate, curate, and analyze the relevant multiplicity of data required to assess and predict impacts of extreme weather events across contexts and scales.
- incorporating vital, mediating societal, institutional, organizational, and relational factors into understandings of the environment. The workshop organizers assume that societal and natural landscape factors are crucial mechanisms in managing mitigation and adaptation, but until now these factors have been incompletely incorporated into models that link climate change or weather to human and societal well-being;
- catalyzing research that more completely models weather- and climate-human linkages and accelerates knowledge and provision of tools for policymakers, emergency managers, businesses, and the public.
Prior to the workshop, participants will be expected to participate in activities related to preparing for the workshop, which may include a short presentation before or during the workshop on the participant’s workshop-related expertise, or topical research or methodological focus. Participants may also be asked to contribute to the post-workshop development of products, such as research papers, grant proposals, websites, and shared integrated tools and data.
Teams
D4 Hack Week participants collaborate in small group project teams, usually 3-6 people, to integrate specific social and weather datasets in order to address a research question of their choice.
Learn more about teams’ datasets, expected challenges, and member biosketches here.
Team members: Joan Casey, Lauren Wilner, Vivian Do, Heather McBrien, and David Coomes
Institutions: University of Washington, Columbia University
This project plans to integrate data on wildfires and wildfire boundaries, FEMA household claims, and demographics, with CalEnviroScreen data to ask several questions about FEMA assistance for wildfire disasters: What are the population characteristics of those actively applying for FEMA assistance for wildfire disasters? Within this group, what are the individual (e.g., owner vs. rental status, level of disaster preparedness, reliance on electrical medical equipment) and area-level (e.g., neighborhood poverty) factors associated with successfully receiving FEMA aid or the amount of FEMA aid? Do the factors associated with successfully receiving FEMA aid vary by disaster type (i.e., wildfire disaster vs. wildfire disaster co-occurring with extreme heat)?
Team members: Elizabeth Fussell, Jack DeWaard, Katherine Curtis, James Done, and Sara Ronnkvist
Institutions: Brown University, Population Council, University of Wisconsin – Madison, NSF National Center for Atmospheric Research
This project plans to integrate IRS county-level migration data with data on tropical cyclones, wildfire events, flood events, wet bulb temperatures and air pollution, county health rankings and roadmaps data, and national neighborhood data archive data, to ask the broad question: How do tropical cyclones affect county-level migration systems, where migration systems are the counties connected through in- and out-migration flows, and do these dynamics differ for age, sex, race/ethnicity, and nativity groups?
Team Members: Fabien Cottier, Mona Hemmati, Andrew Kruczkiewicz, and Kytt MacManus
Institution: Columbia University
This project is integrating data on tropical cyclones, migration, and vulnerability to examine to what extent flooding and hurricanes contribute to shape migration flows in the US. The project also asks how early warning systems can be appropriately leveraged to inform risk reduction with a view towards decreasing disproportionate impact from different types of floods / hurricanes.
Team members: Deborah Balk, Dylan Connor, Melanie Gall, Lori Hunter, Jenna Tipaldo, and Helen Wilson Burns
Institutions: City University of New York, Arizona State University, University of Colorado Boulder
This team plans to integrate Spatial Hazard Events and Losses Database (SHELDUS) data with American Community Survey (ACS) microdata to examine how the burden of hazards on the U.S. population has changed over the past 15-years and what implications this has for risk mitigation and emergency preparedness efforts. SHELDUS data include property damage (in dollars) and fatality counts from natural hazard events.
Team members: ChangHoon (Chang) Hahn, Sharif Islam, and Lidia Cano
Institutions: Princeton University, Massachusetts Institute of Technology
To address the broad questions of “How will existing flood management tools deal with future climate scenarios?” and “Which communities will be the most affected/most protected as a result?”, this team is planning to integrate Federal Emergency Management Agency (FEMA) data from the National Flood Insurance Program, American Community Survey data, data from First Street, Meta data on perceptions of future events/risks, and possibly international data (EM-Dat and the Extreme Weather dataset).
Team members: Mathew Hauer, Alexis Santos, and Sunshine Jacobs
Institution: Florida State University and Pennsylvania State University
This project plans to integrate ACS/Census Data and IRS migration data with novel demographic change modeling and data on environmental hazard events to ask: How do populations change in association with environmental events? How do different race/ethnic groups change in association with environmental events? And what are the long-range impacts of environmental change on demographic change?
Team members: Ethan Sharygin, Justin Stoler, and Mary Angelica Painter
Institutions: Portland State University, University of Miami, University of Colorado Boulder
This team is considering addressing the following questions:
1) How are people displaced after a wildfire and what are their characteristics? How are water insecurity and distrust in water utilities and services related and shaped by hazard experiences? What is the relationship between government capacity, community resiliency, and weather events? Data under consideration for integration to address these questions include 2024 nationally representative survey data with modules on water insecurity and institutional trust, migration destinations post-wildfire, rural capacity index, local hazard mitigation plan status, and community resiliency estimates.
2) Estimate population displacement effects from sudden onset disaster and compare network of migration destinations after sudden onset disaster to regular migration pathways, commute patterns, and parametric migration models.
Team members: Andrea Schumacher, Julie Demuth, DJ Gagne, Jorge Celis, Amy McGovern, Sara Curran, Sameer Shah, Masha Vernik, Ann Bostrom
Institutions: NSF National Center for Atmospheric Research, University of Oklahoma, University of Washington
Collaborating Across Threads Demonstration Project: Integrating multidisciplinary data to investigate changes in driving behaviors before and during a southern California atmospheric river flooding event
The motivation for this project is to investigate whether people change their driving behaviors as an atmospheric river (AR) flooding event is threatening and occurring and, if so, 1) who changes their driving behaviors, 2) how driving behaviors change, and 3) when those changes occur. This case study focuses on an AR event that occurred in Southern California from 29-31 March 2024, and integrates data from multiple disciplines such as longitudinal panel survey data, American Community Survey (ACS), weather, and mobility data to examine changes in driving behavior surrounding this event through various disciplinary lenses. This project also seeks to address questions related to the process of integrating data from different disciplines, including: How do different datasets, individually and together, answer these questions? Where are there similarities and differences? How can we have a focused examination on population disparities? What types of data are needed to meaningfully explore these groups? What can we learn in doing so?
Team members: Jason Stock, Tyler Fricker, Patrick Greiner
Institutions: Colorado State University, University of Louisiana Monroe, Vanderbilt University
Floaters are scientists who bring their expertise to tackle data integration challenges, model development, AI, machine learning solutions, and more. They will hold office hours and dedicate time to answering team questions.
Jason Stock is a Computer Science PhD candidate at Colorado State University (CSU) advised by Professor Chuck Anderson. His current research interests are in neuro-inspired attention methods, generative diffusion models, creating interpretable-by-design machine learning algorithms, and modeling weather and climate change.
Tyler Fricker is an Assistant Professor of Geography in the School of Sciences at the University of Louisiana Monroe. His research interests are, broadly, in the environmental impacts of natural hazards on society and the connection between weather and climate. Much of his work bridges the gap between climate change science and climate-society interaction.
Patrick Greiner is an environmental sociologist at Vanderbilt University. His research and teaching address questions at the intersection of structural inequality, development processes, and environmental change.
Workshop Support
This workshop is supported by a partnership between the University of Washington (UW)’s Center for Studies in Demography and Ecology (CSDE), the National Science Foundation AI Institute for Research on Trustworthy AI in Weather, Climate, and Coastal Oceanography (AI2ES), and the University of Washington eScience Institute.
Funding for the workshop derives from a grant from NOAA to AI2ES (Award NA23OAR40505031) and a center grant to CSDE from the Eunice Kennedy Shriver National Institutes of Child Health and Human Development via the P2C HD042828 mechanism.
In his role as a John C. Garcia Term Professor, Dr. Shah recognizes the generous financial support made possible by Carole Garcia.
Support for the workshop organization is also provided through the Weyerhaeuser Endowed Professorship in Environmental Policy at the University of Washington Evans School of Public Policy & Governance.
