projects
Effects of Green Space on Health and Wellbeing
This National Science Foundation (NSF) funded project is led by Principal Investigator Dr. Jon Goodall and brings together researchers at UVA and Norfolk State University to design and implement a series of green infrastructure interventions in Norfolk and Portsmouth located within Virginia’s Hampton Roads region. As part of NSF’s Coastlines and People Hubs for Research and Broadening Participation (CoPe) program, this work aims to meaningfully integrate community participation to advance equity in coastal resilience and consists of six inter-related tasks:
- Task 1: Asset Mapping
- Task 2: Stormwater Modeling
- Task 3: Equitable Public Policies
- Task 4: Community Engagement
- Task 5: Co-Design of Green Stormwater Interventions
- Task 6: Assessment of Green Stormwater Co-Benefits
Working closely with Dr. Jenny Roe, my role on this project has entailed designing and implementing a household survey where residents of a neighborhood in Norfolk and a similarly situated neighborhood in Portsmouth are recruited to and compensated for responding to a series of questions regarding access to an usage of nearby public green space, perceptions of the safety and quality of nearby public green space, mental and physical health outcomes, as well as demographic characteristics. To support Task 6: Assessment of Green Stormwater Co-Benefits I will analyze the results of the survey using structural equation modeling to explore the relationship between factors like physical proximity, frequency of usage, and wellbeing indicators. The research team has selected these two neighborhoods because one is more conventional in terms of its stormwater management infrastructure while the other reflects green infrastructure approaches to flood mitigation.
I have also designed and am supervising a complementary data collection effort based on the methodology established by Chen et al. (2024) that documents usage patterns—including the degree and nature of social interaction—within public green space like parks in an unobtrusive manner that recalls the pioneering work of William Whyte. Like the Chen et al. study, member of our team will visit the selected public green spaces six times—two times (morning and afternoon) on three different days (a weekday, a Saturday, and a Sunday). The ESRI Survey123 tool and mobile app have been used to operationalize the Chen et al. (2024) methodology, which facilitates mapping and spatial analysis of the behavioral observations recorded during these visits. The results of this component of Task 6: Assessment of Green Stormwater Co-Benefits will complement the household survey and lay the foundation for a bookend analysis at the end of the project.
Integrating Scenario Analysis, Simulation, and Design Techniques to Inform the Scaling and Siting of Nature-based Solutions
This project has been funded for two years by the Environmental Institute and aims to develop and pilot a framework for the co-production of equity-centered resilience planning that brings rigor and local relevance to the selection, siting, and scaling of nature-based solutions (NbS). Through design techniques utilizing participatory mapping approaches, the project seeks to incorporate community input and capture new knowledge that is often missing in decision making, including acceptance by residents, and alignment with local government implementation capacity—in the selection of NbS in different contexts. Such an approach can help bridge the inclusion gap and may enhance uptake and implementation of NbS in the most vulnerable neighborhoods by building trust between residents and institutions like local government or universities. More broadly, the project will identify strategies for increasing the visibility and meaningful integration of equity dimensions into climate resilience initiatives that rely on NbS. Broader impacts also extend to innovative education and outreach activities that mentor students and foster collaboration between UVA and community partners.
The study area for this pilot consists of the Meadow Creek and Moore’s Creek watersheds, which span Charlottesville and Albemarle County in central Virginia. Identifying NbS relevant to flood and heat mitigation as well as air pollutant reduction, then engaging residents in a deliberative process that presents inherent tradeoffs in ecosystem services provided by candidate NbS interventions is the centerpiece of this work. The ecological modeling that quantifies expected benefits will be conducted using i-Tree Tools and led by Dr. Charity Nyelele.
Using participatory mapping techniques, participants are also asked to identify areas of concern related to stormwater management, urban heat, and air pollution in their areas as well as potential sites for future NbS. These insights will be used to develop scenarios that reflect plausible interventions and where they should be implemented within each study area to yield flood mitigation and co-benefits such as heat island and air pollutant reduction. By creatively integrating environmental justice into an integrative framework for ecosystem service assessment and valuation, results could provide an example roadmap for other cities to follow, potentially transforming NbS planning and management into a more just system that is both transferable and scalable. Finally, this project will identify strategies for increasing the visibility and meaningful integration of equity dimensions into climate resilience initiatives that rely on NbS.
Predicting Groundwater Contamination to Guide Testing and Mitigation
This ongoing project brings machine learning to bear on an important, yet understudied issue—contamination of drinking water in rural communities and areas that are not served by EPA-regulated public water supply systems. More than one million households in Virginia rely on wells, springs, and cisterns for their drinking water supply. While there is widespread perception that well water is safe, the quality of drinking water in Virginia homes that use private wells varies widely. Lead levels in homes supplied by private wells are influenced by the chemistry of the groundwater due to geology and the make-up of plumbing components. Lead is a pervasive environmental contaminant with well documented adverse health effects. Groundwater in the Piedmont—where Fluvanna, Albemarle, Nelson, and Buckingham counties (i.e., the study region) are located—is particularly corrosive, leading to higher incidence of lead contamination even in relatively new homes. Unlike homes served by regulated municipal water supplies, maintenance and monitoring of these systems is solely the homeowner’s responsibility. Given that the burden of testing and maintenance is privately held (i.e., the responsibility of residents), community members and researchers hypothesize that water contamination may disproportionately impact people of lower socioeconomic status—which correlates with racial minority status in the region studied. A series of pilot studies based on water quality testing of samples drawn from homes in the region has supported this expectation.
The central research question motivating this study is “how can predictive models be used to inform groundwater testing and remediation initiatives in rural areas where data are scarce” and potentially replicated within similarly situated rural areas outside Virginia. We apply a machine learning approach to model the relationship between selected water contaminants and a variety of predictors derived from structural characteristics, land use, soil, and geologic factors within the immediate vicinity at the scale of the individual housing unit. The trained models are then used to predict which untested housing units in the study region are most likely to have drinking water that exceeds regulatory thresholds for selected contaminants. Next, we use spatial statistics to identify “hotspots” where model predictions exhibit clustering of housing units with a higher probability of exceeding acceptable thresholds for drinking water contamination. Finally, we map the clusters of housing units with higher predicted probability of contamination alongside demographic and economic indicators in order to begin examining the equity implications of this issue. This exploratory research highlights a role for predictive models in targeting scarce testing and remediation resources to those housing units and resident populations that are at greatest risk of negative health impacts and builds upon an ongoing initiative led by Virginia Cooperative Extension.
