This pilot project proposes to supplement ongoing high-frequency data collection in Monrovia, Liberia with remotely sensed flooding measurements. Since June, we have been conducting monthly surveys in six poor communities to learn about households’ flooding exposure, displacement, health, and adaptive behaviors. The final endline survey will be completed in October 2025. This pilot will support the acquisition and analysis of satellite imagery to enhance exposure measurement, validate self-reported data, and evaluate which types of remote sensing products offer the greatest added value for future population research.
The proposed remote sensing strategy follows a stacked design, layering multiple types of satellite data to assess their ability to measure flooding at the sub-community level. Each image will be aligned with survey data collected during the same time period, allowing for direct comparison across sources. The pilot includes acquisition of a newly tasked radar image, ideally timed with the October endline survey, to evaluate the feasibility and utility of on-demand imagery acquisition.
This pilot project proposes to supplement ongoing high-frequency data collection in Monrovia, Liberia with remotely sensed flooding measurements. Since June, we have been conducting monthly surveys in six poor communities to learn about households’ flooding exposure, displacement, health, and adaptive behaviors. The final endline survey will be completed in October 2025. This pilot will support the acquisition and analysis of satellite imagery to enhance exposure measurement, validate self-reported data, and evaluate which types of remote sensing products offer the greatest added value for future population research.
The proposed remote sensing strategy follows a stacked design, layering multiple types of satellite data to assess their ability to measure flooding at the sub-community level. Each image will be aligned with survey data collected during the same time period, allowing for direct comparison across sources. The pilot includes acquisition of a newly tasked radar image, ideally timed with the October endline survey, to evaluate the feasibility and utility of on-demand imagery acquisition.
Aim 1: Integrate and compare multiple satellite imagery products to detect community-level flooding. Using both public and commercial data, I will evaluate each image's ability to detect flooding across the six survey communities. This validation is essential to understand which types of satellite data provide reliable measures of environmental stress relevant to population responses. Comparison metrics will include spatial resolution, cloud/weather impacts, image quality, and cost. Findings will inform satellite data selection strategies and cost-effective study design for future projects.
Aim 2: Use linked satellite-survey data to analyze household responses to flooding. I will combine satellite-derived flood exposure with our survey data to examine how flooding affects reported displacement, health, service access, migration aspirations, and coping strategies. In addition to validating self-reported exposure, I will assess whether satellite-observed flooding explains variation in household impacts and captures aspects of exposure missed by survey data alone.
Aim 3: Identify health and migration outcomes most responsive to satellite-derived flood exposure to inform K01 development. Preliminary analyses from waves 1-3 show that an additional day of flooding in the past week is negatively associated with self-reported physical health and subjective well-being and positively associated with stress levels. I will use the health measures in our survey data to determine which outcomes show the strongest associations with satellite-observed flood exposure. A focused literature review on remote sensing applications in health research (including e.g. Grace and Davenport 2021; Gray and Thiede 2024; Pinchoff et al. 2020) will help guide this analysis and support the development of my planned K01 application.
This project will result in linked satellite-survey datasets, validated flood exposure measures, and a structured comparison of remote sensing data sources. By evaluating the reliability, interpretability, and cost-effectiveness of each data layer, the findings will generate preliminary data essential for future career development applications and inform optimal methods for studying the intersection of weather shocks, migration, and population health.