Student Research Reports

View Research Report

Urban Heat Islands (UHIs) are areas in cities that experience significantly higher temperatures than their surroundings. This phenomenon is especially concerning in places like Austin, Texas, where socioeconomically disadvantaged communities are often the most affected. These low income neighborhoods typically have less greenery, which provides shade and cools the surrounding air, and more impervious surfaces like concrete and asphalt, which can trap heat. This leads to higher temperatures and increased health risks, like poor water quality, drought, disease, and overheating, particularly as climate change makes heat waves more common. Understanding what contributes to these higher temperatures is critical so that city planners can take steps to take away from the effects of UHIs and better protect people living in urban areas. To explore these questions, we analyzed twelve specific 3km by 3km areas, called areas of interest (AOIs) in Austin, Texas, collecting data on greenness and impervious surfaces from satellite imagery, income data from the U.S. Census, and temperature data from a NASA sensor called ECOSTRESS. We also ensured that the satellite data was accurate using ground photos taken by local volunteers in our AOIs. Using this satellite data and machine learning - a process where a coding model is fed large amounts of data and trained to produce a certain result - we found that lower-income areas often have more impervious surfaces and less green space, which increases the risk of UHIs. Our findings also showed a strong correlation between lower income and higher levels of impervious surfaces, leading to more severe heat in disadvantaged parts of the city. Using clustering algorithms - another type of code model that singles out patterns in data and separates them into groups based on these patterns - we confirmed that lower-income areas are more likely to experience UHI effects due to less vegetation and more heat-absorbing surfaces. In the future, we plan to expand our study to include areas in California in order to analyze a wider range of data and combine our methods with more advanced machine learning techniques to create even more accurate predictions on how UHIs affect communities. Our ultimate goal is to help info