-
- ED41A-0525: A Spatiotemporal Machine Learning-Based Approach for Predicting Urban Heat Island Growth in the Research Triangle, North Carolina
-
Board 0525‚ Hall EFG (Poster Hall)NOLA CC
Author(s):Generic 'disconnected' Message
Jacob Aronow, NASA SEES Internship Program (First Author, Presenting Author)
Catherine Cardenas, NASA SEES Internship Program
Lorenzo Beltran, NASA SEES Internship Program
The urban heat island effect (UHIE) is a phenomenon in which cities and urban areas experience hotter temperatures when compared with surrounding rural areas. This is attributed to greater infrastructure including buildings and roads and fewer green spaces, causing these areas to absorb and keep more heat. In this study, the Research Triangle region (Raleigh-Cary-Durham) of North Carolina was analyzed to predict how these urban heat islands would evolve over time. The area was divided into smaller grid squares in which variables including land surface temperature, vegetation, impervious surfaces (such as concrete) and population between the years 2016 and 2024 were analyzed. A random forest, a type of machine learning model, was used to find patterns in the data and predict where the hottest areas would be. The model could also predict temperatures fairly accurately for the years 2022 and 2023. We used the model to estimate how these hotter areas could develop in 2025 and 2026. The most important factor contributing to higher temperatures was the amount of impervious (non-absorbent) surfaces, like roads and buildings. This approach could help other cities prepare for growing urban temperatures due to continued development and climate change.
Scientific DisciplineNeighborhoodType
Enter Note
Go to previous page in this tab
Session
