Machine Learning-Based Urban Nighttime Light Intensity Analysis in Sri Lanka Using VIIRS Images
摘要
Urban expansion drives a sharp increase in electricity consumption, particularly for household lighting. Using Visible Infrared Imaging Radiometer Suite (VIIRS) Nighttime Light (NTL) data from 2015 to 2024, this study analyzes trends in urban energy use influenced by economic growth, energy policies, environmental factors, and global events like climate change and the COVID-19 pandemic. Radiance levels from NTL data serve as a proxy for electricity consumption, highlighting patterns of urbanization and socioeconomic development. The results show that major urban areas in Sri Lanka experienced an upward trend in NTL intensity, with interruptions caused by events such as the 2017 power disruptions, the 2020 pandemic-induced slowdown, and the severe economic crisis in 2022. Monthly trends reflect seasonal variability, while anomalies, like a − 78.9% drop in July 2020, underscore the impact of transient events and data inconsistencies. Population grid cell data from 2015 to 2020 correlates moderately with NTL intensity, highlighting the link between urban growth and energy use. Urban population share rose steadily, reaching 19.21% in 2023, mirroring increasing urban energy demands. The study emphasizes the utility of high-resolution NTL data for monitoring urban dynamics but notes limitations in rural and low-light areas, requiring complementary datasets. Insights from this research support urban planning and energy policy by advocating for enhanced satellite data reliability, investments in renewable energy, and strategies to improve resilience against external shocks. Policymakers can leverage integrated geospatial and socioeconomic data to foster equitable and energy-efficient urban growth as Sri Lanka continues to urbanize.