Enhancing Urban Ecological Risk Assessment by Integrating Spatial Modeling and Machine Learning for Resilient Environmental Management in UNESCO World Heritage Cities

Rapid urbanization in George Town, Malaysia, a UNESCO World Heritage city, has led to significant ecological degradation over the past three decades. This study enhances the Remote Sensing Ecological Index (RSEI) by integrating water resources as a new parameter, providing a comprehensive assessment of the city's ecological health from 1992 to 2022. Utilizing multi-temporal Landsat data, ecological assessment parameters such as land cover, soil moisture, surface temperature, and greenness patterns were analyzed. The integration of these parameters into the RSEI revealed correlations between forest cover and water body degradation, with a 54.90% and 46.94% reduction, respectively, leading to increased surface temperatures and negatively impacting soil moisture. The analysis shows that 37.64% of George Town experienced ecological degradation over three decades, with areas of excellent ecological health declining from 11.13 to 4.45%. A hybrid machine learning algorithm combining Cellular Automata and Artificial Neural Networks projected increased ecological vulnerability by 2032, with a further decrease in areas of good (12.20%) and excellent (0.25%) ecological health. Directional change analysis suggests that areas from the center to the eastern region experienced the highest levels of ecological degradation, a pattern projected to persist. The enhanced RSEI facilitates accurate ecological monitoring, guiding conservation efforts to maintain and restore ecological corridors and greenspaces within vulnerable ecosystems. This research provides an innovative, integrative methodology to support the global sustainable development agenda, advancing ecological change assessment in rapidly developing urban areas and informing urban planning for ecological resilience.