Assessing the three-dimensional vegetation carbon sink of urban green spaces using unmanned aerial vehicles and machine learning
Wei Wei, Junqiao Li
Abstract
• Evaluated the urban green spaces’ carbon sink functions for carbon management. • Compared 2-D (GSA, GCR) vs. 3-D (3DGV, 3DOR) metrics for assessing green spaces. • Introduced UAV modeling, machine learning, and allometric equations for analysis. • Demonstrated multi-dimensional approaches improve ecological monitoring accuracy. As cities pursue decarbonization and carbon neutrality, urban green spaces play a crucial role as primary carbon sinks, warranting comprehensive quantitative assessments. This study compares traditional two-dimensional green space indicators, such as green space area and GCR, with advanced three-dimensional metrics, including 3DGV and 3DOR, as well as commonly used remote sensing indices like NDVI and NPP, for evaluating the carbon sink potential of urban green spaces. By integrating vegetation allometric growth equations, this paper introduces a novel methodology for assessing the carbon sink function of urban green spaces using UAV-based modeling and machine learning techniques for feature recognition. The results show that three-dimensional metrics provide a more accurate representation of the carbon sink capacity of urban green spaces, while traditional two-dimensional indicators fail to capture the spatial and functional variations effectively. This research contributes to the development of more robust ecological indicators for urban carbon management and highlights the role of innovative technologies, such as AI, in advancing environmental monitoring and management practices. The findings underscore the importance of multi-dimensional approaches in ecological assessment, demonstrating their potential to inform policy and management strategies for sustainable urban development.