Remote sensing of urban tree carbon stocks: A methodological review

Detailed and accurate assessments of urban tree carbon stocks (UTCS), achieved through the combination of field surveys and remote sensing techniques, are crucial for understanding the ecosystem services provided by urban trees, informing the terrestrial carbon cycle, and supporting sustainable urban planning. Since the 2010 s, breakthroughs in Earth observation and artificial intelligence have revolutionized UTCS remote sensing, transforming it into a dynamic field characterized by diverse methodological approaches. However, the methodological diversity complicates the selection of optimal approaches, and a standardized solution remains elusive. Furthermore, existing UTCS remote sensing methodologies face limitations in accuracy and scalability, particularly for large-scale, multi-city assessments. In response to these challenges, this study reviews UTCS remote sensing from a methodological perspective. Building on advances in remote sensing of urban trees, we integrate recent UTCS remote sensing methodologies into a unified framework and classify them into three categories: (1) land use stratification methods, representing early-stage strategies that estimate carbon stocks using field-sampled carbon densities assigned to different urban strata; (2) area-based inversion methods, which predict UTCS at the area level (e.g., plot-based units) by statistically modeling the relationship between field-measured carbon stocks and remote sensing metrics; and (3) individual tree detection methods, which quantify UTCS at the tree level using high-resolution remote sensing data. Within this framework, we present an application-oriented overview covering methodological definitions, underlying principles, modeling processes, data foundations, advantages and limitations, performance differences, and usage across the reviewed literature. We also summarize critical advances in supporting components, including allometric equations, classification and regression techniques, uncertainty analysis, and accuracy assessment. Through an in-depth discussion, we aim to enhance understanding of UTCS remote sensing and further propose a two-step framework to guide method and data selection for scientific research and management applications. Overall, this review offers timely insights into the integration of field surveys and remote sensing for UTCS assessment over more than a decade. The proposed tripartite methodological classification offers a novel perspective and a practical framework for advancing UTCS remote sensing research and practice.