High‐Resolution Mapping of Regional NMVOCs Using the Fast Space‐Time Light Gradient Boosting Machine (LightGBM)

Abstract

AbstractAccurate spatiotemporal estimation of non‐methane volatile organic compounds (NMVOCs) plays a pivotal role in establishing sophisticated early warning systems and formulating strategies to combat air pollution. Despite these critical applications, robust estimation of high spatiotemporal resolution NMVOCs concentrations remains a challenge. In this study, we develop a space‐time Light Gradient Boosting Machine (STLGB) model, which successfully renders hourly maps of NMVOCs concentrations across Shanghai from 2019 to 2022 by integrating spatiotemporal information. After extensive training, the STLGB model demonstrates remarkable estimation performance for NMVOCs, accounting for multiple spatiotemporal variables (R2 = 0.92, RMSE = 34.52 ppb). With the developed model, we provide first high‐resolution (1 km) hourly NMVOCs concentration maps, uncovering previously overlooked spatiotemporal variations. Further, SHapley Additive exPlanation (SHAP) regression values reveal significant local interpretation capabilities of the STLGB model, emphasizing the strong influence of emissions on NMVOCs estimation, whilst acknowledging the important contribution of space and time term. Our study of the pandemic lockdown further showcases the model's adaptability to unique events influenced by policy changes. The superior performance of the STLGB model, with its minimal computational memory requirements and faster speed, makes it an ideal tool for air pollutant estimation, adaptable to any region with NMVOCs monitoring capabilities.