Machine learning-based climate zoning and asphalt selection for pavement infrastructure under changing climate: A focused study of Ningxia, China
Feipeng Xiao, Zhitao Zhang, Zichao Wu, Wentao He, Li Jin
Abstract
Climate change poses significant challenges to the durability and performance of asphalt pavements. This study presents a comprehensive analysis of climatic factors in Ningxia, China, to establish a robust climate zoning framework for asphalt pavements. Utilizing machine learning techniques, specifically the fuzzy c -means (FCM) algorithm, three distinct climate zones within Ningxia were divided considering climatic features such as maximum temperature, minimum temperature, average temperature, maximum temperature difference, cumulative precipitation, and cumulative radiation. Based on the historical climate data and long-term pavement performance (LTPP) model, five asphalt performance grade (PG) zones were classified in Ningxia Province. Besides, six climate sub-zones, which integrated the asphalt PG zones into climate zones, provided a more refined strategy for the asphalt selection. The study also projected future climate scenarios using the NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP-CMIP6) dataset provided by the National Aeronautics and Space Administration (NASA) to assess the impact of climate change on asphalt selection in Ningxia. The significant changes in pavement temperature indicated the necessity to adapt asphalt pavement designs to future climate scenarios. Overall, this research contributed to the construction of more climate-resilient pavement infrastructures and provided an analysis framework for other regions facing similar climate-induced challenges.