Analysis of the relative uncertainty in asphalt pavement rutting prediction in China based on CMIP6 models

Anthropogenic greenhouse gas emissions exacerbate global warming, threatening asphalt pavement performance due to its strong temperature sensitivity. Most existing studies adopt only one global climate model (GCM), ignoring inter-model uncertainties when evaluating pavement performance. This study uses six CMIP6 models to explore how inter-model temperature discrepancies affect asphalt rutting predictions across China under three SSP scenarios. Near-surface air temperature projections are converted into asphalt mixture equivalent temperatures and incorporated into a rutting prediction model. The study compares temperature and equivalent-temperature biases, as well as rutting results at national, spatial and city scales, and quantifies relative uncertainty via the generalized three-cornered hat (GTCH) method. Results show obvious inter-model differences in temperature projections, and absolute uncertainty increases after equivalent-temperature conversion. Under SSP5-8.5 (2051–2060), rutting predictions vary by as much as 4.1 mm. Model selection may affect whether predicted rutting exceeds the 15 mm expressway maintenance threshold in certain cities. CanESM5 exhibits the highest rutting relative uncertainty among the six models, while the other models show lower or intermediate levels with varying responses across scenarios. These findings highlight the importance of accounting for inter-model variability in climate-informed pavement performance assessment and maintenance planning.