Examining weather impacts on urban travel demand in Nanjing, China during the COVID-19 pandemic

Accurately characterizing variation in public transport demand is important for urban traffic management, especially when weather conditions and public-health disruptions jointly reshape travel patterns. Using monthly transportation statistics and weather observations for Nanjing, China, from February 2019 to December 2022, this study examines associations between weather conditions and passenger flow across buses, taxis, and the Rail Transit System (RTS) during the COVID-19 period. Random Forest (RF) and Shapley Additive Explanations (SHAP)-based interpretation were used as exploratory tools to examine the relative associations of weather variables with passenger-flow variation. The results show that RTS remained the dominant travel mode in Nanjing, followed by buses and taxis, and that passenger flow for all three modes declined sharply at the onset of the pandemic before gradually recovering. During the study period, temperature generally increased, whereas other weather variables showed changes in varying directions. Although the importance rankings derived from the RF and SHAP methods are not entirely consistent, both identify temperature and precipitation as important weather factors influencing changes in traffic demand. In addition, the pandemic was associated with a marked shift in modal structure: before the pandemic, passenger-flow shares were relatively balanced across modes, whereas afterward taxi and RTS accounted for a larger share and bus ridership declined. These findings provide may offer a preliminary reference for adaptive transport planning.