Public Climate Sentiment and Urban Air Quality: Evidence From China

Public climate sentiment reflects societal perceptions and emotional responses to climate-related risks, yet its relationship with objective environmental outcomes remains insufficiently understood. This study aims to examine whether public climate sentiment is systematically associated with urban air quality. Using a large-scale dataset of 256,400 geo-tagged Weibo posts from 2010 to 2023, this study constructs a city-level public climate sentiment index for 287 Chinese cities through supervised text classification and machine-learning-based continuous sentiment scoring. The resulting index is integrated with city-year data on PM _2.5 concentrations and analyzed using two-way fixed-effects models, robustness checks, instrumental variable estimations, heterogeneity analyses, and channel analyses. The results show that more positive public climate sentiment is consistently associated with lower PM _2.5 concentrations. This association is more evident in non-resource-based cities, southern cities, provincial capitals, and environmental-protection priority areas. Further analyses show that public climate sentiment is systematically associated with public transport use, green patent activity, and environmental investment, suggesting that it is closely connected to broader behavioral, technological, and governmental responses. This study contributes by developing a scalable AI-enabled measure of public climate sentiment, linking digital social signals to objective city-level environmental outcomes, and demonstrating how social sensing can complement conventional environmental indicators in urban sustainability governance.