Background: Previous studies have demonstrated an association between short-term exposure to ambient temperature and mortality. However, the long-term effects of elevated temperature and temperature variability on mortality have remained somewhat elusive in epidemiological studies.
Methods: We conducted a comprehensive epidemiological study utilizing Chinese population census data from 2000 and 2010. Census-derived demographic and socioeconomic factors were paired with temperature data from the European Re-Analysis Land Dataset across 2823 counties. We employed a difference-in-difference approach to quantitatively examine the relationship between all-cause mortality and annual exposure to mean temperature and diurnal temperature range (DTR). Additionally, we evaluated the potential effects of socioeconomic and environmental covariate modifications on this relationship and calculated the attributable mortality. Lastly, we projected excess deaths attributable to annual temperature exposure under various shared socioeconomic pathways (SSPs, e.g., SSP126, SSP370, and SSP585).
Results: For each 1℃ rise in annual mean temperature and DTR, the mortality risk could increase by 6.12% (95% CI: 0.84%, 11.69%) and 7.72% (95% CI: 3.75%, 11.84%), respectively. Counties with high labor-force ratios and high NO2 and O3 concentrations appeared to be sensitive to the annual mean temperature and DTR. Climate warming from 2000 to 2010 may have resulted in 5.85 and 14.46 additional deaths per 10,000 people attributable to changes in annual mean temperature and DTR, respectively. The excess mortality related to changes in annual mean temperature and DTR is expected to increase in the future, with special attention warranted for long-term temperature changes in Southwest China. 
Conclusions: Our findings indicate that long-term mean temperature and DTR could significantly impact mortality rates. Given the spatial heterogeneity of increased mortality risk, the formulation of region-specific strategies to tackle climate change is crucial.