Objective To assess the interactions between the daily mean air temperature, inhalable particulate matter (PM₁₀), and fine particulate matter (PM_2.5) on mortality of people in Harbin, Heilongjiang province, China.

Methods The daily mortality data, meteorological data, and atmospheric pollutants concentration data, including PM₁₀, PM_2.5, etc. in Harbin from 2014 to 2016 were collected. The distributed lag nonlinear model (DLNM) with quasi-Poisson link, bivariate response surface model, and stratified model were used to detect and measure the potential interactions between atmospheric particulate matter (PM₁₀ and PM_2.5) and air temperature on the daily mortality in Harbin.

Results The bivariate response surface model showed that there were significant interactions between PM₁₀, PM_2.5 and air temperature on mortality (P < 0.05). The stratified analysis showed that at high temperature, the longer the accumulative lag days, the greater the influence of PM₁₀ and PM_2.5 on the total daily mortality. When the accumulative lag was 12 d, the accumulative risks of the total daily mortality were increased to 10.4% (95% CI: 2.7%-18.8%) and 18.2% (95% CI: 1.8%-37.1%), respectively, with the increase of PM₁₀ and PM_2.5 by 10 μg/m³ . At low and moderate temperatures, no significant correlations between PM₁₀/PM_2.5 and total mortality were found. Compared with the moderate temperature layer, PM₁₀ and PM_2.5 had significantly greater influence on the mortality at high temperature (P < 0.05). A subgroup analysis showed that at high temperature, atmospheric particulate matter had a greater influence on the mortality of males and people over 65 years of age.

Conclusion There are interactions between PM₁₀, PM_2.5 and air temperature on mortality of people in Harbin. At high temperature, PM₁₀ and PM_2.5 have greater influence on the risk of mortality.