The Spatial-temporal Patterns of COVID-19 Lockdown Effects and Diurnal Variations of Air Quality in Beijing, China due to Reduced Human Activities

Abstract:

Unprecedented travel restrictions and non-essential factories closures due to the COVID-19 pandemic caused a remarkable reduction of anthropogenic emissions and improvements in air quality. Estimates of the changes in pollutant concentrations during the lockdown periods are likely to be biased if the influence of meteorology and seasonal signals are not considered. The Generalized Additive Models (GAM) has been demonstrated the ability to identify the contrasting holiday effects on PM_2.5 and NO₂ in the Beijing area. In this study, a GAM model is developed that includes multi-temporal cycles and the non-linear influence of meteorological observations in order to quantify the lockdown effects at 34 sites in the Beijing area. The model is further used to determine diurnal variations during different lockdown stages.

The model showed that there are clear geographical patterns due to different local land-use types. The downtown area shows the largest reduction and most sites did not return to normal until June. The suburban area between the Fifth Ring Road and the Sixth Ring Road experienced a relatively smaller impact than downtown and was mostly back to normal by April. The area close to factories recovered quickly in March. The diurnal pattern from the model shows that the peak of people’s travel for necessities occurred from 1 to 5pm, and this was most pronounced in the suburban area during the strictest control periods. This study provides insights for quantifying the changes in air quality due to modified human activities while accounting for meteorological variability. The model improves the estimates of the effects on air quality of COVID-19 on small spatial scales and short temporal scales.