Skip to main content
. 2021 May 4;291:112676. doi: 10.1016/j.jenvman.2021.112676

Table 1.

The type of functions to characterize the relationships between predictors and pollutant concentrations, and the definition of each factor including in GAM analysis. Categorial functions represent the variables are dummy vectors with values of 1 during the time interval otherwise the values are 0. Variables in brackets indicate the list of meteorological candidates. More details can be seen in the main text.

Predictors Functions Factors
Interannual Linear Time series of days since January 1, 2015.
Monthly Smooth January to June.
Weekly Category Weekdays, Weekends, Holidays in the heating season, Tomb Sweeping Day, Holidays in the non-heating season.
Lockdown Category Pre-lockdown, every month from February to June in 2020.
Diurnal Category Three factors (Pre-lockdown, lockdown, Post-lockdown) for each hour of the day.
Meteorology Smooth Boundary layer height, winds, (relative humidity, dew point temperature, air temperature, surface pressure and precipitation).