Development of PLS–path model for understanding the role of precursors on ground level ozone concentration in Gulfport, Mississippi, USA

Abstract

Ground-level ozone (GLO) is produced by a complex chain of atmospheric chemical reactions that depend on precursor emissions from natural and anthropogenic sources. GLO concentration in a particular location is also governed by local weather and climatic factors. In this work an attempt was made to explore a Partial Least Squares Path Modeling (PLS–PM) approach to quantify the interrelationship between local conditions (weather parameters and primary air pollution) and GLO concentrations. PLS path modeling algorithm was introduced and applied to GLO concentration analyses at Gulfport, Mississippi, USA. In the present analysis, three latent variables were selected: PRC (photochemical reaction catalyst), MP (meteorological factor), and OPP (other primary air pollutants). The three latent variables included 14 indicators for the analysis; with PRC having two (extraterrestrial radiation on horizontal surface, and extraterrestrial radiation normal to the sun), MP having nine (temperature, dew point, relative humidity, pressure, visibility, maximum wind speed, average wind speed, precipitation, and wind direction) and OPP having three (NO2, PM2.5, and SO2) parameters. The resulting model revealed that PRC had significant direct impact on GLO concentration but very small overall effect. This is because PRC had significant indirect negative impact on GLO via MP. Thus, when both direct and indirect effects were taken into account, PRC emerged as having the weakest effect on GLO. The third variable (OPP) also had a positive impact on GLO concentration.