Geospatial and Machine Learning Driven Air Pollution Modelling Using Vehicular Load and Climatic Variables in Agbarho, Delta State, Nigeria

Abstract

Modeling air pollution concentration is a key aspect of environmental quality assessment and ecological resources management. In-situ measurements of three pollutants VOC, CO, and NO2 mainly determined by a combination of climatic variables and vehicle load were collected during field campaigns in three phases. Also, meteorological data (temperature, wind speed, and precipitation rate) was collected from an online weather forecasting platform. Results from the mean values computed and graphs plotted showed Five junction which has the highest concentration of vehicular movement (50 vehicles every two minutes) and human activities have the highest mean concentration of 980 ppm and a standard error of 40.60 ppm for VOC and 11.04 ppm for CO and a standard error value of 0.41 ppm, NO2 concentration of 0.35 ppm. Express junction which has the next highest vehicular concentration (47 vehicles every 2 minutes) has the next ranked mean concentration of 713.91 ppm and a standard error value of 59.56 ppm, CO value of 8.05 ppm, and NO2 value of 1.88 ppm. Ekwere road along was the next in pollution concentration ranking for VOC with 524.18 ppm and a standard error value of 20.64 ppm. it also had a high CO mean concentration of 9.20 and a standard error value of 0.69 ppm. which was higher than the express junction. Ughughelli road had a VOC mean concentration of 426 ppm, a standard error value of 31.53 ppm, a CO value of 4.14 ppm, and a standard error of 0.32 ppm. CDI and Agofure park which has similar vehicular loads had values of VOC at 323.05 ppm and 327.08 ppm respectively. They also had CO values of 4.03 ppm and 3.29 ppm. Ekwere less residential which has the lowest vehicular and human presence was the least polluted, with a VOC mean concentration of 193.42 ppm, CO mean concentration of 0.25 ppm, and NO2 concentration of 0.14 ppm.