Machine learning based quantification of VOC contribution in surface ozone prediction.

The prediction of surface ozone is essential attributing to its impact on human and environmental health. Volatile organic compounds (VOCs) are crucial in driving ozone concentration; particularly in urban areas where VOC limited regimes are prominent. The limited measurements of VOCs, however, hinder assessing the VOC-ozone relationship. This work applies machine learning (ML) algorithms for temporal forecasting of surface ozone over a metropolitan city in India. The availability of continuous VOCs measurement data along with meteorology and other pollutants during 2014-2016 makes it possible to deduce the influence of various input parameters on surface ozone prediction. After evaluating the best ML model for ozone prediction, simulations were carried out using varied input combinations. The combination with isoprene, meteorology, NOx, and CO (Isop + MNC) was the best with RMSE 4.41 ppbv and MAPE 6.77%. A season-wise comparison of simulations having all data, only meteorological data and Isop + MNC as input showed that Isop + MNC simulation gives the best results during the summer season (RMSE: 5.86 ppbv, MAPE: 7.05%). This shows the increased ability of the model to capture ozone peaks (high ozone during summer) relatively better when isoprene data is used. The overall results highlight that using all available data doesn't necessarily give best prediction results; also critical thinking is essential when evaluating the model results.

Tracer-based characterization of source variations of ambient isoprene mixing ratios in a hillocky megacity, India, influenced by the local meteorology.

The ambient biogenic volatile organic compounds (BVOCs), mainly isoprene, are potentially involved in the formation of secondary pollutants, hence, they are significant in terms of air quality and climate. Although the largest sources of BVOCs are tropical regions, the measurements of isoprene in the Indian subcontinent are limited. We conducted the measurements of isoprene, benzene, and toluene at an urban site in a hillocky megacity of India using a high-sensitivity proton transfer reaction quadrupole mass spectrometer (PTR-QMS). The mixing ratios of isoprene were compared with those of aromatic compounds like benzene and toluene, which represent typical anthropogenic VOCs. Isoprene and isoprene/benzene (>5 ppbv ppbv-1) showed higher levels in the pre-monsoon months, most likely due to large emissions by urban vegetation during physiological activities in plants which was enhanced by the high ambient temperatures and solar radiation. While Benzene and toluene showed higher mixing ratios during winter, which were due to shallower boundary layer depths and transport of air masses from polluted Indo-Gangetic Plain during this season. The mixing ratios of VOCs show significant diurnal variation as a result of their different origins and the role of different meteorological parameters. The robust emission ratios of isoprene/benzene obtained from nighttime data were used to separate the non-anthropogenic and anthropogenic isoprene emissions. ∼30% enhancement observed in non-anthropogenic emissions to isoprene from winter to pre-monsoon season when temperatures and solar radiation were stronger, although traffic in the city. Isoprene/benzene ratio at lower temperatures (<25 °C) and solar radiation (<100 W m-2) was predominantly controlled by anthropogenic sources. Overall, toluene and isoprene are the most frequent species in terms of having the highest ozone-forming potential (OFP) values but biogenic isoprene became more important to ozone formation during the afternoon hours in the pre-monsoon months with high air temperatures (>25 °C).