Development of a recurrent spatiotemporal deep-learning method coupled with data fusion for correction of hourly ozone forecasts.

Ambient ozone (O) predictions can be very challenging mainly due to the highly nonlinear photochemistry among its precursors, and meteorological conditions and regional transport can further complicate the O formation processes. The emission-based chemical transport models (CTM) are broadly used to predict O formation, but they may deviate from observations due to input uncertainties such as emissions and meteorological data, in addition to the treatment of O nonlinear chemistry. In this study, an innovative recurrent spatiotemporal deep-learning (RSDL) method with model-monitor coupled convolutional recurrent neural networks (ConvRNN) has been developed to improve O predictions of CTM.

Source and sectoral contribution analysis of PM based on efficient response surface modeling technique over Pearl River Delta Region of China.

Identifying and quantifying source contributions of pollutant emissions are crucial for an effective control strategy to break through the bottleneck in reducing ambient PM levels over the Pearl River Delta (PRD) region of China. In this study, an innovative response surface modeling technique with differential method (RSM-DM) has been developed and applied to investigate the PM contributions from multiple regions, sectors, and pollutants over the PRD region in 2015. The new differential method, with the ability to reproduce the nonlinear response surface of PM to precursor emissions by dissecting the emission changes into a series of small intervals, has shown to overcome the issue of the traditional brute force method in overestimating the accumulative contribution of precursor emissions to PM.

Real-time source contribution analysis of ambient ozone using an enhanced meta-modeling approach over the Pearl River Delta Region of China.

The nonlinear response of O to nitrogen oxides (NO) and volatile organic compounds (VOC) is not conducive to accurately identify the various source contributions and O-NO-VOC relationships. An enhanced meta-modeling approach, polynomial functions based response surface modeling coupled with the sectoral linear fitting technique (pf-ERSM-SL), integrating a new differential method (DM), was proposed to break through the limitation. The pf-ERSM-SL with DM was applied for analysis of O formation regime and real-time source contributions in July and October 2015 over the Pearl River Delta Region (PRD) of Mainland China.