Shift of surface O3-NOx-VOC sensitivity with temperature in the Guangdong-Hong Kong-Macao Greater Bay Area, South China

This study examines the response of O3-NOx-VOC sensitivity to temperature under the windless (<3.3 m/s) and "no precipitation" daytime conditions from 2014 to 2019 in the Guangdong-Hong Kong-Macao Greater Bay Area. A non-parametric regressing DPO3-NOx (or NO2) curve-based method was validated that the ozone formation regime (OFR) spatial pattern was generally consistent with the previous satellite-derived HCHO/NO2 ratio (FNR)-based studies. Three "no precipitation" screenings (ISD- and ERA5-50% cloud cover, and ERA5-zero precipitation (<2 mm)) were included; and their diagnostic VOC-limited probabilities varied by about 23% in the Pearl River Delta (PRD) region, <8% in Hong Kong and >50% in Macao, respectively. The ERA5-zero precipitation-screening was selected for further investigation, as it provided much more data for analysis, and was more effective for diagnosis with generally lower deviations (<20% under 75% percentile) between the observation and CAQRA datasets. During the MDA8-daytime hours, the ozone pollution elevation coincided with higher temperatures. A total of 147 out of 305 CAQRA grids were diagnosed under the VOC-limited regime, with the highest probability along the cities from the eastern Pearl River Estuary (PRE) to the northwestern Greater Bay Area. The critical temperature points of 33 degrees C, 28-30 degrees C and 26-29 degrees C were recognized for the PRD region, Hong Kong and Macao, respectively, above which the majority of the study area was under the NOx-limited to transition regime (with the VOC-limited probability of 27% for PRD and 20% for Macao), except Hong Kong (63%). Whereas, it became much more VOC-limited (PRD: 60%, Hong Kong: 87%, Macao: 53%) below these critical points. Hence, the NOx mitigation should receive more attention at temperatures above those critical points, while the VOC emission control is likely more effective at the relatively lower temperatures.