Sensitivity analysis of isoprene and aerosol emission in a suburban plantation using long short-term memory model

Abstract

Plants emit volatile organic compounds (VOCs), which undergo photochemical reactions to generate secondary organic aerosols and cause ozone (O3) accumulation; thus, they are the focus of urban pollution control strategies. Although some studies on biogenic VOCs and aerosols have been conducted, their environmental effects at the community scale are uncertain, particularly in suburban areas where abundant vegetation and pollutants coexist. Simple data processing methods cannot well address complex environmental conditions, whereas the long short-term memory (LSTM) model can learn relationships between research objects autonomously, thus enabling the analysis of influencing factors. Therefore, we collected 11 indicators of isoprene and submicron aerosol (PM1) concentrations, as well as other pollutants and meteorological factors, for six consecutive months in a suburban plantation in Shanghai. On the basis of correlation and redundancy analyses, LSTM was used to determine the main environmental factors influencing isoprene and PM1 concentrations. The results showed that nitrogen oxides (NOX), O3, photosynthetically active radiation (PAR), and temperature had the highest influence. The reactions between NOX, O3, and isoprene, PM1 are complex, and high levels of both isoprene and NOX in suburban areas may favor O3 production. PAR and temperature control the emission and photooxidation of isoprene and influence their conversion from a gaseous state to a particulate state. Considering the complex atmospheric environment, this study analyzed the sensitivity of isoprene and PM1 to environmental factors by artificial intelligence method. It demonstrated the effectiveness of LSTM for analyzing complex time-series data with unknown mapping relationships, which provided solutions for large-scale studies. And the results provided a scientific basis to further explain the mechanism through which influencing factors affect suburban communities and to control suburban air pollution.