Investigation of nonlinear dynamics and stochastic characteristics of fine particulate matter in urban environments

Fine particulate matter (PM_2.5) is a global environmental issue and a serious threat to human health. Reducing PM_2.5 emissions is particularly crucial for China and India, which have the highest mortality rates associated with PM_2.5 pollution. Prediction and modeling as a vital tool for making accurate PM_2.5 concentration reduction policies require understanding the time series behavior of these pollutant. As the PM_2.5 concentration is influenced by multiple pollutant and meteorological parameters, it is needed to investigate that if the PM_2.5 has chaotic, periodic, or stochastic behavior. Hence, this paper investigates the dynamics of the daily time series of the PM_2.5 in Beijing and Delhi during 2014–2021 implementing the most common nonlinear methods of chaotic data analysis: false nearest neighbors, average mutual information, local divergence rates, correlation exponent, and the recurrence plot (RP). Based on the results, the correlation exponent does not saturate with increasing m, and the value for maximal Lyapunov exponent is negative. Hence, the PM_2.5 daily time series in Beijing and Delhi is non-chaotic. According to the RP analysis, the PM_2.5 daily time series in Beijing and Delhi has white noise and auto-regressive pattern, respectively. Understanding these data characteristics will be useful for designing accurate environmental policy and pollution reduction programs, as it helps to improve the estimation power and the robustness of the model.