CONDITIONAL LINEAR RANDOM PROCESSES BASED MODELLING OF THE NITROGEN DIOXIDE AIR POLLUTION TEMPORAL DYNAMICS

Abstract

Mathematical modeling and analysis of air pollution play a critical role in predicting pollutant dispersion, tracing emission sources, evaluating health and environmental impacts, and designing efficient monitoring and control strategies. Models grounded in theoretical investigations and data-driven analysis yield high-resolution spatial and temporal information, supplement costly monitoring networks, and enable “what-if” scenario testing to assess prospective policies, thereby supporting evidence-based decisions in air quality management.

Nitrogen dioxide (NO₂) is a major atmospheric pollutant primarily generated by combustion processes associated with energy production and transportation. Beyond its direct adverse effects, NO₂ is a key precursor in the formation of photochemical smog and acid deposition, amplifying risks to environmental integrity and human health.

Mathematical models, complemented by time series analysis, constitute essential tools for monitoring NO₂ concentrations. Both deterministic and stochastic approaches have been developed to represent NO₂ dynamics. In this study, we examine a mathematical model of NO₂ concentration along highways formulated as a conditional linear cyclostationary random process, explicitly accounting for cyclostationarity and the physical mechanisms underlying NO₂ emissions. The proposed framework draws inspiration from Gaussian plume models, which are widely employed to simulate pollutant dispersion from point sources, providing a simplified yet practical basis for air quality modeling. We develop a model of NO₂ highway pollution that characterizes emissions as a superposition of stochastically dependent plumes, where point sources arise in time according to a non-homogeneous Poisson process.