Answers are grounded in this paper and cite their sources.
Statistical modeling analysis reveals distinct heavy-tailed patterns in coal plant nitrogen dioxide emissions, highlighting the need for segmented distributions to manage extreme pollution events.
This study uses three relatively heavy tailed statistical distributions commonly applied in air pollution, namely, the Weibull, Lognormal and Pareto distributions, to characterise the NO 2 emissions data (tons) from Lethabo coal fired power station, owned by Eskom, in South Africa. The study uses the Quantile–Quantile (QQ) and derivative plots as the main descriptive tools. The selected distributions cater for different tailed distributions relative to the Exponential distribution. Their tail heaviness is considered and ranked to arrive at the best fitting distribution for both the main body and the tail of the data distribution. The extreme value index (EVI) estimates (the generalised Hill and moment) and the generalised QQ plot are used to determine the tail heaviness of the data and to justify the selection of the distributions in this study. The adopted approach allows for the fitting of more than one distribution to different segments of the same dataset. The Lognormal distribution given a good fit for the bulk of the data and the overall upper tail of the NO 2 emissions data. However, the top six largest emissions observations are heavier tailed and better described by the Pareto-type tail as evidenced by the derivative plot and the EVI. Practically, the Pareto distribution models heavier tails of the NO 2 emissions data better than the Weibull and Lognormal distribution. The modelling methodology adopted herein can aid in quantifying and explaining NO 2 emissions more accurately along the spectrum of the emission. Thus, the Lognormal distribution model can be utilised to conservatively quantify and guide rigorous policy and management solutions for the control and reduction of NO 2 emissions without underestimating the very extreme emissions of primary concern.
No takes yet. Share an insight, caveat, or question.
Mamba et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: