Statistical Distribution
Models for Airborne VOCs (Xylene, Toluene and Benzene) in Visakhapatnam using Burr
XII 3P, Log-Logistic 3P and Dagum-I 3P Distributions
Saripalli Arun Kumar, Akiri Sridhar, Sarode Rekha, Siripurapu Adilakshmi and Ramanaiah
M.
Res. J. Chem. Environ.; Vol. 29(3); 39-51;
doi: https://doi.org/10.25303/293rjce039051; (2025)
Abstract
This study investigates the application of advanced statistical models for airborne
Volatile organic compounds (VOCs)-specifically xylene, toluene and benzene-using
air pollution data collected from the city Visakhapatnam from the year 2018 to 2022.
To capture the non-normal distributional behavior of VOC concentrations, three flexible
probability distributions were employed: Burr XII 3P, Log-Logistic 3P and Dagum-I
3P. Parameter estimation was performed via maximum likelihood estimation (MLE) while
model validation was achieved through P–P and Q–Q plots. In order to identify the
most suitable distribution for modeling the air pollutant data, three distinct goodness
of fit test statistics and five model selection criteria were applied.
The results demonstrate that the Dagum-I 3P distribution best fits xylene and benzene,
while the Log-Logistic 3P model is optimal for the moderately skewed toluene concentrations.
Cross-validation confirms these findings, highlighting the reliability of tailored
distributional models for VOCs. The proposed distributions provide a robust framework
for predicting air quality information and conducting accuracy assessments with
all calculations and visual indications carried out with R-software. This work underscores
the potential of pollutant-specific modeling for improved air quality assessment
and management strategies, contributing to environmental health planning in urban-industrial
areas.
