| Environmental pollution studies conducted to monitor ambient levels and to quantify the concentration of various pollutants entering a given environmental area are of great interest for possible adverse-health effects. Of particular importance in environmental data analysis is to select appropriate probability models. The previous studies indicated that none of the probability models, including the classical lognormal, has been identified to be superior to others in a general sense. To address this problem, the purpose of this study is twofold. Firstly, we introduce a generalized log-logistics distribution as a general model in fitting environmental pollutant data, and develop maximum-likelihood techniques for estimating parameters of the proposed distribution. The family of the generalized log-logistics distribution includes several well-known distributions In modeling data of environmental pollutant concentrations, such as lognormal, weibull, and gamma as special cases. Secondly, by applying the proposed model to seven data sets, we explore the possibilities of using this model as a general probability model for representing environmental-quality data.;The results of applications indicate that the generalized log-logistics distribution could be a good alternative to the classical lognormal distribution for fitting environmental quality data. For all of data sets, generally, the four-parameter GLL distribution fits better than the lognormal log-logistic, and three-parameter GLL distributions, or provides at least as good a fit as the other distributions. |
