An overview of forecast models evaluation for monitoring air quality management in the State of Texas, USA

Abstract

Purpose

The purpose of this study is to investigate forecast models using data provided by the Texas Commission on Environmental Quality (TCEQ) to monitor and develop forecast models for air quality management.

Design/methodology/approach

The models used in this research are the LDF (Fisher Linear Discriminant Function), QDF (Quadratic Discriminant Function), REGF (Regression Function), BPNN (Backprop Neural Network), and the RBFN (Radial Basis Function Network). The data used for model evaluations span a 12‐year period from 1990 to 2002. A control chart of the data is also examined for possible shifts in the distribution of ozone present in the Houston atmosphere during this time period.

Findings

Results of this research reveal variables that are significantly related to the ozone problem in the Houston area.

Practical implications

Models developed in this paper may assist air quality managers in modeling and forecasting ozone formations using meteorological variables.

Originality/value

This is the first study that has extensively compared the efficiency of LDF, QDF, REGF, BPNN and RBFN forecast models used for tracking air quality. Prior studies have evaluated Neural Networks, ARIMA and regression models.

Citation

Abdullah, A.B.M., Mitchell, D. and Pavur, R. (2009), "An overview of forecast models evaluation for monitoring air quality management in the State of Texas, USA", Management of Environmental Quality, Vol. 20 No. 1, pp. 73-81. https://doi.org/10.1108/14777830910922460

Download as .RIS