Heuristic statistical models for carbon monoxide in El Paso, Texas.

Abstract

On the basis of existing data it was determined that while traffic is the major source of CO in the area, meteorological factors provide the dominant influence in the elevation of these concentrations to levels that exceed the National Ambient Air Quality Standards. Extensive comments regarding the results of the study and the implications and application of these results are also included.

The title of this study is "Heuristic Statistical Models for Carbon Monoxide in El Paso, Texas". The stated objectives of the study are (1) to produce such an heuristic statistical model(s) and (2) to examine the historical and current CO ambient air profile in El Paso, Texas, using the method of descriptive statistical analysis. A literature review and bibliography are given to provide greater understanding of the multiple facets of CO behavior and atmospheric modeling. Descriptions of the sampling sites, sampling, analytical methods, and the data bases used in this study are given. The methods section discusses in detail data collection and management practices and problems, particularly as related to data quantity and diversity of initial data bases. Also included are diagrams and a discussion for the sequential development, calibration, and verification of two heuristic statistical models for CO concentration predictions.

Descriptive statistical analyses of the data indicated strong, persistent seasonal and diurnal patterns for carbon monoxide, traffic, wind speed, wind direction, temperature, mixing height, and transport wind. Tables and figures are provided to demonstrate the types of conditioning used in the model development and calibration processes which were performed using 4 of the 6 years contained in the data base. Model verification results using 2 of the 6 years of data in a 20-term quadratic form model and a 5-term general linear model show both models to be effective in predicting ambient air CO concentrations.