| Ground-level Ozone (O3) and Nitrogen Dioxide (NO2) air pollutants adversely affect public health in many urban areas nationwide and some rural areas of Southeastern states. The objective of this research was the development of decision-making tools for air quality management. Regression equations were developed for predicting daily maximum 8-hr average O 3 and daily average NO2 concentrations using 1996--2000 data as functions of meteorological variables, roadway traffic data, emissions from mobile and point sources, and aircraft operations. The predictions of O3 for the hottest day in Tupelo and Hernando in Summer 2001 are reasonable and within 12% of the measured values. A new imagery-based surface classification (IMAGES) algorithm was developed using the spectral reflectance data of high-resolution multispectral imagery. Integer nonlinear optimization formulations were solved, which improved the overall accuracy of the surface class results by 6%. The IMAGES analysis of 1-m resolution satellite imagery taken in March 2000 showed 15.2% built-up area for Oxford in Mississippi. A methodology for estimating surface class areas based on the traditional topographic map and aerial photo provided comparable results with the IMAGES results for Oxford. The surface class areas for Tupelo and Hernando in Mississippi were used to calculate historical surface temperatures and enhance O 3 and NO2 regression models. An estimated 14.4% growth in built-up areas and traffic near the MS Highway 6-Jackson Avenue intersection in Oxford increased the surface temperature by 1.9°C, O3 by 0.006 ppm, and NO2 by 0.002 ppm in 2003. These models can be used to assess air quality degradation in rural areas considering growth in built-up areas, traffic volume, and congestion. |
