Date of Award


Document Type


Degree Name

Master of Science (MS)


College of Science and Mathematics


Mathematical Sciences

Thesis Sponsor/Dissertation Chair/Project Chair

Andrew McDougall

Committee Member

Thomas Devlin

Committee Member

Kirk Barrett


This research applies time series methods to determine relationships among a set of weather variables which are continually monitored in the Hackensack Meadowlands region of northern New Jersey. Weather data includes chemical and atmospheric factors. Chemical factors are Nitrogen Oxide, atmospheric Ozone, Carbon Monoxide, and Carbon Dioxide. Weather factors are wind speed, barometric pressure, air temperature, humidity, and solar radiation. Additionally, traffic density and time of week are brought in as categorical factors. This research attempts to (a) introduce the reader to various time series methodologies, (b) find a significant and efficient model for forecasting Nitrogen Oxide levels, and (c) answer several research questions related to relationships between the above factors.

We have found significant effects of rush hour and weekend on atmospheric NOx and Ozone levels. Specifically, we have shown that NOx levels are higher during the morning rush hour than during any other time on weekdays. As for weekends, overall NOx levels are lower than on weekdays, but there is no difference between morning rush and non-rush hour NOx levels. The effects of rush hour and time of week on Ozone are the opposite. Where NOx was found to be significantly highest during the weekday AM rush hour, Ozone was found to be at its lowest at this time. Weekends in general recorded significantly higher levels of Ozone than the weekday readings.

We have shown that a NOx prediction model is adequately modeled by statespace procedures based on atmospheric factors, but chemical factors can greatly enhance estimation capabilities. We have shown several relationships between chemical factors, as well as relationships between atmospheric factors. State space procedures have allowed us to produce past-value autoregressive descriptions for all factors. With this method, we have uncovered the close relationship between relative humidity and the Nox. Also, we have shown evidence to conclude that Meadowlands NOx levels are strongly influenced by the Carbon Monoxide emitted from automobiles, since Carbon Monoxide is one of only two factors with a positive relationship to NOx levels. Finally, with all methods, we have explained the reasoning and usefulness for log transformations on the chemical data.

File Format