From the year 2000 to the year 2010, the total population in the United States increased by 12.1%. About ~80% of the total population resides in the urban areas. The growth in the urban population influenced the urban sprawl, congestion, and, subsequently delays on the existing road infrastructure. Urban sprawl is directly linked to the land use developments and has a significant influence on the operational performance of the neighboring links, leading to congestion and delay. Further, traffic condition, day-of-the-week, time-of-the-day, and network characteristics of the upstream, downstream, cross streets, and intersecting links also influence the operational performance of the link. Therefore, one needs to consider spatial dependency and the influence on links within the proximity (based on the distance decay effect), over time, to compute travel time variability or reliability. The goal of this dissertation is to model the influence of developments on travel time variations to improve mobility of people and goods. The objectives of the dissertation research are: 1. To identify the predictor variables which could influence the operational performance of link in terms of travel time and travel time variations,2. To identify to what extent the influence of proximal land use developments persist on travel times, 3. To compare before and after travel times and travel time variations on neighboring links of new developments, and, 4. To develop the relationship between land use developments on travel times and travel time variations on neighboring links by land use type, area type [Central Business District (CBD), CBD fringe, and urban area], and by speed limit categories (speed limit < 45 mph, 45 – 50 mph, > 50 mph).Data for 259 road links were selected within the city of Charlotte, North Carolina (NC). The land use developments and network characteristics were collected from the local agencies, while real-world travel time data were collected from the private agency. Three years of data, from the year 2013 to the year 2015, were considered in this research. Thirty-five different types of land use developments were considered in this research. The spatial dependency was incorporated by considering the land use developments within 0.5 miles, 1-mile, 2 miles, and 3 miles of the selected link. Network characteristics of the upstream, downstream, upstream and downstream cross street, and intersecting links were also considered to address the spatial dependency.Pearson correlation coefficients were computed by considering before-and-after data to investigate the relationship between land use developments and travel time measures. Forty-eight models were developed in this research. Of these, twelve models were developed by considering different buffer widths, eighteen models were developed by classifying the links by area type [Central Business District (CBD), CBD Fringe / Other Business District (OBD), and urban area], and eighteen models were developed by classifying the links based on the speed limit (< 45mph, 45 to 50 mph, and > 50mph). Each of the developed models were validated using the Root Mean Square Error (RMSE), the Mean Absolute Percentage Error (MAPE), and the Mean Percentage Error (MPE) considering data for links, which were not used for model development.Log-link with Gamma distribution model was observed to be the best-fitted model for the data used in this research. Models were developed by incorporating all the predictor variables at a time (backward elimination) and also by selecting the independent variables based on Pearson correlation coefficients. The results obtained indicate that land use developments have a significant influence on the travel times. Different land use categories contribute to the average travel time based on the buffer width, area type, and the link speed limit.Developing the models by classifying the links based on the speed limit (< 45 mph, 45 to 50 mph, and > 50 mph) was observed to be the best approach to examine the relationship between land use developments and the average travel time. However, capturing the land use developments within 1-mile from a link was observed to be the best approach to examine the relationship between the land use developments and the average travel time by buffer width and area type.Typically, travel times on a selected link is higher during the evening peak period compared to the morning peak and the afternoon off-peak period. The results obtained indicate that, typically, the number of lanes and the posted speed limit are negatively associated with the travel time of the selected link. Some of the important findings are listed next.1. Car wash, convenience store, department store, multi-family, office, fast food, funeral home, hospital, and supermarket type land uses within 0.5 miles from a link increase the average travel time.2. In the CBD area, department store, government and multi-family type land uses within 1-mile from a link increase the average travel time.3. In the CBD fringe / OBD area, daycare, multi-family, shopping mall, and supermarket type land uses within 1-mile from a link increase the average travel time.4. In the urban area, convenience store, department store, fast food, funeral home, multi-family, recreational, retail, and supermarket type land uses within 1-mile from a link increase the average travel time.Such findings help professionals and planners in land use planning decisions and can reduce the congestion through proactive implementation of mitigation measures. In addition to the procedure followed in the traffic impact studies, the developed relationships could be helpful to quantify the influence of land use developments on the travel time based on the type of land use development, area type, and the speed limit of the link.
