Analysis of seasonal and day-of-week traffic patterns at national parks

The National Park Service (NPS) is currently contemplating the implementation of a system-wide traffic monitoring program. While several of the national parks within this network collect continuous vehicle data at multiple stations within each park, these programs have not been examined for their efficiency and cost effectiveness. Therefore, as the NPS looks to expand their count program, this thesis investigates potential improvements using a sample set of five parks. To determine whether the national park seasonal and day-of-week traffic patterns exhibit consistency from one year to the next, the seasonal and day-of-week factors were compared across all five years. Using the Kruskal-Wallis test, it was determined that the seasonal and day-of-week factors were not statistically different from 2002 to 2006 for all five national parks. Therefore, it is recommended that the NPS consider reducing the amount of data that they collect by using short-duration counts in conjunction with a modest number of permanent counts. To determine whether data collection efforts can be shared amongst various entities, the national park traffic counts for 2002 to 2006 were compared to those of nearby state highway automatic traffic recorder (ATR) locations using correlation analyses. While the correlation values ranged from ?high? to ?negligible?, the distance between the park and ATR location had a direct effect on the magnitude of the value. Therefore, in order to achieve the greatest probability that the correlation will be ?high?, it is suggested that the NPS share data collection efforts using ATR locations within 20 miles of the park. To determine which design volume calculation method was most appropriate for the parks, design volumes were computed using two methods. Using the traditional Kfactor plot, it was determined that the 30th highest hourly volumes should be used for urban parks as this is where the ?knee? occurs. Although this is not the case for rural parks, there is no compelling evidence to suggest a more appropriate design hour. Additionally, the method recommended by AASHTO for recreational roadways resulted in volumes that were frequently exceeded. Therefore, the K-factor plot method is most appropriate for both the urban and rural parks.