摘要: |
In the first part of this project, a Transportation Research Center (TRC) team gathered automated hourly pedestrian counts from a sidewalk in downtown Montpelier, Vermont to determine if temperature, relative humidity, precipitation and wind affect the number of walkers. The researchers found that, after adjusting for time of day and day of week, weather and seasonal variables explained 30 percent of the variations in pedestrian volume -- and that bad weather such as cold temperature or precipitation consistently affected walking traffic, but by only a moderate amount (less than 20 percent). For the next part of this project, hourly distributions of non-motorized traffic data at 9 locations along shared-use paths in Chittenden County, Vermont were investigated for a linkage between total daily volumes, daily distributions, and surrounding land-use. The analysis failed to reveal significant variations in the hourly distributions relative to the land-use proximate to the count location. The findings were then used to identify temporal and spatial gaps to provide a robust, heterogeneous data set for non-motorized travel modeling and exposure estimation. Additional regression analyses was also conducted on a separate set of intersection-based non-motorized traffic counts to determine more generally if a connection exists between pedestrian and biking volumes and proximate land use. A geographically-weighted regression was performed which was sensitive to spatial autocorrelation in the dependent and independent variables. The University of Vermont (UVM) TRC is currently developing a new method of collecting non-motorized travel counts in rural locations, using a closed-circuit camera. This new method is being used to collect counts at the 18 new locations in support of the calculation of total bike and pedestrian miles of travel (BPMTs) in the County. |