资源环境科技发展态势分析平台

This study presents new trail demand models based on data collected between January 1, 2014 and February 16, 2016 at 32 locations in the seven major climatic regions in the continental U.S. We contribute fourfold to the literature on analysis of trail traffic demand. First, we develop a set of econometric models to predict average daily pedestrians (ADP), average daily bicyclists (ADB), and average daily mixed-mode traffic (ADM) using the 5 D's of the built environment (i.e., density, diversity, design, distance to transit, and destination accessibility), and socio-economic characteristics. Second, we test the performance of trail demand models in predicting ADB, ADP, and ADM using the leave-one-out cross-validation technique and compare the relative accuracy of the models. Third, we assess the performance of separate bicycle and pedestrian demand models in predicting mixed-mode travel demand. Fourth, we introduce a post-validation technique to advance the prediction accuracy of trail traffic demand models. The results indicate: (1) with only a few exceptions, ADP and ADB are correlated with different variables, and the magnitude of effects of variables that are the same varies significantly between the two modes; (2) The mean relative percentage error (MRPE) for bicyclist, pedestrian, and mixed-mode models equals 65.4%, 85.3%, and 45.9%; (3) Although using separate but integrated sensors to monitor bicycle and pedestrian traffic enables us to juxtapose the bicyclist demand with pedestrian demand, there is not a significant improvement in predicting total demand using these more expensive sensors; (4) A new post-validation procedure improved the demand models, reducing the MRPE of bicyclist, pedestrian, and mixed-mode models by 27.2%, 32.1%, and 14.1%. Overall, our models confirm that different variables are correlated with bicycle and pedestrian traffic volumes and that these modes need to be modeled separately. Our models can be used in practical applications such as selection of trail corridors and prioritization of investments where order-of magnitude estimates suffice.