View/hide extracted text (may have errors)
Schoner 2 43 INTRODUCTION 44 Cities are increasingly promoting cycling as a transportation alternative to driving; arguing that 45 mode shift away from the private auto provides region-wide congestion, environmental, and 46 health benefits. (FHWA 2012a) Between 1999 and 2011, total federal and state funding on 47 bicycling and pedestrian infrastructure exceeded $7 Billion (FHWA 2012b). The Federal 48 Highway Administration just completed the federal non-motorized transportation pilot program, 49 which allocated $25 M to each of four pilot cities to measure the impacts of new infrastructure 50 on mode shift to bicycling and walking. (FHWA 2012a) Many of these projects are explicitly 51 targeted at closing “gaps” in bike routes to form a more cohesive network (Byers 2002) This idea 52 of a “network” of bicycle routes connecting the region is an important indicator of the shift in 53 transportation priorities from auto dominance to accommodation of nonmotorized modes. While 54 bicycles are permitted to use most components of the road network, bicycle-specific 55 infrastructure provides safe, comfortable routes that many bicyclists prefer over riding in general 56 mixed traffic. The network formed by bicycle-specific infrastructure in any given city, however, 57 is not as expansive or complete as the underlying road network, so cyclists often have to detour 58 to stay on dedicated facilities or ride with mixed traffic in order to complete their trip. 59 While numerous studies have looked for relationships between rates bicycling and 60 infrastructure, this concept of “networks” in bicycle infrastructure is nascent. This paper aims to 61 fill this gap in the literature by developing a protocol for evaluating bicycle infrastructure 62 network structure and testing its predictive power on bicycle commuting mode share. 63 Understanding these relationships between bicycle commuting and network features will enable 64 transportation and planning agencies to target investment in infrastructure components for 65 optimum impact on existing riders and the “interested but concerned” population of would-be 66 cyclists. 67 This paper analyzes bicycle facility networks from 74 mid- to large-sized US cities to 68 identify, quantify, and evaluate the underlying network structure. These network variables are 69 positively and significantly correlated to the journey-to-work mode share for from the 2005-2009 70 American Community Survey (ACS). Network variables factor into five scales that describe the 71 network’s size, connectivity, directness, cohesiveness/fragmentation, and average fragment size. 72 Regression models are used to test the relationship between these factors and bicycle commuters 73 per 10,000 workers, controlling for city population, land area, household structure, auto 74 ownership, and median income, showing that network connectivity is a significant predictor of 75 bicycle commute mode share in the model. This paper is organized as follows: Section 2 76 discusses literature on the relationships between infrastructure and bicycling mode share. 77 Section 3 explains the data collection process and methodology. Section 4 describes results from 78 measuring the network, factoring network results, and regression analysis. Finally, section 5 79 outlines implications for practice and opportunities for further study. 80 81 LITERATURE REVIEW 82 Planners and researchers have long sought evidence that infrastructure induces bicycling, and 83 that certain infrastructure features serve the needs of beginner bicyclists better than others. 84 Wilkinson, Clarke, Epperson, and Knoblauch (1994), in their FHWA manual on roadway design 85 treatments for bicyclists, recommend providing bicycle facilities that both serve existing 86 bicyclists and encourage new riders. They argue that a supply-driven approach of facilities 87 targeted at Group B (basic cyclists) and Group C (children) riders will encourage mode shift to 88 cycling and use of the facilities. Group B/C cyclists make detours to avoid uncomfortable road TRB 2013 Annual Meeting Paper revised from original submittal.
