Measuring and Predicting Multimodal Level of Service Measures for Urban Arterials

TEA-21 and its predecessor, ISTEA, call for mainstreaming transit, pedestrian, and bicycle projects into the planning, design, and operation of the U.S. transportation system. But the current chapters of the Highway Capacity Manual 2000 that deal with urban streets essentially address level of service (LOS) only for automobile users. Nationally recognized analysis techniques exist for the highway and transit modes, but analysis techniques for the pedestrian and bicycle modes are not as well established. At present, no nationally accepted method exists for combining the automobile, transit, bicycle, and pedestrian modes in an integrated analysis. Multimodal analysis is difficult also because most evaluation techniques have been developed from a modal perspective; LOS thresholds may not match well when one mode is compared with another. This paper discusses some research based on data from of a recent project (NCHRP 3-70) to develop and test a framework and enhanced models for predicting levels of service for automobile, transit, bicycle, and pedestrian modes on urban streets that take into account the interaction among the modes and result in consistent LOS definitions across the modes. LOS measures for auto, bike, and pedestrian facilities were based on responses of several panels of focus group subjects to films of auto, bike, and pedestrian facilities under varying conditions. Transit LOS was measured by on-board passenger surveys. At the time of this writing, data sets were available only for auto and transit. Several different auto and transit LOS models were estimated using disaggregate ordered choice models. The results of the analyses show how different aspects of urban arterial facilities and traffic affect users’ perceptions of LOS.