Prior to the peer review, NYMTC staff identified a list of issues to guide the discussion during the peer review. These issues are mentioned briefly below, and discussed in further detail in the following section.
Time Required to Complete an Analysis
The NYBPM is a complex model in every respect. As a result, the time to complete a modeling analysis and share results is often measured in weeks to over a month. The sheer amount of required input data spread across 28 counties and 3 states is a real obstacle to prompt response times for certain analyses. Complex models are also more difficult to use regardless of the geographic scale.
Additionally, while improved computer processing has improved run times dramatically, the complexity of the hardware and software has resulted in installation challenges for NYMTC's planning partners and consultants, creating an initial hurdle to effective and broad application of the model.
Flexible TAZ/Network Detail
The NYMTC staff is interested in exploring options to create a model that is "flexible" or scalable in terms of geographic and network detail. For example, perhaps TAZs could be aggregated on the fly within the model application (based on user-input) in areas well outside the corridor of interest. As another example, perhaps for certain applications the transit network does not need to be as detailed in terms of the number of modes, which would reduce the network coding burden and model run-time.
Non-residents/Visitors
NYBPM does not currently represent visitors explicitly. However, the region, and New York City in particular, attracts a substantial amount of visitor travel for business and leisure purposes and their travel patterns and mode usage are clearly different relative to residents of the region.
Network Consistency
The roadway and transit networks in the NYBPM are stored in different GIS layers and are not related other than by physical proximity. That is, the operating speeds for buses, for example, are not a function of modeled roadway speeds for autos. Similarly, the impact of buses on traffic operations is not incorporated dynamically (only through static preloaded bus equivalents).
Distinguishing Attractiveness of Land-Uses
The NYBPM uses five employment categories for modeling activity participation and destination choice - office, retail, industrial, institutional and other. These categories are too broad to explain the wide variety of attractions in the region with the necessary level of accuracy, For comparison ABMs recently developed for San-Diego, Phoenix, and Chicago operate with 20-25 employment categories (2-digit NAICS codes).
The core models of the NYBPM focus on the typical weekday travel of residents observed in the sample survey. The models do not necessarily address precisely the magnitude or nature of certain types of land uses in the region, such as airports, hospitals, and universities, for which, ideally, site specific data can be developed that will more accurately profile travel to and from these facilities.
The NYBPM core models do account for a significant amount of travel by residents of the region to and from airports, medical complexes and universities/colleges - most notably work travel. In addition, visiting medical facilities is a common activity reported in the survey data and included in the core model forecast. In addition, attending a University or college is one of the six activity types that are explicitly modeled in the NYBPM.
With respect to sporting facilities, major recreational attractions and tourist sites, the NYBPM models do not address travel to and from these sites in an adequate manner. Only a small portion of this travel is captured by retail and/or office employment variables. Air passenger travel, whether for residents or visitors, is not captured.
Effect of Parking Supply & Parking Pricing
The availability, price and location of parking can be significant issues affecting mode choice. The NYBPM incorporates parking considerations in several ways, most of which are standard approaches. NYMTC is interested in approaches to more realistically represent parking, particularly in Manhattan where parking constraints are unique.
The presence of on-street parking is a roadway network attribute affecting roadway operations. Parking cost for auto trips is modeled in the NYBPM, but parking supply is not directly factored into the NYBPM choice models for auto trips. Instead, an assumed terminal time is utilized consistent with area type. For transit trips, parking supply at stations is factored into transit path-finding but not as an explicit constraint.
Effect of Urban Form on Mode Choice
Similar to the parking discussion above, urban form generally can have an observable impact on mode choice. Urban form encompasses many factors, such as the presence and connectivity of mode-specific infrastructure, the diversity and density of nearby land-uses, and factors such as parking supply and convenience, and pedestrian and bike friendliness, among other things.
Throughout the NYBPM model region, there are locations with extensive non-motorized travel both as an access/egress mode and as the sole means of travel. Additionally, urban form can greatly affect roadway operations and can affect transit ridership depending on the relationship between these networks and the surrounding land-use. The NYBPM uses area type variables to assign road network operational parameters and estimate parking costs and terminal times. Additionally, size/density variables influence destination choice and trip length. However, NYMTC is interested in methods to potentially improve these aspects of the NYBPM. This is essential in view of highly diverse travel and urban conditions in the NY metropolitan region.
Time-of-day Choice
The current version of the NYBPM has a simplified timing model based on a set of predetermined look-up tables (often referred as time-of-day factors) with percentage of journeys by 30-min time intervals. The look-up tables are stratified by journey purpose, leg, mode, and some aggregate spatial categories with a substantial level of detail. However, these factors are static and not sensitive to congestion and/or pricing. This simplified technique has produced reasonable results in terms of the aggregate zone-to-zone mode matrices by four specified periods of a day. However, for further development of the time-of-day models for the NYBPM, it would be desirable to incorporate more flexible timing considerations. This would allow for better replication of individual travel patterns (in terms of journey sequencing and scheduling) as well as make the modeling system more sensitive to policy measures aimed at congestion relief.
Road Pricing
The NYBPM incorporates directional toll pricing applied at the link-level, with different tolls for trucks, and these tolls are incorporated into path-finding. Auto operating cost is a fixed input that is applied on a per-mile basis. NYMTC is looking to enhance these capabilities by potentially being able to consider complex pricing structures and reflect policies such as managed lanes with dynamic pricing. Further, NYMTC is interested in mechanisms to make network coding more efficient related to network pricing generally, including roadway pricing, transit fares and operating costs, among other factors. More broadly, NYMTC is interested in defensible validation strategies related to pricing.
Freight/Commercial Vehicles
The commercial travel models of the NYBPM address two components of commercial travel - trucks moving goods and other commercial vehicle traffic. Trucks are defined as vehicles with at least 2-axles and 6-tires, while other commercial traffic consists of commercial delivery vehicles, sometimes also referred to as "vans" in NYBPM documentation. While addressing commercial traffic as part of the overall NYBPM regional models was considered essential, the emphasis for the initial NYBPM was clearly on developing an advance set of private passenger travel models. The resources for development of the commercial travel element were significantly more limited. Consequently, rather than grounding these models in the overall framework of freight or goods movement analysis, the methodology targeted an empirically oriented modeling of truck and other commercial traffic that would make maximum use of vehicle class traffic count and origin-destination (O-D) data in the region. The freight and commercial vehicle models utilize simplified trip-based models followed by matrix estimation to match observed truck counts. Future truck and commercial vehicle demand is factored from the base year using fratar methods.
Dynamic Traffic Assignment
NYMTC currently utilizes a static aggregate equilibrium assignment model. NYMTC is interested in whether a Dynamic Traffic Assignment model could be implemented in the region at the mesoscopic level, and wants to better understand the pros and cons. With a static model, it is challenging to reflect real operating conditions in the urban core of the NYBPM model region given the high degree of congestion and variability in operating characteristics by time-of-day. In particular, significant queuing on the most critical facilities (bridges and tunnels around Manhattan) has to be addressed.
Land-use Models
NYMTC does not currently rely on a land-use model to forecast sub-area land-use allocation. While the idea of a land-use model is appealing, jurisdictional issues present an obstacle along with the sheer size of the geography. Like many MPOs, NYMTC has a desire to do scenario-based planning in a theoretically consistent manner.
