16-Civ-A6 Highway Design, Construction, and Maintenance · May 2016
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Paper format: 98-Civ-A6 Transportation Planning & Engineering, National Examination May 2016. Seven questions, all of equal value (20 marks); any five constitute a complete examination. Closed book, one two-sided aid sheet permitted. Three hours. All seven questions are solved below as a study resource.
Reference texts. Mannering, Washburn & Kilareski, Principles of Highway Engineering and Traffic Analysis (Wiley) — queueing, shock waves and traffic-stream models; Papacostas & Prevedouros, Transportation Engineering and Planning (Prentice Hall) — the four-step demand model; Ortuzar & Willumsen, Modelling Transport (Wiley) — trip generation, distribution, mode choice and assignment; Roess, Prassas & McShane, Traffic Engineering (Pearson) — signalised-intersection delay.
Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.
Low-density suburban development changes travel behaviour through two linked mechanisms: it lengthens the distance between where people live and where they work, and it destroys the density thresholds that any mode other than the private automobile needs in order to function. The two effects reinforce each other, which is why the automobile share of suburban work trips in Canadian metropolitan areas is routinely above 80 per cent while the same figure in the pre-war inner city is often below half of that.
The distance effect follows directly from the geometry of dispersal. Employment in a metropolitan region remains far more concentrated than housing — central business districts, hospital and university precincts, and suburban office and industrial parks each pull thousands of workers from a catchment measured in tens of kilometres. When new housing is added at eight to fifteen dwellings per hectare on greenfield land at the urban edge, the average worker is placed farther from every one of those employment clusters than an equivalent household in an established neighbourhood. Average commuting distance therefore rises, and because low-density subdivisions are typically served by curvilinear collector streets that funnel onto a small number of arterial and freeway access points, the actual route distance exceeds the straight-line distance by a larger margin than in a gridded street network. The same street pattern lengthens the beginning of every trip, which is the part of the journey a pedestrian or cyclist would otherwise perform.
The mode-choice effect follows from the economics of transit supply. Conventional fixed-route bus service needs a corridor demand of roughly forty to fifty dwellings per hectare, or a comparable concentration of employment, before service can be offered at headways short enough that waiting time is not punitive. A suburb built at a third of that density can only be served at thirty- or sixty-minute headways, which converts what would be a five-minute wait into an out-of-vehicle penalty that mode-choice models weight at two to three times the value of in-vehicle time. Walk access compounds the problem: a low-density street pattern with cul-de-sacs and no continuous sidewalk network puts many dwellings more than the 400 m that most riders will walk to a bus stop. In utility terms, the out-of-vehicle time coefficient — the largest negative coefficient in the mode-choice models used in Questions 7 of this paper — multiplies a large access and waiting time, driving the transit utility far below the automobile utility and pushing the predicted transit share towards zero. Ample free parking at both the suburban origin and the suburban workplace removes the last cost disadvantage of driving.
Three consequences follow for the transportation planner. Vehicle-kilometres travelled per household rise more than proportionally to the number of households, because both trip rate and trip length increase — suburban households own more vehicles, and vehicle ownership is itself one of the strongest explanatory variables in the cross-classification and regression trip-generation models of Question 3. Peak-hour demand becomes concentrated on a small number of radial arterials and freeway ramps, producing exactly the oversaturated approach analysed in Question 2. Finally, because trip lengths are long and the mode is almost exclusively the single-occupant automobile, energy consumption and greenhouse-gas emissions per worker are substantially higher than for a comparable worker in a transit-served neighbourhood. The policy response, discussed in part (b), is either to raise density and mix land uses so that transit and active modes become viable, or to attack the occupancy of the vehicles already on the road.
Travel demand management (TDM) seeks to change the amount, timing, routing or mode of travel rather than to add capacity. The subset of TDM measures that raises average vehicle occupancy works by making the shared ride cheaper, faster or more convenient than the solo drive, and the effective programmes always combine an incentive with a corresponding disincentive.
Ridesharing programmes. Employer-sponsored carpool and vanpool matching, supported by a guaranteed-ride-home provision that removes the fear of being stranded by a late meeting, converts single-occupant trips into shared ones directly. A vanpool carrying seven commuters removes six vehicles from the peak flow. Because the marginal cost of an added passenger is close to zero, ridesharing is the cheapest occupancy measure available and it is the first one Canadian employers adopt.
Priority infrastructure for high-occupancy vehicles. High-occupancy vehicle (HOV) and high-occupancy toll (HOT) lanes give the shared ride a travel-time advantage that the solo driver cannot buy at any price short of the toll. This is the strongest lever available, because it changes the relative in-vehicle travel times that enter the mode-choice utility directly rather than acting on cost alone. Queue-jump lanes, transit signal priority and preferential parking spaces at the destination work the same way at smaller scale.
Pricing measures. Parking cash-out (paying an employee the cash value of the parking space they decline), workplace parking charges, congestion pricing on the approach corridor, and transit-pass subsidies all raise the private cost of the solo drive relative to the shared or transit trip. Since the travel-cost coefficient in a mode-choice model is negative, any of these shifts the utility balance measurably; parking price is the most powerful single instrument because the commuter perceives it as a direct out-of-pocket cost paid every day.
Work-schedule measures. Compressed work weeks, staggered start times, flexible hours and telework do not raise occupancy in themselves but they remove peak-period trips altogether and they make ridesharing easier to organise around a common schedule. A four-day compressed week removes twenty per cent of a worker's commuting trips.
The effects on the three quantities the question asks about are as follows. Travel patterns change first in mode and occupancy rather than in destination: the number of person-trips is essentially unchanged, but the number of vehicle-trips falls, so peak-hour vehicle demand on the corridor drops. Schedule measures also spread the arrival profile, flattening the peak and reducing the duration of oversaturation. Some trips are re-timed to the shoulder of the peak, and telework removes them entirely. Travel time improves in two distinct ways. Participants in the rideshare or HOV programme gain directly through the priority lane. Everyone else gains indirectly: because delay on a congested facility is a strongly convex function of the volume-to-capacity ratio, a small reduction in vehicle demand produces a disproportionately large reduction in delay for the vehicles that remain. Against this must be set the detour and pick-up time the carpool passenger accepts, which is why the guaranteed ride home and the priority lane matter so much. Fuel consumption falls for three compounding reasons: fewer vehicle-kilometres are driven for the same number of person-kilometres; the remaining traffic runs at speeds closer to the fuel-efficient range instead of in stop-and-go flow; and the number of cold starts, which dominate emissions of hydrocarbons and carbon monoxide, is reduced in proportion to the number of vehicle-trips eliminated. Canadian evaluations of comprehensive employer TDM programmes typically report drive-alone shares falling by five to fifteen percentage points, with the larger reductions where parking pricing and an HOV facility are applied together.
Trip generation is the first step of the four-step demand model, and it is deliberately split into two halves that are estimated from different variables. Trip production is the number of trips generated at the home end of a home-based trip, or at the origin of a non-home-based trip; it is a property of the households living in the zone. Trip attraction is the number of trips drawn to the zone at the non-home end; it is a property of the activities located in the zone.
The dominant production factors are household size, household income, vehicle ownership, and the number of employed residents, with dwelling type and residential density as secondary variables. Household size is the strongest single predictor because trips are made by people: a five-person household simply contains more workers, more students and more shoppers than a one-person household, and the cross-classification table in Question 3 shows the trip rate rising monotonically from about 2.2 to about 12 trips per household across that range. Vehicle ownership acts as both a cause and a symptom — owning a car lowers the generalised cost of every discretionary trip and so induces additional travel, while households that already travel a great deal buy cars. Income raises trips per household through both of those channels and by adding discretionary shopping, recreation and social trips. Employment status determines the number of home-based work trips, the most stable and most peaked component of demand. Residential density and dwelling type act in the opposite direction: at equal size and income, an apartment household in a dense mixed-use zone makes fewer vehicle trips than a detached-dwelling household in the suburb described in part (a), because destinations are closer and some trips are chained or walked.
The dominant attraction factors are employment by type, floor area by land use, school and university enrolment, and retail sales or commercial floor space. Employment is the primary driver of home-based work attractions; retail floor area and sales drive shopping attractions; enrolment drives school attractions. Zonal accessibility — a measure of how many opportunities can be reached from the zone within a given travel time — modifies both halves, and is the mechanism through which the land-use pattern of part (a) feeds back into the demand model.
These factors matter for the forecast in a specific way. Production and attraction models are estimated separately because their explanatory variables are different in kind (household characteristics versus land-use and employment characteristics) and because the planning agency forecasts them from different data sources: population and household projections on one side, employment and floor-space projections on the other. Trips are also stratified by purpose — home-based work, home-based other, non-home-based — because each purpose responds to a different set of variables, has a different temporal profile and a different sensitivity to cost, so pooling them would blur every coefficient. Finally, because the aggregate totals produced by the two sub-models will not agree, the model must be balanced (conventionally by scaling attractions to match productions) before the trip table can be distributed in step two, which is exactly the singly constrained structure used in Question 5.
