Question 1 of 7: Land Use, Trip Purpose and Advanced Vehicle Technology

Question 1: Land Use, Trip Purpose and Advanced Vehicle Technology (20 marks)

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.

Part (a) — low-density suburban residential development. Residential development at low density changes travel behaviour through three linked mechanisms: it lowers the density of trip ends, it separates land uses, and it makes the automobile the only mode that can serve the resulting pattern at a tolerable cost in time. Density is the variable that transit economics turn on. A conventional bus route needs of the order of 25 to 40 dwellings per hectare within a 400 m walk of the stop before service can be offered at headways short enough that riders stop consulting a timetable; a subdivision built at 8 to 15 dwellings per hectare on curvilinear streets and cul-de-sacs simply cannot generate that boarding density. The transit agency responds with long headways, indirect routes and park-and-ride rather than walk-up service, and each of those responses raises the generalised cost of the transit trip relative to driving. Walking and cycling are suppressed by the same geometry from the other direction: the cul-de-sac pattern inflates the ratio of network distance to straight-line distance to typically 1.4 or more, so a destination 800 m away as the crow flies becomes a 1,200 m walk, which is beyond what most commuters will accept.

The second mechanism is the separation of home from work. Low-density residential development is almost always single-use development, so the employment that suburban residents commute to is somewhere else — the central business district, a suburban office park, or an industrial area on the far side of the region. The resulting work trip is long. Canadian journey-to-work data show the pattern consistently: commuters living in low-density outer suburbs travel substantially farther to work than residents of the inner city, and they make almost all of those trips by car, most of them driving alone. The third mechanism is self-reinforcing. Because households in these areas must own one or more cars simply to function, the marginal cost of using the car for any additional trip is very low — the fixed costs of ownership are already sunk — so the car is chosen even for trips that a transit or active mode could in principle serve. Abundant free parking at both ends of the trip completes the picture, since parking price is one of the strongest single determinants of commuter mode choice.

The engineering consequences follow directly. Vehicle-kilometres travelled per capita rise, peak-period demand concentrates on a small number of radial arterials and freeway ramps because the local street network offers no parallel capacity, and the transportation authority faces pressure to widen those facilities — which improves accessibility at the fringe and induces still more low-density development. This is the land-use/transport feedback cycle: transport investment changes accessibility, accessibility changes land value and development location, and the new development changes travel demand. Breaking the cycle requires acting on the land-use side (minimum densities, mixed use, transit-oriented development around stations, connected street grids that shorten walking distances) rather than only on the supply side.

Part (b) — work trips against non-work trips. Work trips and non-work trips differ systematically in length, in frequency, in time-of-day distribution and in how sensitive they are to travel cost, and travel-demand models treat them as separate purposes for exactly that reason.

Work trips are long and highly regular. A commuter makes essentially two of them per working day, to the same destination, at nearly the same time, five days a week; the trip is the longest routine trip most people make, because the workplace is chosen once, from the whole regional labour market, and is then fixed for years. Work trips are also strongly peaked: they create the morning and afternoon peaks that size the network, and they are the least elastic trips in the system in the short run, since the commuter cannot easily decide not to go to work. What the commuter can change is mode, route and departure time, which is why work trips dominate mode-choice and assignment models.

Non-work trips — shopping, personal business, escorting children, recreation, medical appointments, social visits — are individually shorter but collectively more numerous. In most Canadian household travel surveys they make up roughly two-thirds to three-quarters of all daily person trips, even though they contribute a smaller share of peak-period vehicle-kilometres. They are short because the destination is chosen from a much smaller opportunity set: a grocery store, a school or a clinic is selected from those nearby, so the deterrence effect of distance is far stronger for these purposes than for the journey to work. Their frequency varies greatly between households and from day to day, they are spread across the midday, evening and weekend rather than concentrated in the peak, and they are far more elastic — a discretionary trip can be deferred, combined with another, or suppressed altogether if travel becomes expensive or unpleasant.

Two further distinctions matter in practice. Non-work trips are much more often chained: a single tour leaves home, drops a child at school, stops at a store and returns, so the individual "trips" are not independent and a simple trip-based model misrepresents them — this is a principal motivation for activity-based and tour-based modelling. And non-work trips are much more likely to carry passengers, so their average vehicle occupancy is higher than the near-1.1 occupancy typical of the solo commute. Calibrated trip-generation rates and friction factors therefore differ sharply by purpose, with a steeper deterrence function for shopping and personal business than for work.

Part (c) — connected and autonomous vehicles. Advanced vehicle technology acts on travel demand through the generalised cost of travel, and the direction of the net effect is genuinely uncertain because the technology pushes on that cost in two opposite ways.

The effects that increase travel are the better established. An automated vehicle converts driving time into time that can be used for work, rest or entertainment, which lowers the perceived value of in-vehicle time; in a mode-choice model this is a reduction in the travel-time coefficient for the auto alternative, and it will draw trips from transit and lengthen the distances people are willing to commute. Automation also extends mobility to people who cannot drive today — older adults, people with disabilities, and travellers below driving age — which is a genuine and desirable increase in accessibility but still an increase in vehicle trips. Connected vehicles using cooperative adaptive cruise control can follow at shorter headways, which raises lane capacity materially; that capacity gain reduces congestion in the short run and then induces additional travel in the long run through exactly the feedback cycle described in part (a). Finally, an automated vehicle can travel empty — to a cheaper parking location, or home again after dropping its owner — adding zero-occupancy vehicle-kilometres that no conventional model accounts for.

The effects that reduce travel run the other way. Shared automated fleets, if they replace private ownership, restore a per-trip marginal price and remove the sunk-cost logic that makes the private car so attractive for short trips. Automated feeder service can solve the first-and-last-kilometre problem that keeps suburban commuters off rapid transit, making the transit alternative competitive where it is not today. Better real-time information — which is already deployable through connected-vehicle infrastructure and advanced traveller information systems, without full automation — lets travellers shift departure time or route away from congestion, which flattens the peak and improves the use of existing capacity. And if automated vehicles reduce the demand for parking in city centres, the land released can be redeveloped at higher density, which pushes the land-use system in the direction that shortens trips.

For a practising transportation engineer the operative conclusion is that the outcome is a policy variable, not a technological inevitability. If automated vehicles arrive as privately owned, single-occupant vehicles with free road space, the reasonable expectation is a substantial increase in vehicle-kilometres travelled and no lasting congestion relief. If they arrive as shared, high-occupancy services complemented by road pricing, parking pricing and transit priority, the same technology can reduce vehicle-kilometres. Planning practice should therefore test both futures as scenarios, keep the pricing and land-use instruments in view, and avoid committing irreversible capacity investments on the assumption of a single technological trajectory.