1. Land Use Accessibility

Land use patterns have important impacts on accessibility. The location of activities and destinations affects accessibility, and different types of land use patterns are most suitable for different transport modes. Land use accessibility is often described as convenience, that is, the ease with which they can reach activities and destinations. A shop that is relatively accessible to consumers is called a convenience store, and a home near common destinations is said to have a convenient location.

Several land use factors affect accessibility:

• Density (number of people or jobs per unit of land area) increases the proximity of common destinations, and the number of people who use each mode, increasing demand for alternative modes such as walking, cycling and transit.
• Land use mix (locating different types of activities close together, such as shops and schools within or adjacent to residential neighborhoods) reduces the amount of travel required to reach common activities.
• Clustering (commercial districts, shopping malls, recreational centers) allows more destinations to be visited during each trip.
• Network connectivity (more roads or paths that connect one geographic area with another) allows more direct travel. The Interconnectivity Index [Ewing, 1996] can be used to evaluates how well a path or road network connects destinations.

Most people rely on commercial and public services they can reach within 10-minutes, and try to choose jobs that they can reach within a 40-minute commute (of course there are many exceptions, but these are reasonable reference values). This means, for example, that shifting travel from automobile to nonmotorized modes requires that commercial and public services be available within a convenient 10-minute walk or bike ride, and suitable worksites be located within a 40-minute walk or bike commute. Access can be evaluated at different geographic scales.

• At a very fine-grained scale, accessibility is affected by the quality of the pedestrian conditions and the clustering of activities within a site, mall, campus or small district. For example, strip commercial development tends to be less accessible than a commercial center because customers, clients and employees can walk between businesses rather than needing to drive to each destination.
• At the neighborhood level, accessibility is affected by the quality of sidewalks and bicycle facilities, street connectivity, geographic density and mix. For example, a more accessible neighborhood will tend to have shops and public services (e.g., schools) within or adjacent to residential areas so some errands can be made by walking, cycling, transit, or short car trips.
• At the regional level, accessibility is affected by street connectivity, transit service, geographic density and mix. A more accessible region will have a network of many roads (rather than just a few major arterials) and efficient transit service that makes it convenient to travel within the region by car or transit.
• Interregional accessibility refers to the quality of highways, air service, bus and train service, and shipping services.

Some transportation professionals argue that increased land use density is harmful because it increases traffic congestion. Whether this argument is right or wrong depends on how transportation is measured. Traffic-based measurement units, such as level-of-service or average traffic speed on a particular section of roadway, indicate that increased density reduces transportation system performance. However, accessibility-based measurements, such as the generalized cost required to reach common destinations, indicate that increased density can improve overall accessibility. Doubling population and business density may reduce average traffic speeds by 25% (e.g., from 40 to 30 mph), average trip distances can be cut in half because there are more activities nearby, so residents are better overall as a result.

2. Trade-offs Between Different Types of Accessibility

There are inherent conflicts between different forms of accessibility. This occurs because a vehicle's space requirements, risk and noise impacts increase with speed, and because land use patterns optimal for one mode are generally not optimal for other modes. These conflicts are manifested in many specific ways.

• Limited access highways designed for maximum vehicle mobility have poor accessibility (few offramps, driveways or cross-streets), while road designed for maximum accessibility (many driveways and intersections) cannot safely accommodate higher-speed traffic.
• Land use patterns that maximize automobile access (low density development with activities located along arterials and highway intersections) tend to have poor transit access, while transit-oriented development (clustered development with limited parking and good pedestrian accessibility) tends to have traffic and parking problems.
• Wide roads, and higher traffic speeds tend to create barriers to walking, so vehicle and pedestrian street design objectives often conflict.

For example, from a traffic-based perspective, the best location for a public school (or any major destination) is adjacent to a major roadway at the urban fringe where land is cheap in order to provide abundant, free parking to all staff and students. This assumes that most students will arrive by automobile or school bus. From a mobility-based perspective, the best school location is on a busy urban street with adequate parking, frequent public transit service, and perhaps a bike lane. This assumes that most staff and students will arrive by automobile, but some will bicycle or use public transit. From an accessibility-based perspective, the best location for a school may be within a residential neighborhood, even if automobile access is inconvenient and parking supplies are limited. This assumes that most students and perhaps some staff will walk or bicycle.

Although careful design can mitigate these conflicts, they are to some degree unavoidable. For example, a pedestrian bridge can improve nonmotorized accessibility across a busy highway, but such facilities are too expensive to build at everywhere they might be used, and they are often inconvenient to use. Similarly, structured rather than surface parking can improve walkability by reducing the amount of land devoted to parking around activity centers, but it significantly increases facility costs.

Because of these inherent trade-offs, planning decisions that favor one form of access over others can create a self-fulfilling prophecy. For example, if school planners choose a location that maximizes automobile access it probably will have high vehicle trip and parking generation rates. However, if the school is located within the residential neighborhood and designed for nonmotorized accessibility, a much larger portion of students and staff may arrive without a car.

Current transport planning practices are often biased in favor of automobile access at the expense of other modes. Increased road and parking capacity is often called an "improvement," while the negative impacts on walking and cycling access are ignored. Objective language uses neutral terms, such as "added capacity," "additional lanes," "modifications," or "changes".

3. Reference Units

Reference units are measurement units normalized to help people understand and compare impacts. Common reference units include per capita, per mile, per trip, per vehicle and per dollar. For example, a city's transport budget might be measured per capita to compare them with other expenditure categories, other years, and other communities. Roadway project costs may be measured per lane-mile, to compare with other highway projects, or per additional peak-period person trip, to compare with other ways to accommodate increased travel demand. Which reference units are used can affect how problems are defined and which solutions are considered, as described below.

• Vehicle-mile units reflect a traffic perspective that gives high value to automobile travel.
• Passenger-mile units reflect a mobility perspective that values automobile and transit travel, but gives less value to nonmotorized modes because they tend to be used for short trips.
• Per-trip units reflect an access perspective which gives equal value to automobile, transit, cycling, walking and telecommuting.
• Travel time units reflect an access perspective that gives higher priority to walking, cycling and transit travel, because they tend to represent a relatively large portion of travel time.
• Exposure time reflects the amount of time that a particular person or groups uses a particular facility or is exposed to a particular impact. Slower modes and people who stay along a street have greater exposure time than motor vehicle passengers.

Similarly, a stretch of road might carry 5,000 cars with 6,000 passengers, 100 transit buses carrying 2,000 passengers, 500 pedestrians, 200 bicycles, and have 100 adjacent homes and businesses. Traffic-based analysis, measured in vehicle-trips, considers motorists the dominant road user group, justifying roads designed to maximize vehicle traffic volumes and speeds. Mobility-based analysis, measured in person-mile, also considers motorists the primary road user group, but gives greater value to transit buses and rideshare vehicles, and so may justify giving priority to high-occupant vehicles (public transit, vanpools and carpools). Access-based analysis, measured in person-minutes-of-exposure, gives greater value to pedestrians, cyclists and residents, since they spend more time on the roadway. This justifies far greater emphasis on walking and cycling improvements, traffic calming and vehicle restrictions in urban areas.

Measurement units should reflect incremental impacts. For example, the costs of a project to increase urban road capacity intended to reduce congestion should be measured based on the costs per peak-period trip, rather than based on total trips made on the road each day, since the project does not directly benefit off-peak travelers (if roadway projects also improve safety or provide other benefits to off-peak travelers they can be assigned an appropriate portion of costs). Similarly, if a shopping mall's parking lot is only filled about 50 days per year, the cost of adding capacity should defined in dollars per additional peak-period parker and compared with alternatives, such as operating a shuttle bus service specifically at those times. When costs occur during different years their values should be adjusted to reflect inflation and discounting.

4. Modeling

Current planning practices are primarily intended to measure traffic. They often do a poor job measuring mobility and access. Travel surveys and traffic monitors often undercount short trips, non-work travel, travel by children, recreational travel, and nonmotorized links of motorized travel. Trips that are classified as "auto" or "transit" trips are often "walk-auto-walk," or "walk-bus-walk" trips, but the walking component is not counted.

The actual number of nonmotorized trips is usually much greater than what conventional surveys indicate. For example, if travel surveys only measure the portion of trips that consist entirely of walking, then walking typically represents about 5% of total trips. But if surveys measure the portion of trips that include a portion of walking, then walking typically represents about 20% of trips. Rietveld [2000] found that the actual number of nonmotorized trips is about six times greater than what conventional surveys indicate.

Similarly, traffic models can predict with precision the impacts that a road improvement may have on vehicle travel, but are generally unable to predict impacts on pedestrian mobility or land use patterns. As a result, data and models for evaluating traffic tends to be more precise but less accurate than what is available for evaluating mobility or access. Evaluating mobility and access may require sacrificing precision in order to improve overall accuracy.

Vehicle traffic and mobility can be measured and evaluated as physical activities. Access is more difficult to measure because it is affected by a variety of factors, including proximity, transportation choice and travel costs. As a result, evaluating access requires integrated transportation/land use models that incorporate both physical and economic factors and so can calculate and compare the changes in total costs that result from a change in the transportation systems and land use patterns [Abraham, 1998].