Literature Reviews

The expansion of on-demand delivery services has been made possible by ghost kitchens and dark stores – grocery fulfillment centers which are located near consumers but are not open to customers [1]. These fulfillment centers have created new real estate opportunities. Several major ghost kitchen operators are known for building large portfolios out of warehouses, empty strip malls, or other storefronts near areas with growing on-demand food-delivery markets [1]. Restaurants are dispersing away from ground-floor locations in popular retail districts as ghost kitchens increase their urban real estate [1], [2].

One emerging area of study is the impact of on-demand delivery services on restaurant formation and viability. The services charge participating restaurants delivery fees as high as 30 percent of order value, though some cities have imposed caps of 15 percent [3].

Scholars have posited that freight transfer hubs will be placed near interstate highways and on the fringes of regions where automated trucks drop trailers to be picked up by human-operated trucks [1], [2]. However, this is an emerging area of practice and available research has not yet considered implications for land use.

Ride-hail use varies both by land use and demographics. In general, people are more likely to use ride hail services in dense, urban areas [1], [2]. Ride-hail users in the United States tend to own fewer cars, and are more likely to use public transit, than the average resident [2]. There are exceptions, notably Los Angeles, where ride hailing is popular in both urban and lower-density neighborhoods [3]. A separate study from California found that people in lower density suburban and rural areas who used ride hail services tended to earn higher incomes; in contrast, urban ride hail users tended to earn lower-incomes [4].

Given that ride-hail trips are more frequent in urban areas, it is unsurprising that places with high rates of ride-hail use also tend to have high rates of street parking occupancy [5]. Ride-hail has the potential to alleviate curb congestion if a sufficient threshold of car trips are replaced. Ride-hail users may select the service specifically to avoid cruising for parking where few curb spots are available, and thus free up a longer-term parking spot [5]. However, those freed up spots may quickly be taken up by drivers who would otherwise have parked elsewhere, parked at a different time, or not made the trip by private vehicle at all. Moreover, ride-hail drivers must compete for curb access when dropping off riders, and thus temporarily congest the curb. Additional research is needed to better understand the impacts of ride-hail on land use and curb congestion.

Many studies show that Autonomous Vehicles (AVs) could change the layout of urban areas [1], [2], [3], potentially leading to dispersed development or densification of cities. By lowering travel expenses, AVs could influence residential and work locations, potentially leading to more pronounced urban sprawl. For example, Moore et al., [4] used a web-based survey of commuters in 2017 in the Dallas-Fort Worth Metropolitan Area (DFW) and predicted a substantial extent of urban sprawl up to a 68 percent increase in the horizontal spread of cities due to AVs. AVs could also increase urban density by decreasing the need for parking, leading to more dense and mixed use development.

AV could increase trip lengths and induce suburban and exurban development [5], [6], [7], [8]. Nadafianshahamabadi et al., [9] utilized an integrated model of land use, travel demand, and air quality. The modeling is designed for the Albuquerque, New Mexico metropolitan area to demonstrate that AVs encourage development at the urban fringe. While jobs and population typically migrate outward in tandem, trip lengths and overall travel demand continue to rise due to the relatively low density in these emerging areas compared to traditional urban employment centers. Similarly, Gelauff et al. [10] used equilibrium model to simulate spatial effects of AVs and found that population tends to increase in large metropolises and their suburbs, at the expense of smaller cities and non-urban regions given high automation with good public transport systems in Netherlands. Carrese et al. [11] used discrete choice modeling and traffic simulation to study the residential relocation due to different time perception. Results show that about 40 percent of respondents would move to the suburbs under the AV regime in Rome, Italy, and travel time would increase by 12 percent for suburban resident commuters.

Besides contributing to the development of new peripheral centers, AV has the potential to densify the existing urban landscape by reallocating space for residential, economic, and leisure activities [12]. Zakharenko [1] concluded that with the introduction of AVs, the need for daytime parking may shift to outlying areas, which would allow for denser economic activity and increased land rents in downtown areas. As AVs potentially reduce car ownership, it's anticipated that less space will be required for parking, which could give rise to more high-density and mixed-use developments [13], [14], [15]. Zhang and Guhathakurta [16] developed a discrete event simulation model to assess the impact of Shared AVs (SAVs) on urban parking land use in Atlanta, Georgia and concluded that SAV can reduce parking land by 4.5 percent at a 5 percent market penetration level and each SAV can emancipate more than 20 parking spaces. However, some research indicates that vehicles are traveling longer distances daily, and there could be an increase in parking space on the outskirts [17], [18].

In general, most studies found that private AVs can potentially lead to dispersed urban development, while SAVs are expected to contribute to densification of city centers. Current areas for future research include: 1) AV effects on people's residential and employment location decisions, recreation spaces and supply of infrastructure. 2) long-term effects of AVs on urban land use patterns to promote AV adoption with efficient use of land. 3) infrastructure adaptation to fully accommodate the new traffic dynamics and parking needs introduced by AVs [19].

A review of the literature using Google Scholar and ProQuest yielded no applicable research, indicating a probable gap in the literature.

Carshare is particularly useful for people who either live in car-dependent areas and cannot afford a private vehicle, or for urban car-less people who enjoy a diverse array of transportation options that they supplement with driving for trips that require longer-distances, multiple stops, or more storage capacity [1], [2]. Carshare benefits from economies of scale in densely populated cities, and firms can more easily adjust their prices and grow their network flexibly according to consumer demand. Carshare is particularly effective when parking is scarce, there are transit hubs nearby, and land uses are mixed [3]. In lower-density areas, carshare can be more challenging to implement, as people are more likely to enjoy plentiful parking, own their own cars, and have fewer alternatives to driving that would make them more likely to choose to forgo private vehicles [4].

The success of demand-responsive transit (DRT) and microtransit programs, measured by ridership, depends in part on land use. Whereas traditional fixed-route public transit services are most efficient in densely-populated areas, DRT/microtransit programs offer a smaller-scale alternative that can prove a more cost-effective solution in lower-density suburban and rural areas [1]. There are some exceptions: DRT/microtransit programs in urban areas are sometimes designed to complement fixed-route transit by providing a first-last mile solution to connect riders to transit, or by offering supplemental services for gaps in the transit network (during off-hours, or expanding the service area) [2]. However, evidence that DRT/microtransit can increase transit ridership in cities is mixed [3]. In rural and suburban areas, however, DRT/microtransit may serve particular demographic groups, such as older adults as a form of paratransit, and commuters who can collectively share a service to a specific jobs center [1]. In areas where vehicle ownership and use is high, and the DRT/microtransit service operates in a limited area, ridership can be low [1].

Micromobility works best when the land use and transportation system supports it. The typical scooter share or bikeshare trip is under two miles and takes 11-12 minutes [1]. Micromobility - both manually-powered or electric-powered - may be faster than walking, but nonetheless slower than driving, and leaves users exposed to the elements and street traffic. Streets that are well-connected [2] and dense with a mix of establishments and residences, and robust transit options shorten trip distances and times, and, in turn, facilitate micromobility trips. A meta-analysis of shared micromobility programs found that ridership increased with population density, employment density, bus stops and metro stations, and bike infrastructure [3]. In contrast, low-density neighborhoods with few young people and zero-car households have less access to micromobility services [4]. In the longer run, micromobility may ultimately impact land use by providing more transportation nodes and extending the reach of shared transportation services [5]. A floating bikeshare or carshare service, for example, may enable residents in outlying urban areas to connect to a city’s fixed-route transit system.

There is a body of research related to connected autonomous vehicles (CAVs) and land use, but with a focus on the implications of automating vehicles as opposed to the connective technology. There is little research focusing on how connected vehicles (CVs) and Vehicle to Everything (V2X) technology will impact land use demand and what policies may be needed to better integrate (CAVs) into existing transportation systems.

By bundling multiple modes into one interface and payment scheme, Mobility-as-a-Service (MaaS) can both induce mode shift [1] and generate new trips [2], which has implications for urban land use. Early research suggests that those most likely to use MaaS services are those who already use public transit frequently [3]. However, price structure specifics, like the number of discounted rides, geofencing, and unlimited options, can determine MaaS user mode choice [4], which can then impact congestion and parking demand. MaaS schemes that incentivize private auto drivers to switch to public transit may ease parking demand and congestion, but schemes that incentivize the switch to services like ride-hail or carshare may exacerbate congestion. MaaS services may also induce mode changes from active transportation modes like biking and walking towards public transit and ride-hail [1], with unclear implications for congestion and use of infrastructure like bike lanes and sidewalks. Future research should consider the impact of MaaS on parking demand in dense urban areas.