Estimating energy bounds for adoption of shared micromobility

Highlights

• Shared micromobility can reduce energy consumption of passenger travel by 2.6%.

• Micromobility induced transit trips offer highest energy saving potential.

• Micromobility energy impacts are tradeoff between energy intensity & service range.

• Sensitivity Analysis shows distance threshold has strong influence on energy impact.

Abstract

Shared micromobility has garnered widespread popularity in recent years, but limited attention has been given to the energy impacts of trips replaced by micromobility. This paper investigates the energy bounds of shared micromobility adoption. Travel demand data at the national and city level were analyzed to identify trips that can be served through micromobility, and scenarios with varying levels of micromobility adoption were evaluated. Results show that peak adoption of shared micromobility can reduce energy consumption from reported passenger travel by 1% at the national level and 2.6% at the city level, with micromobility-induced transit trips identified as the largest contributor for energy reduction. Sensitivity analysis was carried out to show how the energy impacts would change with various levels of key micromobility-related parameters, and results show distance threshold having a stronger influence on the energy impacts, compared to redistribution energy intensity.

Introduction

The past decade has seen many exciting advancements in transportation, from major progress in self-driving vehicle technology to the introduction of the mobility-as-a-service (MaaS) model. The MaaS model refers to a suite of services in which mobility is served on demand through crowdsourced or privately owned vehicles, such as ride-sharing services that can be best described as app-based shared taxi services. More recently, the MaaS model has been extended to offer on-demand mobility through smaller vehicles. These small, lightweight mobility options (commonly referred to as micromobility) build on a foundation of shared station-based manual bicycle systems (Duvall, 2012), and have been extended in the past few years to include additional vehicles such as dockless bikes, electric bikes, and electric scooters. Though micromobility vehicles are available for private ownership, this paper focuses on energy impacts from the use phase of shared micromobility, fully recognizing the fact that additional energy benefits or disbenefits are possible through personal ownership of this emerging mode. Micromobility caters to trips that are shorter in distance, compared to ride hailing (CB Insights, 2020), and is touted to reduce car trips (Crowe, 2019) and increase transit demand (Kim and Kim, 2019, Lee et al., 2019, Zarif et al., 2019). On the flip side, micromobility has raised concerns regarding rider and pedestrian safety (Mohn, 2020), as well as increased curb congestion (Hunstable, 2019). With all its advantages and disadvantages, the market share for this mode is on an increasing trend, given that shared micromobility was available in over 90 U.S. cities by the beginning of 2019 (Srivastava, 2020). Even in the wake of reduced travel due to the COVID-19 pandemic, micromobility is forecasted for a strong recovery, compared to many other transportation modes (Heineke et al., 2020).

As micromobility adoption is on an increasing trajectory, it is important to investigate the benefits, constraints, and impacts associated with the mode. Understanding this need, many researchers have tried to address operational, safety, policy, and management aspects of micromobility. However, there is currently a lack of clarity on the energy and sustainability impacts of this emerging mode. For comparison, there has been ample literature on the energy impacts of ride-hailing (Wu et al., 2018, Circella et al., 2019), as well as bounding studies on the energy impacts of future automated vehicle technologies (Stephens et al., 2016, Taiebat et al., 2019). As we foray into the next decade of integrated, electric, and sustainable mobility, it would be beneficial to have a clear understanding of the magnitude of energy impacts from the adoption of micromobility in general and shared micromobility in particular. Addressing this research need, this paper presents an in-depth analysis of the energy bounds for adoption of shared micromobility using publicly available micromobility data and the National Household Travel Survey (NHTS). For this analysis, it is important to first describe the types of vehicles included under the umbrella of micromobility. The Society of Automotive Engineers (SAE) made an initial attempt to describe and categorize micromobility vehicles (Chang et al., 2019). This study adopts the SAE terminology and defines micromobility as app-based shared mobility offered through powered bicycles (referred to as e-bikes for the rest of the paper), powered standing scooters (e-scooters), and electricity-powered seated scooters (seated scooters), along with traditional manual-powered bicycles (m-bikes).

National- and city-level data sets are analyzed to compute the energy bounds for micromobility adoption. This approach helps determine if the size of the energy impact varies with the geographic extent of analysis, which can in turn aid in decision-making around this mode at the local and federal levels. Because micromobility has been in the limelight only for a few years, travelers’ preferences toward this mode are not yet solidified. Therefore, a scenario-based approach is adopted for quantifying the energy bounds by developing a generic analysis framework and implementing the framework in the context of various levels of market adoption. The intent of the analysis conducted in this paper is to: (1) quantify the range of possible energy impacts from different levels of micromobility adoption, (2) compare and contrast the energy impacts of using micromobility for various types of trips (e.g., short trips, transit access/egress trips), (3) carry out a comparative analysis of energy savings from various types of micromobility vehicles, and (4) generalize the results of the scenario analysis by exploring sensitivity of the energy impacts to key parameters associated with micromobility.

The remainder of the paper is organized as follows. Section 1.2 details the current state of knowledge on various aspects of micromobility and identifies key gaps this study is attempting to address. Section 2 is organized into three parts. Section 2.1 presents the analysis of publicly available e-scooter data, Section 2.2 discusses the data description and the analysis framework, and Section 2.3 presents the scenario development process. Results from the national-level analysis are presented in Section 3.1, and Section 3.2 dives into an in-depth analysis of the city-level results. Section 3.3 further identifies the sources of energy impacts and Section 3.4 presents a sensitivity analysis with respect to selected analysis criteria. Section 4 presents some concluding thoughts and directions for future research.

Early studies on micromobility primarily focused on bicycle-sharing systems (Shaheen et al., 2010, Zhao et al., 2014, Yang et al., 2016). With the widespread availability and increasing market share of app-based shared micromobility, literature on micromobility has expanded to cover additional modes such as e-bikes and e-scooters. Current research on micromobility focuses on usage patterns, policy and management, safety, and—to a lesser extent—sustainability aspects.

Studies that explore usage patterns of micromobility rely heavily on data served through an application programming interface (API). Using an API, a micromobility service provider can provide access to a variety of information such as geolocations of scooters/bikes, status (available/unavailable), and duration of use. Researchers leverage publicly available APIs to conduct spatiotemporal exploration of micromobility, focusing on factors such as trip length (Lee et al., 2019, Lazarus et al., 2020), trip distribution (Duke et al., 2019, McKenzie, 2019, McKenzie, 2020, Lazarus et al., 2020), correlation with sociodemographic (Shaheen et al., 2019), and built-environment attributes (Maiti et al., 2019, Bai and Jiao, 2020, Zou et al., 2020). Coupling micromobility data with other spatiotemporal information, many studies have investigated the factors influencing demand for micromobility. Factors such as employment density, location of micromobility rebalance point (Arnell, 2019, Caspi et al., 2020), sociodemographic characteristics (Lee et al., 2019), and facility density (Lazarus et al., 2020) are all shown to affect the demand for micromobility. Studies exploring the impacts of micromobility on mode shifts reveal interesting insights such as increase in transit adoption (Kim and Kim, 2019, Lee et al., 2019), and reduction in auto usage (Barnes, 2019). Lee et al. (2019) note that micromobility adoption has the potential to reduce New York City taxi mode share by 1%. Many of the studies discussed above are mostly in the context of dense urban locations (Kim and Kim, 2019, Lee et al., 2019) and college campuses (Eccarius and Lu, 2020).

Studies focusing on policy and management aspects of micromobility identify critical issues pertaining to equity of access (Six, 2019), regulation (Goodman et al., 2019, Herrman, 2019, Shaheen et al., 2019), and infrastructure (Gössling, 2020, Zagorskas and Burinskienė, 2020) and provide suggestions through case studies (Herrman, 2019), interviews of city government officials (Goodman et al., 2019), and analysis of local news (Gössling, 2020). Among studies focusing on safety-related aspects of micromobility, some leveraged crowdsourced data to analyze encounters between pedestrians and e-scooters (Maiti et al., 2019), whereas others analyzed news reports (Yang et al., 2020) and hospital visit data (Vernon et al., 2020) to identify key factors and severity of injuries associated with crashes involving e-scooters. Attention has also been given to the privacy (Li et al., 2019), charging strategy (Masoud et al., 2019), accessibility (Six, 2019), and public space usage (Tuncer et al., 2020) perspectives of micromobility.

Although research on usage patterns, policy, and safety facets of micromobility has grown significantly over the past few years, there is limited literature on sustainability and energy-related aspects of micromobility. Even though there is a reasonable amount of literature focusing on the energy impacts of emerging transportation technologies such as transportation network companies (TNCs) and automated vehicles (AVs), the same cannot be said for micromobility. In estimating energy impacts of AVs, literature indicates that changes to travel demand, vehicle design, and operations-related factors influence energy consumption and carbon emissions within a wide range—anywhere from half to double current values (Wadud et al., 2016). Due to lack of information on AV adoption and resulting consequences, Wadud et al. (2016) base their estimates on existing literature and engineering and economic analysis. They adopt an ASIF (Activity Level, Mode Share, Energy Intensity, and Fuel Economy) framework and evaluate emissions impacts of AVs under a variety of mechanisms (or scenarios) such as platooning, right-sizing, and eco-driving. Studies on this topic have found that full automation increases the chances of higher (i.e., adverse) energy and emissions outcomes unless higher rates of sharing are adopted (Wadud et al., 2016, Ross and Guhathakurta, 2017). Similar to Wadud et al., 2016, Ross and Guhathakurta, 2017 adopt a scenario-based approach (to overcome the lack of data availability in the context of AVs) and project the energy consumption for highways for the entire United States under different levels of vehicle automation and sharing. Liu et al. (2019) use a bottom-up approach to evaluate the impact of passenger vehicle automation on greenhouse gas emissions in China. Based on four scenarios considered in their study, they report that fuel economy will be the deciding factor between a net increase or decrease of emissions relative to non-AV scenarios. Henao and Marshall (2019) collected real-world data on TNC operations in Denver, Colorado, through a quasi-natural experiment. Based on their analysis of 416 actual TNC trips, they estimate that TNC operations led to 83.5% more vehicle miles traveled (VMT) relative to the no-TNC baseline. Wenzel et al. (2019), through their analysis of the RideAustin data, report a 41–90% increase in energy use for TNC trips relative to prior mode. They use data from existing surveys across the United States to obtain shares of motorized and nonmotorized trips and derive a “bounding case” for travel shifts induced by TNCs.

Among relevant research on the sustainability aspect of micromobility, Barnes (2019) used VMT reduction as a proxy measure to estimate greenhouse gas emission benefits due to mode shift from traditional modes to micromobility. They estimated the annual auto VMT reduced by micromobility could reach 1,278,790 miles in San Francisco. Hollingsworth et al. (2019) used life cycle analysis to quantify the environmental impacts of micromobility and report that materials and manufacturing contribute about 50% of life cycle emissions, and daily collection (by a larger vehicle) for charging of e-scooters amounts to 43% of life cycle CO₂ emissions.

Aside from these two studies that foray into the emissions aspect of micromobility, no peer-reviewed research currently exists regarding energy-related outcomes of micromobility usage (to the best of our knowledge). Given the continued increase in use of shared micromobility and the fact that micromobility consumes less energy per mile of travel compared to driving, understanding the energy impacts of micromobility can help cities and transportation agencies plan and regulate micromobility operations. Put simply, knowing an answer to the question “What are the energy bounds for adoption of (shared) micromobility” can help decision makers encourage or discourage widespread deployment and adoption of micromobility. To address this question, this research effort leverages national- and city-level data sources to develop a bounding analysis for energy impacts of micromobility adoption. In order to quantify the energy saving opportunities and potential risks of increased energy use from micromobility-related travel, this study considers energy impacts associated with:

• Micromobility replacing short trips made by other modes

• Micromobility serving as access/egress modes for existing public transit usage

• Possible induced transit demand from micromobility adoption