Understanding Life-Cycle Greenhouse-Gas Emissions of Shared Electric Micro-Mobility: A Systematic Review

Abstract

In recent years, the implementation of shared electric micro-mobility services (SEMMS) enables short rentals of light electric vehicles for short-distance travel. The fast expansion of SEMMS worldwide, promoted as a green mobility service, has raised a debate about its role in urban mobility, especially in terms of environmental impacts such as climate change. This article presents a systematic review of the current knowledge on the environmental impacts of SEMMS, with a special focus on the use of life-cycle assessment (LCA) methods. The study offers a detailed analysis of the global warming potential of SEMMS and its critical phases. It is found that shared e-scooters have the greatest greenhouse-gas emissions during their life cycle, while emissions from shared e-mopeds and shared e-bikes tend to be lower. The literature reveals that the materials and manufacturing phase is the most important one for the environmental impact of shared e-scooters, followed by the daily collection of vehicles for charging. The article also identifies influential factors in the sensitivity analysis and the potential for net-impact reduction accounted for mode substitution. Finally, the article identifies further research areas aimed at contributing to the adoption of environmentally responsible practices in the rapidly expanding field of shared services in cities.

1. Introduction

The term “shared electric micro-mobility” refers to a transportation service model in which users have access to a fleet of lightweight electric vehicles, such as scooters, bikes, and mopeds, for short-distance travel [1] with a speed not exceeding 45 km/h [2]. These shared services offer a convenient, affordable, and environmentally friendly alternative to traditional modes of transport for short distances, with the potential to significantly reduce greenhouse-gas (GHG) emissions by replacing trips that would otherwise be taken by combustion-engine vehicles. However, the growth of shared electric micro-mobility services (hereinafter SEMMS) raises concerns about safety, pedestrian impacts, and their true environmental benefits.

SEMMS are at a crucial juncture in determining how to maximize their benefits to society and how competent public authorities should regulate their services to meet the challenges of sustainable urban mobility [3]. So, it becomes essential to evaluate their environmental sustainability to ensure that their benefits are not offset by unintended negative consequences, such as those that adversely affect public transport and active modes, arise out of induced trips or even an increment of the global warming potential (GWP).

Life-cycle assessment (LCA) emerges as a valuable analytical tool for assessing the comprehensive environmental impact of shared electric micro-mobility services throughout their entire life cycle. LCA provides a holistic approach that encompasses the environmental burdens associated with the production, operation, maintenance, and disposal of these micro-vehicles [4]. By considering a variety of factors, such as raw-material extraction, manufacturing, energy consumption, emissions, and end-of-life management, LCA enables a thorough evaluation of the climate-change performance of these services.

This review paper aims to explore the environmental impacts of SEMMS through the lens of LCA, since it is an innovative and relatively under-researched topic. By synthesizing the existing literature and studies in this domain, we seek to provide a comprehensive understanding of the environmental performance and the potential benefits and drawbacks associated with these services. In contrast to Arbelaez-Velez [5], which provided a broad overview of environmental impact studies of all shared mobility services, the current paper offers a more detailed analysis of the critical phases of the life cycle of GHG emissions employed in the reviewed literature, including vehicle lifespan, vehicle manufacturing, electricity generation source mixes for charging, operations of collection and distribution operations (rebalancing), maintenance, lifetime mileage, and the impact of SEMMS on public transportation.

Furthermore, the objective is to develop a robust understanding of the environmental implications associated with SEMMS, to identify opportunities for improvement and deficiencies and gaps in studies of SEMMS from an LCA perspective, as well as to study the role of policy and regulation in shaping their environmental outcomes. With this in mind, based on the literature, this paper answers three research questions: RQ1—What are the impacts of SEMMS on climate change when analyzed through the lens of LCA and their comparison among different services?; RQ2—What are the main key factors that significantly impact the environmental performance of SEMMS?; and RQ3—What are the gaps in the existing research on the environmental impacts of SEMMS? This paper sheds light on this issue and contributes to the development of sustainable urban transportation systems and promotes the adoption of environmentally responsible policies and practices in the rapidly expanding field of shared electric micro-mobility services.

This study conducts a systematic review of the literature on the recent developments (methodologies, databases, and tools) for analyzing the environmental impacts of SEMMS using LCA. Subsequently, it delves into the GWP of SEMMS and its critical phases of LCA, which include materials and manufacturing, use, rebalancing, and end-of-life considerations. The article also reviews and discusses the sensitivity analyses conducted in the literature by identifying influential factors and the potential for net-impact reduction, accounting for mode substitution. Finally, we discuss policies, point out recommendations, and identify future research priorities.

2. Methodology

The study follows a systematic literature review (SLR) methodology, which is a comprehensive research method designed to locate, select, and analyze all published research related to a specific question or topic [6]. The SLR consists of four phases designed to identify the existing literature related to LCA for SEMMS and to analyze the main trends and research gaps. The review includes studies where emissions are calculated using the LCA method.

Figure 1 represents the SLR flow diagram with the four main phases: identification, screening, eligibility, and inclusion.

In the identification phase, keywords, databases, and search limitations were defined. The substrings used in the systematic search relevant to the scope of this study were identified and classified into three different categories: LCA (A group), sharing (B group), and electric modes (C group). The search focused on identifying studies that make all possible combinations of two words from these categories using the Boolean operators “AND” and “OR” to select only papers that contain at least one of the relevant keywords or terms. The Web of Science (WoS) and Scopus databases were used for the search, yielding more than two thousand records from the databases.

Additionally, we reviewed 21 documents from government institutions, private consulting firms, and specialized forums, such as the International Transport Forum (ITF), which were included in the final review analysis. These documents met the selection criteria developed for the systematic review. At the end of the identification phase, 86% of articles were removed mainly due to duplications, lack of accessibility, or the fact that they did not contain at least one keyword from the defined groups in their title.

Then, during the screening phase, many articles were excluded because they were not related to the specific subjects of interest, such as LCA, shared mobility, and electric micro-vehicles. Most of these papers focused on topics such as user perceptions, usage patterns, health effects, or cost and social assessment. Ultimately, only 41 articles met the initial scope criteria. The final selection indicates that the literature chosen for the analysis focuses primarily on China, North America, and Europe, with less emphasis on Oceania and the rest of Asia.

The eligibility phase was the last filter. Publications that did not fit within the scope of the study were excluded, such as non-shared micro-mobility and non-electric micro-vehicles, as well as hybrid micro-vehicles, smart bikes, and autonomous bikes and scooters. After reviewing the full text of papers, only 14 studies were ultimately selected. The final phase of the SLR involved a general web search to identify additional articles or studies, resulting in the inclusion of two additional papers. Moreover, of all documents found on the Web, only four were added to the body of knowledge.

In total, 20 studies were included in the analysis. They were published between 2019 and 2023, with the majority coming from Europe, followed by North America, Asia, and Australia. Most of the identified literature addresses shared e-scooters, with a smaller number of studies focusing on e-bikes and e-scooters (see Section 4).

3. LCA Software and Databases for SEMMS

The LCA method, governed by ISO standards 14040 and 14044 [4,7], is used to determine the potential environmental effects of a product, service, or system throughout its life cycle, in relation to a functional unit. It includes raw-material acquisition, material production, manufacturing, assembly, transport, use, and disposal. This allows for comparisons between systems based on similar functional units. The identified review studies use different life-cycle impact (LCI) databases, software tools, or life-cycle impact-assessment (LCIA) methods.

Table 1. Summary of databases, software, and methods of LCA on SEMMS.

	Name	Description	Study
Databases and Software	CPCD	China Products Carbon Footprint Factors Database.	[8]
	Ecoinvent	Ecoinvent is a life-cycle inventory database that provides detailed information on the environmental impacts associated with the production and consumption of products and services.	[9,10,11,12,13,14,15,16,17,18,19]
	GaBi	Gabi by Sphera Solutions is a comprehensive resource that includes both software and extensive life-cycle inventory (LCI) data, allowing one to perform life-cycle analysis.	[20,21,22,23,24]
	OpenLCA	OpenLCA is open-source free software for sustainability assessment and/or life-cycle assessment.	[9,10,11]
	SimaPro	SimaPro is a software tool for life-cycle assessment (LCA), allowing users to model, analyze, and interpret the environmental impacts of products and processes.	[15,16,17,18]
	TRACI	“Tool for Reduction and Assessment of Chemicals and Other Environmental Impacts” that provides characterization factors for life-cycle impact assessment.	[13]
	Umberto	Umberto is an LCA software developed by Ifu Hamburg GmbH. It facilitates the assessment of environmental impacts.	[12]
Methods	CML	The CML method developed by the University of Leiden is used to calculate how much impact the product has on the environment.	[10,20,21,22,23]
	CED	The cumulative energy demand (CED) method quantifies the primary energy usage throughout the life cycle of a good or service.	[9,19]
	GREET	‘The Greenhouse Gases, Regulated Emissions and Energy Use in Transportation Model’ is a tool used to assess the effects of various technologies of energy sources, products, and vehicles.	[2,25,26]
	ILCD	‘The International Reference Life Cycle Data System’ serves as a standardized foundation, ensuring uniform, reliable, and quality-assured life-cycle data, methodologies, and evaluations at the midpoint and endpoint.	[12,15,17]
	IPCC	The IPCC 2013 is an environmental assessment method that relies on data published by the Intergovernmental Panel on Climate Change.	[9,10,14]
	Pas 2050	Specification for the assessment of the life-cycle greenhouse-gas emissions of goods and services.	[8]
	ReCiPe 2016	This method is used to conduct life-cycle impact assessments and offers a state-of-the-art approach to converting life-cycle inventories into impact scores at both the midpoint and the endpoint levels.	[9,11,16,18,19,24]

includes a description of the databases, software, and methods used in the reviewed studies.

As can be seen, the Ecoinvent [27] database has emerged as the predominant and most comprehensive foundational database extensively used to conduct SEMMS LCA. More than half of the reviewed studies relied on this database. Furthermore, the Ecoinvent database is widely integrated into prominent software tools, such as SimaPro, GaBi, Umberto, [28,29,30], and even open-source alternatives like OpenLCA [31], encompassing a wide range of uses. Approximately two-thirds of these LCA software were implemented in SEMMS studies, with the GaBi software being the most used.

Regarding the impact assessment, ReCipe [32] and CML [33] in the updated 2016 version from 2016 were implemented in half of the reviewed studies, followed by the ILCD [34] and IPCC [35] methods in one-third of them. Furthermore, Hollingsworth et al. [13] in Raleigh, USA employed the TRACI [36] to convert inventory results into environmental impacts. Finally, only one study used Chinese databases [37]. Figure 2 shows the related databases, software, and methods used in the literature to analyze SEMMS LCA.

As shown in Table 1, several studies used the combination of two or more methods. Wortmann et al. [19] and de Bortoli [9] utilized the cumulative energy demand (CED) [38] approach to quantify the primary energy demand associated with the life cycle of their assessed SEMMS. They also employed other impact-assessment methods, such as ReCiPe or IPCC, to calculate additional impact factors. In another study, de Bortoli and Christoforou [10] calculated climate-change impacts using the CML method, incorporating the IPCC 2013 guidelines. In their turn, Chester [25] in the USA, Krauss et al. [26], and Cazzola and Crist [2] estimated the LCA of SEMMS using the GREET model [39].

4. Global Warming Potential of SEMMS and Critical Phases of LCA

This section shows a thorough examination and comparison of the findings concerning the global warming potential (GWP), which is a common metric that reports grams of carbon dioxide equivalent per passenger kilometer (g CO₂eq./pkm). Identifying the critical phases of LCA specific to SEMMS is essential to gain a deeper understanding of their true environmental impact. These critical phases include factors such as the materials used for manufacturing, the energy sources used for charging, and the overall operational efficiency of the SEMMS.

4.1. Life-Cycle Emissions on SEMMS

After the revision of different research works on the life-cycle emissions of SEMMS, Figure 3 summarizes and compares the GWP of SEMMS from the literature review divided by micro-mobility mode from different case studies around the world. It is shown that e-scooter sharing appears to be the most polluting micro-mobility mode and the one that has been studied the most in the literature. At first glance, there is a large variability in the results, from 55 to 213 g CO₂eq./pkm. One of the reasons for such a great difference is the shorter lifespan of shared e-scooters compared to other vehicles (for a detailed explanation, see Section 5).

The results, as shown in Figure 3, indicate that studies on a shared e-scooter, considering a lifespan of more than 12 months, obtained a GPW between 55 and 109 g CO₂eq./pkm [2,9,10,12,15,23,24,26]. On the contrary, lifespans shorter than 12 months achieved GWPs between 126 to 213 g CO₂eq./pkm [13,14,16,18,20,22,25]. An extreme case from a study carried out in Lisbon reached 803 g CO₂eq./pkm, with a lifespan of 45 days, according to Reis et al. [17]. Note that the study by Reis et al. [17] has not been included in Figure 3 due to its high results. The findings show that the results are quite sensitive to the lifetime of the e-scooter.

Dockless systems or free-floating shared bike services do not have a fixed location to pick them up, and users can park the bikes in many different places within the service area [40]. However, this increases the GWP because of larger distances for collecting and conducting service activities such as rebalancing. In contrast, station-based bike-sharing systems have fixed locations requiring infrastructure, such as stations and docks. As a result, the use of a station-based bike service requires less rebalancing (that is, collecting and servicing the bikes) [41].

Figure 3 illustrates that dockless e-bike sharing systems exhibit higher emission levels, as demonstrated by Krauss et al. [26] and Sun and Ertz [18]. Investigations carried out by Zhu and Lu [8] in China and Felipe-Falgas et al. [11] in Barcelona reveal that station-based e-bike services demonstrate notably lower emissions, even when accounting for the inventory of docks and station infrastructure. These results go hand in hand with shared non-electric bike services. Two studies compared both systems with similar results. The dockless system emitted between 118 and 130 g CO₂eq./pkm, while the station-based emitted between 65 and 68 g CO₂eq./pkm [41,42].

Regarding e-moped sharing services, only five studies have evaluated their GWP, which ranges from 34 to 79 g CO₂eq./pkm [2,9,11,21]. On average, shared e-mopeds exhibit a slightly lower GHG per passenger kilometer compared to shared e-bikes, as shown in Figure 3. This makes e-mopeds the shared electric micro-vehicle with the lowest emissions. Several important factors contribute to this difference. First, rebalancing logistics involving swapped batteries plays a significant role [19]. Additionally, the average occupancy of e-mopeds often exceeds one passenger per vehicle [11,21].

4.2. Critical Phases of LCA

To obtain a more specific understanding of the impacts of SEMMS, it is necessary to explore the critical phases associated with them. This type of LCA examines the environmental impact of different phases of SEMMS, including the materials and manufacturing phase, the usage phase (including rebalancing), the end-of-life (EoL) phase, and other phases with minimal impact.

4.2.1. Materials and Manufacturing Phase

Among the three types of shared electric micro-mobility services examined, e-scooters exhibit the highest share of emissions during the materials and manufacturing phase, accounting for more than 60% of the total SEMMS emissions on average (Figure 3). The two most impactful materials on e-scooters are aluminum and lithium-ion batteries [16,20,23]. According to Hollingsworth et al. [13] and Reis et al. [17], only these two materials emit between 53% and 73% of the total emissions during the materials and manufacturing phase. Studies show that aluminum alone constitutes almost half of the mass of the e-scooter [16,24], but its environmental impact accounts for 40% to 65% of the materials and manufacturing phase [12,20], taking into account that between 24% and 27% recycled aluminum is included in the aluminum mixture used for manufacturing e-scooters [13,16].

Lithium batteries are the second most polluting material in e-scooters, contributing between 10% and 19% of emissions during the materials and manufacturing phase [12,17,25]. However, according to a study by the German Energy Agency [20], first-generation e-scooters with permanently installed batteries have a greater environmental impact than second-generation e-scooters with swapable batteries, as they tend to have a longer lifespan. This finding is consistent with that of Kazmaier et al. [14], who claimed that switching to e-scooters with swappable batteries can lead to a 12% decrease in GWP.

The materials and manufacturing phase account for 41% of the total GWP for e-moped sharing services and 48% for e-bikes. However, e-mopeds have a greater weight compared to e-bikes, and the proportion of materials is different. De Bortoli [9] and Schelte et al. [21] estimated the weight of an e-moped to be around 100 kg. Zhu and Lu [8] reported a weight of 51 kg for a station-based e-bike, and Sun and Ertz [18] reported 20 kg for a dockless e-bike. Despite these weight differences, e-bikes use a higher proportion of aluminum (35%) and other metals and lead alloys (50%) [11] while e-mopeds use 25% aluminum and 37% of other metals [21].

Additionally, several reports have shown that the lifespan of Li-ion batteries differ between shared e-mopeds and shared e-bikes. Specifically, an e-moped requires only one battery replacement during its lifespan [9,11], whereas an e-bike requires battery replacement up to 2.42 times during its lifespan [11].

4.2.2. Rebalancing Phase

The collection and distribution process (rebalancing) of electric vehicles emerges as the second most impactful phase for the environment. The method of charging the batteries and the service distance is crucial for determining the actual impact. First-generation e-scooters, with fixed batteries, have low lifespans and high GWP values due to the need for frequent collection and distribution to charge their batteries at various locations [14,16,17,20,22,24,25]. In contrast, in newer generation e-scooters, the batteries are swappable, requiring only the transportation of the batteries themselves, while e-scooters need to be collected only for re-distribution and maintenance [14]. These e-scooters with interchangeable batteries tend to have longer lifespans, resulting in lower emissions per kilometer traveled [9,10,15,23]. Rebalancing activities have less impact if they are carried out with electric lorries [9,15]. This would allow batteries to be replaced on-site and allow the use of e-cargo bikes, eliminating the need to collect e-scooters for charging operations using large vans [43,44].

The use of e-vehicles and optimized routes decrease emissions during rebalancing activities. Both factors account for 9% of carbon emissions throughout the lifecycle of e-scooters, as demonstrated in a study conducted in Paris [9]. The greater the service distance, the higher the GWP. For example, service distances of e-scooters range from 0.02 km to 2.5 km; the GWPs of their rebalancing phase are 5.5 to 54 g CO₂eq/pkm [9,10,13,18,26].

Regarding shared e-bike services, emissions during the rebalancing phase in station-based systems are found to be less polluting compared to the dockless systems [8,11]. Similarly, the same holds for the distance service to collect and redistribute. For station-based systems, this distance is 0.012 km for every pkm travelled [11], while in dockless systems, the distance ranges between 0.14 and 0.167 km for every passenger kilometer travelled (pkt) [2,18,26].

E-moped sharing services also require rebalancing. These systems are typically dockless sharing systems, where empty batteries are swapped with charged ones on-site. With regard to this type of vehicle, there is no consistent consensus on the rebalancing phase. According to Schelte et al. [21] and Felipe-Falgas et al. [11], the rebalancing phase accounts for nearly half of their greenhouse-gas (GWP) emissions. On the contrary, de Bortoli [9] found that the rebalancing phase contributes only 7% of GWP emissions. The distance traveled by the rebalancing fleet in some studies was 0.02 km per pkt by an e-moped [9,11], while other studies provide figures of 0.083 km per pkt [21].

4.2.3. Usage Phase

The impacts on SEMMS during the use phase are mainly influenced by their lifetime mileage and the electricity power source generation mix used to charge vehicles. In the use phase, the environmental impact is significantly lower than that observed during the materials and manufacturing or rebalancing phases.

Regarding shared e-scooters, electric energy consumption ranges from 1.5 to 2 kWh/100 km, with an autonomy range between 20 and 30 km [10,12,13,16,18,20,22,24]. Doubling the weight of the e-scooters in the second generation does not significantly affect their energy consumption [9]. The effects of electricity emissions from charging are insignificant, as long as the generation of the electricity used had a low carbon footprint. For instance, a case study in Paris [9,10] showed a very low impact of the use phase compared to other studies reviewed on e-scooters, primarily due to the fact that France derives 70% of its energy mix from nuclear sources [45].

Despite the full use of renewable energy, the impact of the use phase remains minimal. Moller et al. [15] showed that the use of 100% renewable energy for charging in Paris resulted in a reduction of only 1 g CO₂eq./pkm of its total GWP. In terms of lifetime mileage, the results indicate that low mileage in the life cycle is closely associated with a shorter lifespan and higher GWP [18,22,25]. On the contrary, higher mileage is associated with low GWP [2,9,12]. Furthermore, a study reflecting the initial years of operation of shared e-scooters in Lisbon assumed a lifespan of 45 days and a daily usage of 2 km, resulting in a lifetime mileage of 90 km, with 803 g CO₂eq./pkm [17], the highest of all studies reviewed.

Although the intensity of usage fluctuates daily, mainly due to weather conditions and the specific day of the week [16], several studies have collected average distances per day. Moreau et al. [16] reported an average of 6.4 km/day, while Gebhardt et al. [12], Kazmaier et al. [14], Severengiz et al. [22], and Severengiz et al. [24] observed distances between 10 and 12 km/day. However, no significant relationship was found between the average distance per day and its respective GWP.

For shared e-mopeds, electricity consumption is the highest among all SEMMS. On average, it is 3.3 kWh/100 km, with a battery capacity ranging between 1.4 and 1.8 kWh [2,9,11,19,21]. Similarly, the actual range of e-mopeds falls between 60 and 90 km in terms of autonomy, while the distance traveled per day is estimated from 15–18 km according to Cazzola and Crist [2] and Schelte et al. [21]. And, it is estimated to be up to 36 km according to Wortman et al. [19]. E-mopeds have a substantially longer lifetime mileage compared to other micro-vehicles, ranging from 48,000 to 62,000 km [9,11,19,21]. However, according to Cazzola and Crist [2], the lifetime mileage of e-mopeds can be much lower, around 19,600 km.

On shared e-bikes, the lifetime mileage ranges from 5500 km [2] to 10,000 km [9] to 13,200 km [11]. The latter reported that the rate of replacement due to vandalism, theft, and misuse in this micro mode is approximately 20% annually. Two studies carried out in Chinese cities found that the average distance per trip was 2.5 km [18] and 3.06 km [8], resulting in a daily mileage of 24.5 km/day. However, according to Cazzola and Crist [2], the daily distance traveled was 8 km/day. Finally, the average energy consumption of e-bikes was determined to be 1.5 kWh/100 km [8,11,18], which is comparable to e-scooters and less than half that of e-mopeds.

4.2.4. Maintenance and End-of-Life Phase

The impact of vehicle lifespan and distance traveled on LCA is typically discussed in terms of manufacturing, usage, and rebalancing. Unfortunately, maintenance and end-of-life (EoL) stages are often overlooked or ignored, with few specific details provided or considered. In most cases, the EoL stage is encompassed within the manufacturing phase and maintenance within the rebalancing phase, i.e., Gebhardt et al. [12] and Zhu and Lu [8]. This exclusion is due to limited data availability and the limited lifespan of (i.e.,) e-scooters, which results in the omission of routine maintenance tasks, such as tire and part replacement, throughout the scooters’ lifespan [13]. Few studies have measured other phases, such as maintenance, end of life, or the infrastructure phase.

According to Moreau et al. [16], the lifespan of e-scooters is primarily influenced by four parameters: eco-design, usage, vandalism, and maintenance. Among these, maintenance plays a crucial role in extending the lifespan of e-scooters. For example, despite the relatively short lifespan of the initial generation of e-scooters, some of the earliest models have remained in service due to effective maintenance practices and the replacement of damaged components [16]. Nowadays, several operators have maintenance workshops where they can reuse an important percentage of the spare parts from the out-of-order e-scooter [10,15]. For instance, the study conducted by Reis et al. [17] in Lisbon showed that each e-scooter was maintained 2.76 times a month. The battery and/or electric motor was changed 0.68 times a month due to rough pavement conditions and steep slopes in the city, accounting for approximately 16% of the total GWP of e-scooters. In Paris, an e-scooter service operated by VOI [44], emphasizes that predictive maintenance software and dedicated local repair teams are crucial for achieving a longer lifespan [15]. In addition, the use of a modular architecture allows easy repairs. Sun and Ertz [18] calculated that the total GWP resulting from maintenance in e-bikes is 10%. De Bortoli [9] says that the maintenance of e-bikes contributes less than 10% to the total GWP, while this value increases to 23% for the case of e-mopeds.

The EoL phase is usually considered in the reviewed studies. However, most of them do not set a residual value at the end of the life of the micro-vehicles. Gebhardt et al. [12] assumed that the main metallic components are recovered, and the Li-ion batteries are recycled. According to Reis et al. [17], the end-of-life stage aims to mitigate 50% of the overall environmental impacts if the battery undergoes 10% recycling and the powertrain 83% recycling. Krauss et al. [26] assumed substantial recycling benefits of around 80% on e-bikes and e-scooters.

The most important findings of all the studies reviewed in LCA of SEMMS highlight the short lifespan of the vehicles, which contributes to a greater environmental impact, particularly in the case of shared e-scooters.

4.2.5. Summary of this Section

In essence, the most relevant aspects regarding the phases of LCA in SEMMS are vehicle material and manufacturing, as well as the rebalancing services. In the case of e-scooters, the material and manufacturing phase stands out as the most important contributor to pollution. This is mainly attributed to the significant use of aluminum alloy. E-mopeds were found to be the least polluting among all options, closely followed by e-bike services. However, it should be noted that the lack of representative data for phases of LCA other than materials and manufacturing, usage, and rebalancing in SEMMS, such as infrastructure, maintenance, and end-of-life (EoL), introduces uncertainty regarding the true impacts of SEMMS on greenhouse-gas emissions.

5. Sensitivity Analysis: Identification of Influential Factors

Sensitivity analysis in LCA allows for a deeper understanding of the factors driving the environmental impacts associated with a certain product or service. In this review, almost 90% of the studies carried out at least one sensitivity analysis.

Table A1. Summary of sensitivity analysis conducted on the reviewed literature based on shared micro-mode, influential factors, and different scenarios.

Author	Shared Mode	Influential Factor	Worst Scenario	Pessimistic Scenario	Base Scenario	Optimistic Scenario
			(g CO2eq./pkm)	(g CO2eq./pkm)	(g CO2eq./pkm)	(g CO2eq./pkm)
Zhu and Lu [8]	e-bike	Lifespan	___	___	(23)	10% longer lifespan (21)
Felipe-Falgas et al. [11]	e-bike	Lifespan	Lifespan halves (96)	Lifespan doubles (60)	10 years (56)	___
Felipe-Falgas et al. [11]	e-moped	Lifespan	___	Lifespan halves (116)	(76)	Lifespan doubles (69)
Moller et al. [15]	e-scooter	Lifespan	___	12 months (121)	24 months (70)	___
Severengiz et al. [24]	e-scooter	Lifespan	___	6 months (237)	24 months (77)	___
Gebhardt et al. [12]	e-scooter	Lifespan	___	6 months (320)	24 months (80)	___
Hollingsworth et al. [13]	e-scooter	Lifespan	___	0.5 years (280)	1 year (126)	2 years (89)
Moreau et al. [16]	e-scooter	Lifespan	___	1 month (829)	7.5 months (131)	2.5 years (51)
Sun and Ertz [18]	e-scooter	Lifespan	1 month (1400)	5 months (220)	12 months (159)	18 months (85)
Kazmaier et al. [14]	e-scooter	Lifespan	___	3 months (300)	6 months (165)	15 months (81)
Dena [20]	e-scooter	Lifespan	___	___	6 months (197)	24 months and improved servicing (59)
Chester [25]	e-scooter	Lifespan		Shorter lifespan (45 days) (420)	1200 km of mileage (197)	1000 charging cycle lifespan (50)
Reis et al. [17]	e-scooter	Lifespan	___	30 days (1156)	45 days (804)	180 days (276)
Felipe-Falgas et al. [11]	e-bike	Rebalancing	___	___	(56)	Sharing logistic reduced by half (56)
de Bortoli [9]	e-moped	Rebalancing	60 km service distance (39)	40 km service distance (36)	20 km service distance (34)	10 km service distance (32)
Schelte et al. [21]	e-moped	Rebalancing	___	___	(51)	Battery swapping with e-vans and solar (20)
Felipe-Falgas et al. [11]	e-moped	Rebalancing	___	___	(76)	Sharing logistic reduced by half (66)
de Bortoli [9]	e-scooter	Rebalancing	90 km service distance (81)	45 km service distance (68)	20 km service distance (61)	10 km service distance (58)
Severengiz et al. [24]	e-scooter	Rebalancing	___	Non-swappable battery (121)	Swappable battery (77)	___
Severengiz et al. [24]	e-scooter	Rebalancing	___	___	Diesel van (77)	Collection using e-vans (68)
Severengiz et al. [24]	e-scooter	Rebalancing	___	___	Diesel van (77)	Collection using e-cargo bikes (64)
Cazzola and Crist [2]	e-scooter	Rebalancing	___	___	(106)	50% lower service distance (64)
Cazzola and Crist [2]	e-scooter	Rebalancing	___	___	(106)	Low carbon servicing vehicles (82)
Cazzola and Crist [2]	e-scooter	Rebalancing	___	___	(106)	50% more vehicles for servicing (98)
De Bortoli and Christoforou [10]	e-scooter	Rebalancing	___	Van travelled 90 km with 50 scooters (250)	Van travelled 90 km with 100 scooters (109)	Car travelled 10 km with 100 batteries (38)
Hollingsworth et al. [13]	e-scooter	Rebalancing	___	___	4.6% fully charged battery collected (126)	No fully charged battery collected (102)
Hollingsworth et al. [13]	e-scooter	Rebalancing	___	4 km collection per scooter (160)	2.5 km collection per e-scooter (109)	1 km collection per scooter (92)
Moreau et al. [16]	e-scooter	Rebalancing	___	6.4 km between charges (146)	(131)	20 km between charges (125)
Moreau et al. [16]	e-scooter	Rebalancing	___	Distribution 50% longer (141)	(131)	Optimized distribution by 50% (121)
Kazmaier et al. [14]	e-scooter	Rebalancing	___	___	E-scooter collected by diesel car (165)	E-cargo bike for battery swapping (145)
Sun and Ertz [18]	e-bike	Lifetime mileage	1500 km (300)	3000 km (200)	5000 km (145)	10,000 km (109)
de Bortoli [9]	e-moped	Lifetime mileage	8000 km (150)	20,000 km (65)	48,000 km (34)	65,000 km (28)
Schelte et al. [21]	e-moped	Lifetime mileage	___	40% less mileage (58)	50,000 km (51)	___
de Bortoli [9]	e-scooter	Lifetime mileage	1000 km (360)	2000 km (190)	7300 km (61)	13,650 km (38)
De Bortoli and Christoforou [10]	e-scooter	Lifetime mileage	___	___	3750 km (109)	5200 km (30)
Severengiz et al. [22]	e-scooter	Lifetime mileage	2100 km (230)	3325 km (164)	4200 km (130)	4900 km (115)
Sun and Ertz [18]	e-scooter	Lifetime mileage	___	1000 km (300)	2000 km (159)	6000 km (80)
Cazzola and Crist [2]	e-scooter	Lifetime mileage	___	___	5700 km (107)	8550 km (76)
Cazzola and Crist [2] *	e-scooter	Lifetime mileage	___	1208 km (225)	2417 km (122)	3625 km (84)
Dena [20]	e-scooter	Lifetime mileage	___	1900 km (197)	3750 km (123)	7500 km (59)
Zhu and Lu [8]	e-bike	Daily mileage	___	___	(23)	10% more trip distance (21)
Schelte et al. [21]	e-moped	Daily mileage	___	___	18 km/day (51)	66% less battery swapping frequency (35)
Gebhardt et al. [12]	e-scooter	Daily mileage	___	4.4 km daily mileage (175)	10.2 km daily mileage (80)	13.7 km daily mileage (60)
Cazzola and Crist [2]	e-scooter	Daily mileage	___	___	(106)	50% higher daily mileage (76)
Hollingsworth et al. [13]	e-scooter	Daily mileage	___	4.2 km/day (305)	(126)	16 km/day (90)
Moreau et al. [16]	e-scooter	Daily mileage	___	1.2 km (593)	(131)	20 km (58)
Reis et al. [17]	e-scooter	Daily mileage	___	1 km (1600)	2 km (804)	5 km (112)
Wortmann et al. [19]	e-moped	Energy	___	___	German electricity mix (32)	100% renewables (19)
Schelte et al. [21]	e-moped	Energy	___	___	German electricity mix (51)	Solar power for charging (41)
de Bortoli [9]	e-moped	Electricity mix	China or Australia (75)	USA or UK (58)	France (34)	Norway or Denmark (32)
Severengiz et al. [23]	e-scooter	Energy	___	___	German electricity mix (55)	Recharging using renewable sources (41)
de Bortoli [9]	e-scooter	Electricity mix	China or Australia (90)	USA or UK (77)	France (61)	Norway or Denmark (60)
Severengiz et al. [24]	e-scooter	Energy	___	___	German electricity mix (77)	Charging batteries with solar power (66)
Gebhardt et al. [12]	e-scooter	Energy	___	___	German electricity mix (80)	Renewable electricity (70)
Cazzola and Crist [2]	e-scooter	Energy	___	___	(106)	Low carbon electricity use phase (100)
Moreau et al. [16]	e-scooter	Energy	___	___	(131)	Renewable electricity for batteries (127)
Kazmaier et al. [14]	e-scooter	Energy	___	___	German electricity mix (165)	Renewable electricity for batteries (147)
Zhu and Lu [8]	e-bike	Materials and manufacturing	___	___	(23)	10% less aluminum alloy (22)
Cazzola and Crist [2]	e-scooter	Materials and manufacturing	___	___	(106)	Low carbon aluminum smelting (82)
de Bortoli [9]	e-moped	Transport from China	USA (air) (68)	Europe (air) (63)	Europa (sea-road) (34)	Europa (sea-rail) (34)
de Bortoli [9]	e-scooter	Transport from China	USA (air) (97)	Europe (air) (92)	Europa (sea-road) (61)	Europa (sea-rail) (61)
Severengiz et al. [24]	e-scooter	Transport from China to Germany	___	Transport by plane (92)	Transport by ship and road (77)	___
Krauss et al. [26]	e-bike	End of life	___	20% losses due to vandalism or theft (85)	No theft or vandalism (68)	___
Moller et al. [15]	e-scooter	End of Life	___	___	(70)	High recycling rates and reuse (35)
Krauss et al. [26]	e-scooter	End of life	___	20% losses due to vandalism or theft (125)	No theft or vandalism (100)	___
Severengiz et al. [22]	e-scooter	EoL battery * second life—SL	___	Swappable and upgraded. No SL (134)	Not swappable without SL (130)	Swappable and upgraded with SL (116)
Reis et al. [17]	e-scooter	End of Life	End of life contributes 1% to climate change. (1680)	Recycling and materials recovery reduce impacts by 40%. (971)	Main materials are recycled (804)	___
Felipe-Falgas et al. [11]	e-bike	Modal change	___	50% walking and cycling of modal change (64)	(56)	50% car and motor bike depend city (23)
Felipe-Falgas et al. [11]	e-moped	Modal change	___	50% walking and cycling of modal change (94)	(76)	50% car and motor bike depend city (63)
Wortmann et al. [19]	e-moped	Fleet size	50,000 vehicles (37)	10,000 vehicles (33)	2500 vehicles (32)	___

* Cazzola and Crist [2] first-generation e-scooter.

in Appendix A compiles the range of GWP for shared electric micro-vehicles, considering different sensitivity scenarios, from the most unfavorable to the most optimistic one, and incorporates descriptions and data derived from the reviewed literature, providing crucial information for the analysis. The most influential factors considered in the sensitivity analysis are lifespan, lifetime mileage, rebalancing, energy used, materials, manufacturing and transport phase, and EoL phase.

The results of the sensitivity analysis of the reviewed studies proved that environmental impacts are highly sensitive to the lifespans of the vehicles. Figure 4a illustrates the relationship between the lifespan of shared e-scooters and GWP in both the baseline and sensitivity scenarios. For instance, extending the lifespan of shared e-scooters from one year to two years results in a significant reduction in the average life-cycle emissions from 30% to 50% [13,18,20]. Moreover, when the lifespan triples or quadruples from a baseline of 6 months, the emissions drop to a quarter of the base scenario [17]. On the contrary, Moller et al. [15] and Kazmaier et al. [14] suggest that, if the lifespan is halved, the emissions could double. Severengiz et al. [24] goes even further, proposing that emissions could triple in such cases. In extreme scenarios, emissions could increase fourfold if the lifespan is reduced to a fourth [12,18].

A comparison between optimistic and pessimistic scenarios was also conducted. In Lisbon, a base scenario with a 45-day lifespan results in an impact of 804 g CO₂eq./pkm for e-scooter services. However, extending the lifespan to 180 days reduces emissions by 80%. If the lifespan remains at 45 days, but with low daily use and no recycling at the end-of-life phase, the emissions could double [17]. Similarly, in e-mopeds, when the lifespan is halved, emissions increase by only a third; in contrast, doubling the lifespan slightly reduces emissions [11]. Opposite tendencies were observed for e-bikes when a 10% longer lifespan scenario was assessed. The emissions decreased by the same magnitude [8], but when it doubled, the lifespan did not have a positive impact [11]. It is worth noting that Felipe-Falgas et al. [11] and Zhu and Lu [8] are the only studies that evaluated lifespan scenarios for both e-bikes and e-mopeds; more in-depth studies of these two shared micro-modes are needed in the future.

Another important factor is the rebalancing phase. Almost half of the review studies conducted sensitivity scenarios for that phase. Small impacts are reported for the case of e-scooters if low carbon servicing vehicles are used, such as e-vans [2,24] or e-cargo bikes to swap batteries [14]; or if service distances are reduced [2,9,10,13]. Pessimistic scenarios were primarily assessed, considering longer service distances [9,13,16]. Furthermore, the use of conventional diesel vans or lorries for vehicle collection was evaluated [9], as well as the assumption that batteries would not be swapped in e-scooters [24]. In the case of e-bikes and e-mopeds, optimistic collection and distribution scenarios showed minimal effects on emissions reduction. This is due to the lower impact of the swappable batteries system on e-scooters and the station-based system on e-bikes [9,11]. When e-vans are used to re-balance and employ solar energy to charge batteries, emissions can be reduced by more than 50%. Alternatively, swapping batteries every 6 days instead of 2 days can reduce emissions by a third [21].

Lifetime mileage refers to the total kilometers traveled by the vehicles during their usage phase. Most studies report that the longer the lifetime mileage, the lower the emissions. Figure 4b shows the relationship between the lifetime mileage of the shared e-scooters and GWP in both the baseline and sensitivity scenarios. The pessimistic scenarios on e-scooters achieve total emissions above 150 g CO₂eq./pkm, with a lifetime mileage of less than 4000 km, while the optimistic scenarios record emissions of less than 100 g CO₂eq./pkm, with a lifetime mileage of more than 4000 km.

With respect to energy, sensitivity analyses were conducted to evaluate the impact of emissions reduction when renewable energy sources were used to charge the batteries of both e-mopeds and e-scooters. In Paris, the impact of using renewable energy sources on both e-mopeds and e-scooters was minimal [9], since approximately 70% of their electricity mix in Paris is derived from nuclear energy, with more than 10% coming from renewables [45]. In contrast, several studies conducted in Germany show emissions reductions ranging from 10% to 25% of total GWP when charging e-scooter batteries with 100% renewable energy [12,14,23,24]. Similarly, for e-mopeds, studies indicated reductions of 20% to 40% [19,21]. Other studies, such as Moreau et al. [16] for e-mopeds and Cazzola and Crist [2] for e-scooters, reported smaller impacts of 3% and 6% of the total GWP, respectively, when evaluating scenarios with 100% renewable energy. Furthermore, de Bortoli [9] conducted the only pessimistic scenario by assessing electricity mixes from different countries. The Norway mix had emissions as low as the baseline French scenario, while the USA or UK mixes had 25% higher emissions. China and Australia, relying mainly on coal generation, showed emissions twice that of the French baseline scenario. In particular, no energy-sensitivity scenario was performed for e-bikes.

The impact of transportation of micro-vehicles by shipping or rail modality between origin and destination remains unaffected [9], except when it is made by plane [17,20,24]. Finally, only two studies [2,8] conducted a sensitivity analysis on the materials and manufacturing phase in e-bikes and e-scooters, focusing on the impact of aluminum production.

End-of-life scenarios have been considered in four studies, including vandalism and theft, recycling, and second-life batteries. According to Krauss et al. [26], the losses due to theft or vandalism are approximately 20% in e-scooters and e-bikes. This results in a 25% increase in total emissions by reducing the lifespans of e-scooters and e-bikes. Moller et al. [15] mentioned that a significant portion of e-scooters in Paris have been recycled at the end of their useful life. Along with the reuse of parts for repair and maintenance, this achieves a 50% reduction in emissions. In contrast, Reis et al. [17] evaluated scenarios without recycling, which doubled the total emissions. Severengiz et al. [22] argue that e-scooters with interchangeable batteries are not sufficient to reduce emissions if the batteries do not have a second life.

To sum up, the most pessimistic scenarios are a result of decreasing the lifetime of the vehicles, whereas the more optimistic scenarios come up from increasing the lifetime of the micro-vehicles, the intensity of daily use per vehicle, and achieving a higher energy and logistic efficiency in the collection phase. There is a trend that indicates that the lifespan and rebalancing phases are crucial for improving the environmental indicators of e-scooters, while no preference or trend was found for performing sensitivity analyzes on e-mopeds or e-bikes.

6. Net Environmental Impacts Accounting for Mode Substitution

The implementation and use of SEMMS is believed to lead to a more sustainable mobility in cities. However, to gain a comprehensive understanding of the environmental impact of SEMMS, it is important to compare its LCA with alternative modes of transportation and to know whether the use of SEMMS replaces active modes or public transport with a low LCA or not. Some studies in the current literature review have addressed this question. First, by comparing the GWP per passenger by mode of transportation, that is, [2,8,9,10,11,15,20,22,23,24,25,26]. Second, by analyzing the effect on GWP caused by mode substitution, that is, [8,11,12,14,16,18,19,21,23].

Regarding the impacts of mode substitution, the use of SEMMS can contribute to improving the environmental impact of the whole transport system by substituting transport modes with a higher GWP, such as private cars. However, if other substitution effects occur, for example, substitutions for public transport or cycling, the opposite can occur [16,18,22].

Research studies on shared e-scooters in Europe and the United States indicate that a too-short lifespan can result in the use of shared e-scooters to generate more CO₂ emissions than the mode of transportation they replace [10,13,16]. In fact, in the case of Paris, under the most optimistic scenario that excludes servicing impacts and assumes longer but attainable lifetime mileages, the carbon footprint of e-scooters could potentially range from 30 to 12 g CO₂eq./pkm. It is important to note that, even at its lowest estimate, this carbon footprint remains higher than that of the metro, RER (Réseau Express Régional), and walking in Paris, which emit 7.6, 8.9, and 1.9 g CO₂eq./pkm, respectively [10]. However, e-scooters have the potential to reduce GHG emissions in Parisian mobility by substituting trips that would have been made using less environmentally friendly modes, such as buses, private cars, taxis, and ride hailing only if e-scooters emit less than 56 g CO₂eq./pkm.

For Germany, Kazmaier et al. [14] conducted a study revealing that shared e-scooter services generate a greater amount of CO₂eq./pkm compared to the modes of transportation they typically substitute. Such research demonstrates that the GWP savings resulting from the modal shift prompted by e-scooters are significantly lower than the emissions produced by the e-scooters themselves (165 g CO₂eq./pkm). Even in the most favorable scenario (46 g CO₂eq./pkm), the emissions from e-scooters are still 8% higher than the average emissions per kilometer of the modes of mobility being replaced by e-scooters (39 g CO₂eq./pkm).

Other research carried out in Brussels by Moreau et al. [16] conducted a comparative study between personal and shared e-scooters, in order to assess their environmental impact. The study reveals that, while shared dockless e-scooters emit 21 g CO₂eq./pkm more than the modes of transportation they replace, personal e-scooters emit 50 g CO₂eq./pkm less, mainly due to their longer lifespan and no need for van collection and deployment. Furthermore, the study highlights that personal e-scooters have a 6% higher GWP than shared e-scooters, suggesting personal electric micro-vehicles are more sustainable than SEMMS, despite the comparable overall impacts.

In cities with low population density, limited public transportation options, or high carbon intensity, shared e-scooter services can indeed have a positive environmental impact [10]. To illustrate this, in Raleigh, USA, a survey of e-scooter riders found that 7% of them would have skipped the trip, 49% would have walked or biked, 34% would have used a personal automobile or ride-share, and 11% would chosen public buses [13].

Hollingsworth et al. [13] also found that the adoption of e-scooters led to a reduction of 94 g CO₂eq./pkm in life-cycle emissions, a 26% decrease compared to the typical shared e-scooter use, aligning with emissions in other scenarios (e.g., high scooter lifespan, low collection distance).

Shared e-mopeds have the potential to reduce road transport emissions if they replace car trips instead of active modes [11,19], as shared e-mopeds have a lower GWP compared to shared e-scooters [21]. When considering substitution rates, e-mopeds primarily impact trips that were otherwise made by private cars, conventional scooters, or public transport, accounting for approximately 81% of cases. This implies that e-moped sharing is most likely to replace transportation modes with a similar or higher GWP in terms of passenger kilometers. In Barcelona, shared e-scooters emit more CO₂eq./pkm than shared e-bikes. However, e-mopeds have a lower impact due to the modal change they prompt, as users previously contributed to greater pollution [11]. In Berlin, the simulation results show that almost 2% of passenger car trips can be substituted, reducing emissions from road transport [19].

Regarding shared bikes, only station-based bike services prove a reduction in GHG emissions, since electric and mechanical dockless bike services have failed to achieve the anticipated decrease in GHG emissions [8,18]. Only station-based mechanical bike services have demonstrated notable success in reducing GHG emissions by approximately 32.25 g CO₂eq./pkm. Conversely, opting for dockless bike services may lead to an increase in emissions of around 48.47 g CO₂eq./pkm, and employing dockless e-bike services may result in an increase of roughly 19.46 g CO₂eq./pkm. These findings are consistent with the results of a study by Zhu and Lu [8] on station-based bike services, which indicated a reduction of 39 g CO₂eq./pkm in GHG emissions.

7. Policy Recommendation

In this section, we summarize a set of proposed strategies found in the present literature review aimed at mitigating the environmental impacts of SEMMS. Through the analyses carried out in different case studies in various cities and regions of the world, the following recommendations are discussed for transferability to other cities or contexts. The recommendation on policies and practices revolves around the key factors identified above in Section 4, Section 5 and Section 6: lifespan and vandalism, materials and manufacturing, rebalancing, mileage, electricity mix, modal shift, and integration with public transport.

7.1. Lifespan and Vandalism

Many research studies reach the conclusion that longer lifespans of the vehicles used in SEMMS could lead to a reduction in the carbon footprint during their entire life cycle and, thus, can be done with the cooperation of all stakeholders. On the one hand, it is recommended that vehicle manufacturers strive to design and produce long-lasting vehicles [16] and strengthen their cooperation with operators and recycling organizations [18]. On the other hand, operators can prioritize scooter repairs over premature end-of-life scenarios, as this will contribute to a rapid reduction in emissions per mile [25]. Additionally, operators should explore the utilization of batteries beyond the lifespan of micro-vehicles to minimize the share of battery production within their GWP. Manufacturers should also adapt production materials and micro-vehicle designs to improve their lifetime [21]. It is crucial to wait until the end of an e-scooter’s lifetime before replacing all of them with new models. It would also be possible to extend their lifetime after their service by selling them to private owners [24]. Furthermore, city authorities could enforce antivandalism policies to reduce e-scooter mistreatment, which can result in short lifetimes [13,16,17,24]. Considering the short lifespan of e-scooters, it is recommended to take into account the characteristics of the infrastructure (pavement material, slopes, etc.) of each city to ensure an optimal environmental design of the SEMMS [17].

7.2. Materials and Manufacturing

Regarding the materials and manufacturing phase, many studies recommended that the vehicle manufacturer should use “green” aluminum, with a high recycling rate in its production and a lower proportion of aluminum components in order to decrease the GWP of e-scooters [24] and e-mopeds [21]. The use of alternative materials with better environmental balance and a life-cycle-oriented construction can also make the production of e-scooters more climate friendly [9]. Moreover, the implementation of closed-loop recycling measures between fleet operators or vehicle manufacturers and raw-material or recycling companies can improve the climate impacts of e-scooter production [20].

Another important point is the location of production; the reviewed studies recommend moving vehicle manufacturing to a location that reduces production and transport emissions, where the energy mix of the industrial sector has more low-carbon or carbon-neutral generation sources incorporated [20,24,25], as well as using renewable energy in the production of electric vehicles and batteries [21].

7.3. Rebalancing

Alternative approaches to collecting and distributing bicycles and e-scooters or swapping batteries on e-mopeds and e-scooters can greatly reduce adverse environmental impacts [16,18]. For instance, limiting e-scooter collection to those with a low battery state of charge could result in a net reduction of the global warming impacts [13]. In addition, the use of low-carbon vehicles (e-cargo bikes or e-vans) to collect shared vehicles or to exchange batteries is strongly recommended [13,17,21,24], along with the installation of docking equipment in high-density use areas that can charge e-scooters with solar power [25].

Furthermore, reducing the average driving distance for collection and distribution reduces the impacts of global warming by e-scooters more than the exclusive use of fuel-efficient vehicles for collection [13]. To mitigate the environmental impacts, operators should either reduce the frequency of battery swapping for e-mopeds and e-scooters [17,21] or deploy decentralized battery-swapping stations [23].

7.4. Mileage

To maximize the efficiency of shared micro-vehicles and minimize carbon emissions per passenger kilometer, it is crucial to maximize passenger use by increasing daily trips and the distance covered [17,26] and constructing reasonably-sized fleets by authorities and operators [18].

7.5. Electric Power Generation Mix

In Brussels, authorities recommend that e-scooter providers opt for renewable electricity for charging [16]. However, the use of renewable electricity for charging does not have a substantial influence on the overall environmental impact of dockless e-scooters in cities with low carbon intensity, such as France [10]. However, it is expected that the importance of using renewable electricity grows as the lifespan of e-scooters increases [16].

7.6. Modal Shift and Integration with Public Transport

Shared micro-mobility services, such as electric scooters and bicycles, have the potential to replace cars and taxis if supported by appropriate infrastructure. By establishing mobility hubs and intermodal travel-planning tools, cities can seamlessly integrate these services into their public transportation strategies, contributing to low-carbon initiatives. Krauss et al. [26] suggest three strategies to encourage a shift away from cars: offering accessible shared micro-mobility options, incentivizing taxi and ride-hailing users to choose micro-mobility for short trips, and integrating micro-mobility services with public transit services. Adequate parking facilities and well-designed bike lanes are also important to facilitate the transition to alternative transportation modes.

According to Moller et al. [15], experts predict a dense network of shared mobility options, such as electric scooters and automated shuttles, which improves access to public transport and alternative mobility networks. In their studies, Severengiz et al. [23,24] suggest that the integration of e-scooter sharing with local public transport can position e-scooters as a complement to public transport rather than a replacement. They also emphasize the importance of promoting public transport, walking, and cycling alongside the introduction of e-scooter sharing to ensure the substitution of less environmentally friendly modes of transportation.

Reis et al. [17] propose that cities expand cycle lane networks to include e-scooters, improving comfort and safety conditions for active transportation modes. Additionally, Kazmaier et al. [14] highlight the need for collaboration between e-scooter providers and public transport authorities to develop a comprehensive framework that allows the complementary nature of these transportation options. Policy decisions should focus on creating a safer and connected micro-mobility infrastructure, which can improve the acceptance and use of, for example, e-scooters [12,15].

Finally, Moller et al. [15] propose that governments should establish national regulations for micro-mobility that encourage cities to allocate more space for it. The authors also emphasize the importance of data collaboration and learning from micro-mobility providers. They advocate for cities to choose operators with responsible employment practices and sustainable operations. SEMMSs have the potential to reduce the carbon footprint of urban mobility, but careful deployment and regulations tailored to the characteristics of each city are necessary [10]. Limiting the business area of sharing e-scooters to the inner city increases the density and reduces battery/scooter collection distances [24].

8. Conclusions and Further Research

Overall, this study provides an understanding of the environmental impacts of shared electric micro-mobility services using the LCA perspective. It can assist transportation planners and decision makers in developing policies and regulations governing SEMMS for the development of sustainable urban transportation systems and foster the adoption of environmentally responsible practices in the rapidly expanding field of SEMMS. The scope of our review is to better understand the potential sustainability consequences of SEMMS, particularly in relation to the reduction of the carbon footprint of urban mobility. Through this review, we have identified limitations and gaps that could be addressed in further research.

For data availability, one of the primary challenges in conducting an LCA for shared electric micro-mobility services is the lack of comprehensive and reliable data. Obtaining data on the manufacturing, operation, maintenance, and end-of-life stages of shared micro-mobility vehicles and infrastructure can be challenging due to the involvement of multiple stakeholders (different operators, public authorities, etc.), sub-activities or combinations of activities, and limited data transparency. Studies are needed to elaborate further on data requirements and mechanisms to share data.

For a standardized methodology, there is a lack of standardized methodologies for conducting LCAs specifically tailored to shared electric micro-mobility services. Existing LCA frameworks often focus on conventional transportation modes or broader vehicle classes, which may not capture the unique characteristics and usage patterns of micro-mobility services. Some studies have different units, making it impossible to compare them directly (e.g., g CO₂eq. and g CO₂eq./pkm). Furthermore, most existing studies focus on e-scooters, while there is a scarcity of studies on shared e-mopeds and e-bike services. Such a standardized methodology could serve to compare differences between countries and cities.

For user behavior and travel patterns, understanding user behavior and travel patterns is crucial for accurately assessing the environmental impacts of shared micro-mobility services. Factors such as trip frequency, daily mileage, vehicle occupancy rates (for e-mopeds), mode substitution, and charging patterns significantly influence the overall energy consumption and emissions associated with these services. It is also important to know the impact that the generalization of micro-mobility may have on private car ownership in the medium- and long-term.

For a dynamic charging infrastructure, SEMMSs rely on a charging infrastructure to maintain the operational availability of vehicles. However, the impact of a charging infrastructure on the overall LCA is often overlooked. Evaluating the environmental implications of charging-infrastructure deployment, including energy sources, grid integration, and infrastructure materials, is necessary to provide a comprehensive assessment in the use and rebalancing phases.

For end-of-life considerations, the LCA of shared electric micro-mobility services should consider the end-of-life phase, including recycling, disposal, and material recovery processes. Understanding the environmental consequences of, for example, the disposal of lithium-ion batteries is crucial for assessing the sustainability of these services. However, there is a lack of comprehensive data and studies on the end-of-life phase of micro-mobility vehicles, their batteries, and the infrastructure they use.

For system boundaries and scope, defining the boundaries of the system and the scope of the LCA is another critical challenge. Shared micro-mobility services often interact with existing transportation systems, public transportation networks, and active modes, such as walking and cycling. Evaluating the broader impacts and potential mode shifts associated with these services requires a careful consideration of the system boundaries, which can be complex and context dependent.

For external factors, the LCA of shared micro-mobility services often focuses on the direct environmental impacts associated with vehicle manufacturing, operation, and disposal. However, indirect impacts, such as changes in land use, infrastructure requirements, and overall travel behavior, should also be considered. These external factors can significantly influence the overall sustainability and benefits of shared micro-mobility services.

In conclusion, efforts should be made to make electric micro-vehicles more environmentally friendly through longer lifespans, recycling, anti-vandalism policies, battery reuse, and eco-friendly materials. It recommends integrating micro-mobility with public transportation, using renewable electricity, and creating a safer infrastructure. Collaboration among stakeholders and continued research on their impact on public transport are essential for sustainability.