Last summer, I wrote about the emergent dockless electric scooter programs in cities across the United States, taking issue with the fact that these programs that seemed to be popping up left and right were purporting to offer clean transportation solutions. Vehicles powered by electricity rather than gasoline are indeed increasingly eliminating transportation-related greenhouse gas (GHG) emissions, which these e-scooter programs use to claim that trips on these electric scooters reduce emissions and make them inherently environmentally-friendly mobility options.
While replacing gasoline-powered trips with electric-powered trips do lead to decreased carbon emissions, not to mention the efficiency of moving a human using something small like a scooter as opposed to a car that needs to transport its own massive weight, the main reason I didn’t buy these companies’ claims about sustainability is because charging isn’t the only source of emissions associated with the use of these dockless scooters. Every night, these scooters are rounded up and collected by employees and freelancers (often by car, truck, or van) to be charged overnight and then redistributed in locations with expected demand for them the next morning. If the emissions of these gasoline-powered vehicles collecting & redistributing the scooters were taken into account, I hypothesized, then the dockless electric scooter programs would not be considered the green option that companies suggest they are. In fact, given that many trips taken on these scooters are replacing virtually carbon-neutral choices (walking, biking, or even public transit), the scooters could in fact be increasing total transportation emissions in a city.
Source: Fortune
After publishing my analysis that I titled The Electric Scooter Fallacy: Just Because They’re Electric Doesn’t Mean They’re Green, it seemed to strike a nerve and generate great discussion across related industries. I haven’t found another analysis looking into this idea in a more scientific or comprehensive manner than my self-described back-of-the-envelope calculations, so the article continued to drive some small amount of attention and debate. One weakness to my analysis that was pointed out to me, though, was that I was limiting the scope of my assessment (just as I had accused the electric scooter companies of doing) because I failed to take into account the entire life cycle of the scooters from manufacture to disposal. The idea with this criticism is that because so much less material and production goes into making a scooter compared with a car, the total embedded emissions per mile might come out much more in favor of the dockless electric scooter programs– even with the daily collection & redistribution.
This point was a compelling one, so I finally decided to revisit my sustainability analysis of dockless electric scooter programs to see how life-cycle considerations would affect the final conclusion. Once again, though, the results aren’t favorable towards electric scooters– but all is not lost. These companies can implement strategies that will tilt the scales to end up helping in the fight against climate change, but before we get into those solutions let’s run the numbers!
Previous Analysis
I don’t want to completely rehash the entire analysis that went into my previous article, so if you want to see the full details then please head to the first article and read through that first.
But as an overview, the steps taken and the conclusions drawn can be boiled down to the following:
Using a Swagtron SWAGGER High Speed Adult Scooter (with a 91.2 watthour battery and 15-mile range) as the basis for dockless electric scooters and the carbon intensity of the Washington DC energy grid to charge the scooters (0.622 grams of CO2 per watthour), along with an assumption (based on statements from Bird) that the average electric scooter is ridden five times per day with an average ride distance of 1.5 miles, I calculated that charging/riding the scooter accounts for 28.4 grams of CO2 emitted per day or 3.8 grams of CO2 per mile ridden.
However, those emissions were only part of the equation. To determine the emissions from collecting & redistributing the scooters via gasoline-powered vehicles, I needed information on the scooter collection process: how many scooters are typically collected in one trip and how far are these collection trips? The companies are notoriously private about this data, so I offered a range of assumptions:
Bird caps the number of scooters a single contractor charger can pick up each night at 20, but the competition to get to scooters quickly is stiff and could limit people to as few as 5 scooters to charge each night– so I modeled collection & redistribution in groups of 5, 10, and 20 scooters.
No data points are available on the distance traveled to collect the scooters, but based on their typically urban nature I modeled short (2 miles), medium (5 miles), and long (10 miles) total round-trip mileage for the collection & redistribution process.
Using these crude estimates, the calculations found that the collection & redistribution process in fact did account for a majority of the emissions of a dockless electric scooter program in each case. The best-case scenario (2 miles, 20 scooters) saw a total of 69 grams of CO2 per scooter per day (9.2 grams per scooter-mile) and the worst-case scenario (5 scooters, 10 miles) saw 606 grams of CO2 per scooter per day (80.8 grams per scooter-mile) compared with 404 grams of CO2 per mile driven by an average car (or taxi/Uber).
While these results appear at first glance to support the claims that these scooters are cleaning up transportation, the other critical point was that these results only look positive when assuming a one-to-one replacement of car-miles for scooter-miles. As stated in the first article:
In the most-efficient case, where each scooter accounts for 69 grams of CO2 per day compared with 3,030 grams of CO2 if the scooter’s 7.5 miles per day were replaced by car, as long as 1 scooter trip among every 44 trips are actually being used as an alternative to taking a car (say, hopping on an electric scooter instead of taking an Uber to get downtown), then the scooters are making a positive environmental impact. This proportion of trips actually displacing cars seems feasible. However, in the least-efficient case, each scooter accounts for 836 grams of CO2 per day, meaning that if less than 27% of the scooter trips are actually replacing car trips then the scooters are actually generating more CO2 emissions than the cars would otherwise. Anecdotally from walking around downtown DC, I’m hard-pressed to believe that more than one in four of the scooter-riders I pass would otherwise be taking a car wherever they’re going (none of this is even to mention how pooling multiple scooter riders into a single car or Uber would tip the scales even further).
So, the main point was: yes, electric scooters seem to be a sustainable addition to the urban landscape, but without data on how efficiently the scooters are collected & redistributed and what portion of rides are directly replacing car trips instead of replacing walking/biking/public transit, then concluding they are definitely more eco-friendly is at best incomplete and misleading.
Adding in Life-Cycle Considerations
To make that analysis more robust, though, let’s see what happens when we account for the full life cycle of the scooter, from manufacture to distribution to use to disposal. The calculations for this analysis are captured in this Google spreadsheet, meaning you can download a version and adjust the inputs to determine how the final results would change (useful if you disagree with any of my assumptions or inputs).
Scooter Specs
To start, a recall of defective scooters from China inadvertently revealed the manufacture and model of scooters that are commonly used across dockless electric scooter programs: the Mi Electric Scooter made by Xiaomi’s subsidiary Ninebot in Changzhou, China. According to the official specs, this scooter weighs 26.9 pounds, has a frame made of aerospace-grad aluminum, uses 8.5-inch rubber tires, and is powered by a total of 30 lithium-ion batteries known as 18650 batteries with a cumulative capacity of 335 watthours.
Source: ZDNet
Researching these component parts finds that the total battery weight would be approximately 2.98 pounds and the combined weight of the two tires would be 2.20 pounds, and we can assume the remaining 21.72 pounds are all aluminum (of course, there are other electrical and mechanical components, but this assumption will give us a reasonable approximation to use in the life-cycle analysis).
Another important life-cycle specification is estimating the scooter’s lifetime. Each scooter will vary in its individual use and performance, and life-time durability is another area where the dockless scooter companies keep the data close to the vest. Typical electric scooters (outside of scooter-sharing programs) have been estimated to last for 300 to 500 lifetime rides, 500 charging cycles, or 1,000 charging cycles. Dockless scooter companies perform regular maintenance to keep them running, but according to the Washington Post:
Scooters that reach their expiration date after being worn down merely by inclement weather, overuse, and other hazardous potholes are the lucky ones. Many others can expect their final moments to be undeniably barbaric and marked by vandalism and destruction. Based on the rougher treatment dockless scooters receive, it seems reasonable to estimate the lifetime of these scooters is on the lower end of an average expected lifetime.
Source: Washington Post
So, this life-cycle analysis will use 500 total lifetime rides (meaning an average of 750 total scooter-miles and 100 overnight charges). This assumption is a good one to play with in the Google spreadsheet if you don’t like my choice. Note, though, that I did share my preliminary results with an industry insider who told me that anecdotally the lifetime for dockless scooters is very low and the typical approximation used within the industry is an average of 45 days, which can drop to as few as 23 days in particularly ‘contentious’ markets. In the Google spreadsheet and at the end of this article, I’ll highlight some side cases that include these shorter lifetimes.
Manufacturing
The first step to the life-cycle analysis is estimating the embedded GHG emissions that went into the manufacture of the component parts– our 2.98 pounds of lithium-ion battery, 2.20 pounds of rubber tire, and 21.72 pounds of aluminum.
While researching how to approach this part, I discovered the Greenhouse gases, Regulated Emissions, and Energy use in Transportation (GREET) Model from Argonne National Laboratory. The GREET Model “is a one-of-a-kind analytical tool that simulates the energy use and emission outputs of various vehicle and fuel components,” advertising its ability to assess the full life-cycle, from raw material mining to vehicle disposal. Despite how intimidating this model appears to be, I reached out to the fine people behind the GREET Model who agreed to walk me through what I needed for my analysis over the phone (side note: there’s a reason those working in the public sector, like the Department of Energy’s National Laboratory System, are called civil servants– I could not have done this on my own and a heartfelt thank you to Argonne National Lab and specifically Qiang Dai for being so willing and available to help me with this).
After inputting the materials into my modified GREET Model, the manufacture of the scooter battery accounts for 19,824 grams of CO2-equivalent (CO2e) of GHG emissions and the scooter body and tires account for 184,247 grams of CO2e, for a total manufacturing emissions of 204,071 grams of CO2e. Spread out over the 750 miles of the scooter’s lifetime, these emissions come out to 272.1 grams of CO2e per lifetime scooter-mile.
Note that this calculation isn’t even including the manufacture and use of spare parts used to repair scooters as they break down. In addition, some scooters can be heavier and use a bigger battery than our baseline model, such as the Segway ES4 Kick Scooter. Per both of these caveats, the actual numbers for manufacturing could even be slightly higher.
Assembly
The next step is assembling the manufactured parts in the factory. According to the GREET Model using the same inputs as before, this step accounts for 5,431 grams of CO2e or 7.2 grams of CO2 per lifetime scooter-mile.
Transport
Another notable aspect of the scooters used across all dockless electric scooter programs is that they are made in China. Not only does that mean the energy mix powering the manufacturing process is more fossil-fuel heavy (for which the GREET Model accounted) than it would be if made elsewhere, but the completed scooters must then be shipped to the United States. By adjusting the appropriate parameters of the GREET Model, the total emissions per scooter estimated for transport (via a combination of truck and barge) come out to 2,796 grams of CO2e or 3.7 grams of CO2e per lifetime scooter-mile.
Charging, Collection & Redistribution
The emissions associated with the operation and use of the electric scooters in these dockless scooter programs were discussed above and in the previous analysis, but since we have adjusted the scooter model used they’re worth quickly revisiting.
For charging one of these scooters in Washington DC, it takes 0.335 kilowatthours (kWh) of energy to fully charge the battery that can then travel 18.6 miles. If we assume the battery is fully depleted each night (the full-charge range extends beyond the 7.5 miles the average scooter travels, but these scooters also come with additional electrical components that drain the battery further and we can assume most people leave them plugged in overnight even if they’re fully charged after just several hours) and model the recharging process with the Washington DC energy mix that emits 0.62 grams of CO2e per watthour, a scooter will account for 208.3 grams of CO2e for each overnight charge, 20,831.4 grams of CO2e over the course of its useful life, and 27.8 grams of CO2e per lifetime scooter-mile.
If we’re then using the same assumptions for nightly collection & redistribution via car, then the best-case scenario (2 mile trip to collect/redistribute 20 scooters) adds 5.4 grams of CO2e per scooter-mile, the medium case (5 mile trip to collect/redistribute 10 scooters) adds 26.9 grams of CO2e per scooter-mile, and the worst case (10 mile trip to collect/redistribute 5 scooters) adds 107.7 grams of CO2e per scooter-mile.
Given that the data for collection & redistribution is still scarce-to-nonexistent publicly, this part of the analysis is the hardest to validate. You can adjust these parameters in the Google spreadsheet for yourself to see how it affects the analysis, and in fact in the time since I first looked into these numbers some reports have come out to suggest my ‘worst-case’ scenario may not have been nearly bad enough:
Nathan Griffith of Austin, Texas, said he drives three to four hours each night in his Chevy Silverado. “I’m busting a- -, like if I really want to make money, it’ll be four to six hours of driving. It’s all stop and go, too.”
Lindsay Miller of Chandler, Ariz., said she and her boyfriend drive 30 to 40 miles each night to hunt for Bird scooters. “We try to collect dozens of scooters a night. We have a Kia Optima. But typically, we use our truck more on the weekends because we can fit more.”
These types of collectors are likely the outliers pushing the average towards the ‘worst-case’ scenario, but the urban collectors who mirror the ‘best-case’ scenario are also real:
To be sure, other chargers said they look for scooters on foot or only drive short distances. “Luckily for me, I live in an area with lots of scooters, so I don’t drive very far,” said Carlos Bates of Hollywood, Calif. “I drive in a 5- or 10-mile radius. I would never drive more than 10 miles looking for scooters. I mean, they’re everywhere here in Melrose and Hollywood.
Disposal
Returning to the GREET Model for the end of the scooter’s life, the average disposal/recycling emissions come out to 2,978 grams of CO2e or 4.0 grams of CO2e per lifetime scooter-mile.
Adding Up the Life Cycle
Assembling all these figures together gives the following results: a dockless electric scooter, over the course of its 500 lifetime rides and 750 lifetime miles, can be expected to contribute between 240.1 kg of CO2e (using the best-case collection & redistribution) and 557.1 kg of CO2e (using the worst case).
Divided over the 750 lifetime scooter-miles, each dockless electric scooter can be approximated as contributing 320.2 to 742.7 grams of CO2e per scooter-mile to the atmosphere.
How Does That Stack Up?
My idea for using the GREET Model in the life-cycle emissions analysis came after reading a similar analysis from the Union of Concerned Scientists (UCUSA) to compare the life-cycle emissions per mile of electric cars compared with gasoline-powered cars. We can simply look at their results and compare with the electric scooter results to determine how effective or ineffective the dockless scooter programs are at cleaning up carbon emissions.
According to the UCUSA data, on a per-mile basis over using a similar life-cycle analysis:
A midsize gasoline car emits 424 grams of CO2e;
A full-size gasoline car emits 577 grams of CO2e;
A midsize 84-mile electric vehicle (EV) emits 206 grams of CO2e; and
A full-size 265-mile EV emits 271 grams of CO2e.
Note that the UCUSA study was conducted in 2015, meaning both EVs and gasoline-powered cars have only gotten somewhat cleaner and more efficient in the years since.
What becomes evident when analyzing the life-cycle emissions of a dockless electric scooter compared with alternative car options is that any use of an EV outperforms even the best-case scenario for the scooters on a per-mile basis. Even more striking, removing any collection & redistribution emissions still allows electric cars to win.
Even though the initial glance might appear to show that dockless electric scooters, on a per-mile basis, are cleaner than gasoline-powered cars, that may not be the case in practice. As noted in the original article:
It would be incomplete to declare that the dockless electric scooters are eliminating CO2 emissions from the transportation sector…It goes back to the idea that this calculation assumes a one-for-one replacement of car trips with scooter trips. How often is this actually the case? Scooters could be, and likely are, replacing options that are less carbon-intensive (biking, walking, public transit) just as often as they are replacing the more carbon-intensive act of driving.
So, even in the narrow cases where dockless scooters are less carbon-intensive than driving or riding in a taxi/Uber, that advantage dries up when even just half of scooter-miles are not direct replacements for cars. The first instance of publicly available data that tracked whether scooter rides took place of actual car-miles came from Portland’s recent survey of their dockless scooter program, which found that scooters were used in place of a personal car ride or ride-sharing app 34% of the time for residents and 48% of the time for visitors to the city. If that data is verified elsewhere and applies widely, it’s not a good look for the so-called sustainability of these dockless scooter programs.
The dockless electric scooters end up looking worse from an environment and climate standpoint compared with cars than they did in my first article, which can be attributed to the vast difference in lifetime miles (i.e., while a scooter creates much fewer emissions in manufacturing than a car, a scooter lasting less than 1,000 miles and a car well over 100,000 miles means the per-mile emissions for cars remain competitive). Such difference in lifetime miles is why this counter-intuitive result arises– when you look at the total emissions version of the above chart, the scooters are barely visible.
As these graphs make clear, it’s the short lifetime of the scooters that make them heavy in per-mile emissions. To further drive home the importance of the lifetime of the scooter, I’ve included in the Google sheet some alternative scenarios: assuming scooters last 1,000 charge cycles and/or carbon-neutral collection & redistribution, as well as assuming a 45-day lifetime and/or double the worst collection & redistribution arrangement. But even in those best-case scenarios, it appears likely that the proliferation of dockless electric scooter programs– as they exist now– offer a tenuous benefit to aggregate transportation emissions depending on how frequently they do serve as one-to-one replacements for driving, but at worst they end up accounting for a greater overall emissions per mile.
What Can Be Done to Reduce Emissions of Dockless Scooter Programs?
All of this analysis is not to just dismiss the benefits dockless scooter programs can provide, such as increasing mobility in low-income areas. The goal here is to simply point out that the companies claiming dockless scooters are providing improvements to transportation when it comes to sustainability and the climate without actually delivering those benefits is greenwashing at its core. But that’s also not to ignore the potential about how these programs can improve transportation emissions. When looking at the side cases, we should strive to get them as close as possible to the best side case as possible.
The manufacture of the scooter materials, the charging, and the collection & redistribution all account for a large portion of the emissions. To truly embrace sustainability, dockless scooter programs can work on the following:
Bring the manufacturing of the scooters to the United States, where the industrial sector’s energy mix has more low-carbon or carbon-neutral generation sources built in, which would also greatly reduce the small portion of emissions that go to transporting the scooters across the world.
Encourage more low-emissions options for collection & redistribution of the scooters. Such goals can be accomplished by:
Paying more for charging contractors who use EVs to pick up scooters;
Paying more for charging contractors who can plug into rooftop solar installations at their home or office;
Paying more for charging contractors who charge the scooters at a closer distance to where they are picked up and/or dropped off; or
Installing in heavy-use areas docking equipment that can recharge scooters with solar power.
Source: Swiftmile
Get creative with recharging solutions– instead of transporting scooters to be recharged each night, why not have people travel (by bike or EV) to replace dead batteries with fully-charged ones, similar to proposed battery-swapping stations on highways for EVs.
As the side cases showed, the per-mile results are quite sensitive to changes in how long a scooter’s lifetime extends. While industry information shows that 45 days might be an average lifespan of a dockless scooter, companies can put more effort into repairing scooters as opposed to letting them reach a premature end of life, which will quickly result in per-mile emissions dropping.
Conclusion
If the transportation sector is going to truly embrace action called for in the latest IPCC report on climate change, then highly-disruptive technologies to upend the entirely car-centric model are necessary. Dockless electric scooter companies envision being a part of such a solution, but they need to operate smartly with real energy and emissions management to realize the changes they’re capable of making.
While Lime advertises that its scooter rides are now carbon neutral through funds provided to outside carbon reduction projects, such actions are only partially helpful and do not actually reduce their direct emissions (not to mention these offsets likely ignore scooters’ significant manufacturing emissions). Rather, these acts are more of an exercise in accounting that provides a bit of positive PR. If Lime (or any other dockless scooter company) puts more efforts into releasing data related to emissions, cleaning up and making more efficient the collection & redistribution process, extending the lifetime of scooters, and minimizing the embedded emissions from the manufacture of scooters, they could push real progress in reducing the carbon emissions of the transportation sector. Given the massive challenges ahead to tackle climate change, we’ve no time to spare– we need to push these innovative companies to make actual change today.