Electric Trains, Electric Cars, or Electric Bikes. Which is best for the environment?

You’ve swapped your petrol car for a plug-in hybrid.

Or perhaps you’ve gone full electric.

Maybe you’ve given up the car entirely to take the train to work instead.

Many of us are playing our part in trying to fix the climate crisis we’re all facing.

Though you might be surprised at the environmental cost of seemingly green modes of transport.

Before you buy that electric scooter, you’ll want read this.

Methodology

travelandmobility.tech have curated a great data set analysing the environmental (carbon) impact of a range of popular transport types.

  • Operation (direct): The environmental impact caused by the direct operation of the vehicle (e.g. abrasion emissions from brake linings, wheels…)
  • Operation (indirect): The environmental impact of indirect operation is determined, which primarily includes the provision of energy (e.g, processes from energy extraction from the environment to delivery to the tank…).
  • Maintenance: All the processes required to keep the vehicle roadworthy during its service life are counted (e.g. changing the tires of cars and replacing consumables in railway trains…).
  • Manufacture & Disposal: This category includes all processes that affect the manufacturing of the vehicle that are not included in maintenance (e.g. raw materials, operating emissions of the production facilities…)
  • Roadway: The construction, maintenance, and disposal of all types of tracks are counted (e.g for road transport these include roads, car parks etc., for rail traffic these include entire lines, safety walls, bridges…)

The impact of each of these factors is measured as carbon emissions in grams per passenger kilometre.

There are a number of assumptions that have been made to compile the data, including average level of occupancy per transport type (although in cases of transport types that carry multiple passengers, this figure if not reported) and the average lifetime (distance travelled) for each transport type.

Results

Operational Emissions

Carbon emissions for transport operation in grams per passenger kilometre

Download chart.

Category Operation (direct) g p/pkm Operation (indirect) g p/pkm Operation total g p/pkm
by Foot 0.00 0.00 0.00
Bike 0.00 0.00 0.00
E-Bike 0.00 1.01 1.01
E-Scooter (Vespa-Like) 0.00 2.28 2.28
E-Kick-Scooter (Dockless) 5.92 0.00 5.92
Tram 0.37 13.63 14.00
E-Bus 1.45 14.31 15.76
Car (Electric) 4.07 12.68 16.75
Car (Plug-In-Hybrid) 20.35 5.68 26.02
Bus (>200km) 32.32 6.31 38.63
Train (Highspeed) 0.03 40.65 40.68
Bus (<200km) 43.30 8.43 51.73
Train (Regional) 9.11 45.15 54.26
Scooter (Gasoline) 75.64 15.15 90.79
Car (Hybrid) 86.22 20.96 107.18
Motorbike (Gasoline) 97.24 24.82 122.05
Car (Diesel) 106.01 20.65 126.67
Autobus 112.25 22.10 134.35
Ferry (<200km) 123.65 23.86 147.51
Car (Gasoline) 130.23 34.11 164.34

Full table.

A gasoline car has the highest direct operating emissions (130.23 grams per pax km) and indirect emissions (34.11 g p/pkm). That’s more than a ferry (123.65 g p/pkm // 34.11 g p/pkm).

High-speed trains are very efficient for day-to-day direct operation (0.03 g p/pkm), though the indirect costs are carbon expensive (40.68 g p/pkm).

Combined, an electric car is more carbon friendly than a train from a direct and indirect operational perspective (4.07 g p/pkm // 12.68 g p/pkm).

Manufacture & Disposal Emissions

Carbon emissions for transport manufacture and disposal in grams per passenger kilometre

Download chart.

Category Manufacture & Disposal g p/pkm
by Foot 0.00
Train (Highspeed) 0.55
Train (Regional) 0.73
Tram 1.38
Bus (>200km) 1.75
Bus (<200km) 1.88
E-Bus 2.80
Autobus 3.28
Ferry (<200km) 3.75
Scooter (Gasoline) 5.40
Bike 5.91
E-Bike 10.96
Motorbike (Gasoline) 16.36
E-Scooter (Vespa-Like) 23.09
Car (Gasoline) 32.69
Car (Hybrid) 37.30
Car (Diesel) 39.48
Car (Plug-In-Hybrid) 42.20
Car (Electric) 62.57
E-Kick-Scooter (Dockless) 63.00

Full table.

Electric powered transport is by far the most expensive to create and dispose of. That said, the carbon cost of this is likely to reduce significantly in future years as technology advances.

Currently, an E-Kick-Scooter is the worst type of transport based on the carbon cost (63g p/pkm) — that’s more than an electric car (62.57 g p/pkm)!

Despite their size, trains and trams have a low carbon cost to manufacture and dispose of (high-speed train 0.55- g p/pkm) – this is almost certainly due to the amount of passengers they carry in comparison to other forms of transport considered.

Lifetime Emissions

Carbon emissions total for transport in grams per passenger kilometre (2019)

Download chart.

Category Total g p/pkm
by Foot 0.00
Bike 7.64
E-Bike 16.12
E-Bus 25.15
E-Scooter (Vespa-Like) 29.84
Tram 37.47
Bus (>200km) 44.64
Train (Highspeed) 49.90
Bus (<200km) 58.20
Train (Regional) 59.64
Car (Plug-In-Hybrid) 82.30
Car (Electric) 92.37
Scooter (Gasoline) 100.57
E-Kick-Scooter (Dockless) 126.00
Motorbike (Gasoline) 145.02
Autobus 145.41
Ferry (<200km) 151.45
Car (Hybrid) 158.06
Car (Diesel) 179.60
Car (Gasoline) 208.28

Full table.

Adding in maintenance and roadway costs, in addition to other factors considered, traditional diesel and gasoline cars are the most polluting over their lifetime (179.60 g p/pkm and 208.28 g p/pkm, respectively).

Plug-in hybrids have half the carbon impact compared to tradition hybrids (82.30 g p/pkm and 158.06 g p/pkm, respectively), and are even more emission friendly over their lifetime than pure electric cars (92.37 g p/pkm).

How far to generate a tonne of C02?

How many km transport type to generate tonne of co2 per pax 2019

Download chart.

Category How many km for tonne co2 / pax?
by Foot
Bike 130,868.61
E-Bike 62,028.67
E-Bus 39,761.43
E-Scooter (Vespa-Like) 33,516.37
Tram 26,685.14
Bus (>200km) 22,401.85
Train (Highspeed) 20,040.19
Bus (<200km) 17,182.13
Train (Regional) 16,767.27
Car (Plug-In-Hybrid) 12,150.77
Car (Electric) 10,826.13
Scooter (Gasoline) 9,943.46
E-Kick-Scooter (Dockless) 7,936.51
Motorbike (Gasoline) 6,895.58
Autobus 6,877.08
Ferry (<200km) 6,602.84
Car (Hybrid) 6,326.73
Car (Diesel) 5,568.01
Car (Gasoline) 4,801.13

Full table.

In a gasoline car it takes on average just 4,800 km for each passenger to contribute a tonne of carbon dioxide. A passenger in an electric car will generate a tonne in just under 11,000 km, and a high-speed train in just over 20,000 km.

Note, it is important to stress, most of the emission are down to manufacturing costs (e.g. a Land Rover Discovery in 2010 required 35 tonnes CO2e for manufacture). See Methodology section for assumptions on lifetime distances.

Improvements

As the authors of the dataset note:

… [the results] not illustrating scientifically-proven results but provides our best guess on average carbon emissions produced by transport type based on existing third-party research that we were able to identify and combine.

It is also clear, air transport is missed. Interestingly, one of the data sources referenced is Lufthansa Innovation Hub.

It is impossible to get true figures for an analysis, there are simply too many variables, that said, the numbers used for analysis in this post could definitely be improved for a more accurate output.

tl;dr

In a gasoline car it takes just 4,800 km for each passenger to contribute a tonne of carbon dioxide. A passenger in an electric car will generate a tonne in just under 11,000 km, and a high-speed train in just over 20,000 km.

Footnotes

  1. Data sources + data used in this post.

The $234,650,000,000 (two hundred thirty-four billion six hundred fifty million) subway

In 2018, I declared the New York City subway system “the best”, based on the factors considered.

Cities like London or Moscow have had their subway networks in place for over 100 years. Although both have grown in the intervening time.

Modern day cities like Shanghai have built comparatively new and large systems at an astonishing rate.

New York is planning to have spent around $35 billion between 2005 and 2030 on subway and commuter rail expansion. But it’s only getting 15 km of new tunnel!

Paris is spending a similar amount over the same period: ‚ā¨40 billion, for a total of 228 km, 187 km underground.

Madrid, a much smaller city, spent ‚ā¨10 billion in 1995-2015 on 234 km, around 180 km underground.

Let’s take a deeper look…

Methodology

Pedestrian Observations (Alon Levy) has compiled a comprehensive presentation titled; “What is the Cost of Building a Subway?“.

The figures used for my analysis are taken from that presentation, which uses data from an unreferenced dataset of different urban rail lines and their costs.

According to the presentation, “Costs cover engineering, contracts, and political factors.”

Results

Approximate construction costs by region per km, in millions (USD)

Download chart.

Country / Region USD million/km
East Asia 100
Turkey 100
Mediterranean / Nordic / Switzerland 120
Chile 150
Iran 200
Western Europe 250
China 250
Mexico / Brazil 330
Thailand 475
US / Canada / Australia / Singapore / UK 500
Philippines 1000

Download table.

Think infrastructure project in the Europe / US are expensive?

In Manilla, Philippines, they’ve spent and estimated $1 billion USD per/km. Twice as expensive as the US / Canada / Australia / Singapore / UK ($500 million USD per/km).

Halve that by building in Western Europe. $250 million USD per/km.

Head to Mediterranean / Nordic countries, or Switzerland, and costs plummet even further. $120 million USD per/km.

Approximate construction costs by era per km, in millions (2019 USD dollars)

Approximate construction costs per subway km, in millions (2019 dollars)

Download chart.

Era New York (mm USD 2019) London (mm USD 2019) Paris (mm USD 2019)
1900s 40 30 30
1910s 55
1930s 140 35 30
1960s-70s 700 150
1990s 1500 500 250

Full table.

Who said things got cheaper over time?

Costs to build subways have spiralled.

$40 million USD per km in New York in the 1990s. $1.5 billion USD per km in the 90’s.

While many of you, like me, might assume this is because of political factors or city density, you’re probably wrong.

The author of the presentation notes:

The cost difference seems to be mostly in stations.

To save money, he has uncovered the following cost items that best reduce construction costs:

  • Shallow cut-and-cover construction, disrupting the street for about 18 months. No mining except at undercrossings.
  • No mezzanines. All circulation, including fare barriers, should be on the platform level or at street level.
  • An island platform, ideally accessed from a street median, to avoid duplicating elevators, stairs, etc.
  • No signature architecture. Station designs should be reused systemwide. If art is desired, put on exhibits.

Rebuilding the London Underground in 2020

Est. Build Cost of London Underground Lines using 1990s costs (USD 2019)

Download chart.

Name Length (km) Cost at 1990s cost (mm USD 2019) Cost at 1990s cost (USD 2019)
Waterloo & City line 2.5 1250 $1,250,000,000
Victoria line 21 10500 $10,500,000,000
Bakerloo line 23.2 11600 $11,600,000,000
Hammersmith & City line 25.5 12750 $12,750,000,000
Circle line 27.2 13600 $13,600,000,000
Jubilee line 36.2 18100 $18,100,000,000
Northern line 58 29000 $29,000,000,000
District line 64 32000 $32,000,000,000
Metropolitan line 66.7 33350 $33,350,000,000
Piccadilly line 71 35500 $35,500,000,000
Central line 74 37000 $37,000,000,000
Total 469.3 234650 $234,650,000,000

Full table.

Assuming 1990’s London costs per/km of subway ($500mm p/km), even the 2.5 km Waterloo & City line would cost an estimated $1.25 billion to rebuild — maybe at that cost it would open on Sundays?

In total, it would cost an estimated $235 billion to build the London Underground today.

The London Underground carries about 1.4 billion passengers a year. If every journey cost a passenger $168 (about 12 times the cost of a travel card, which is already expensive!) the costs to build would pay for themselves in a year.

Built at Swiss prices ($120 million /km), it would be 75% cheaper at about $56 billion — a bargain… when compared to New York costs (@$1.5 billion per/km = $705 billion total).

Comparing to HS2 (UK)

HS2 vs London Underground construction costs per km

Download chart.

For those outside the UK, HS2 stands for high-speed 2, a planned high-speed rail network between London, the West Midlands, Manchester and Leeds.

It’s controversial, like many major infrastructure project, though the ¬£307 million / $398 million USD per mile (¬£190.8 / $247.3 million /¬† per km) it is expected to cost (today) has been widely reported by the press.

Comparing this to the per kilometre cost of rebuilding the London Tube at 1990s prices ($500 million / km), HS2 doesn’t sound as expensive, though you don’t get as many tunnels for your $247.3 million!

Improvements

I’ve taken aggregated prices from the presentation that have already been manipulated by the author. To get a better insight into build costs and where they were sources from would dramatically improve the quality of my reporting.

tl;dr

In total, it would cost an estimated $235 billion to build the London Underground today. 

Footnotes

  1. Data sources + data used in this post.

Patient 0 to the World: How Air Travel Makes it Impossible to Contain COVID-19

Corona.

What was once a summer beer is now synonymous with something far less appealing.

COVID-19, or the Corona virus, has sadly led to over 2,500 deaths and almost 100,000 infections as I write this.

Recently I was reading about the World War 1 flu pandemic that claimed an estimated 16 million lives. It is estimated one fifth of the world’s population was attacked by this deadly virus.

Most researchers attribute the movement of people around the world to the fact the flu virus was able to infect so many.

And this was before the days of commercial aviation.

In 2018 there were 4.8 billion air passengers, total. Add in rail, road and sea journeys, and it’s clear the world is incredibly interconnected. There wasn’t even 4.8 billion people on the planet in 1914 (most estimates put it at between 1.5 and 1.7 billion).

From its origin in Wuhan, here’s a simple analysis for how easily it could have been spread around the world.

Methodology

I used a variety of sources to obtain data on air travel in China to estimate and analyse passenger traffic and aircraft movements.

Results

Air Pax Volume China (2019)

China air passenger volume 2019

Download chart.

In total, there were about 660 million passengers flying from a Chinese airport in 2019.

Almost 90% were flying domestically (586 million pax), with 72 million flying out of the country — the equivalent of around 49 million domestic and 6 million international pax each month.

Where do people fly to / from in China?

Download chart.

Rank Airport Passengers
1 Beijing Capital International Airport 100,983,290
2 Shanghai Pudong International Airport 74,006,331
3 Guangzhou Baiyun International Airport 69,720,403
4 Chengdu Shuangliu International Airport 52,950,529
5 Shenzhen Bao’an International Airport 49,348,950
6 Kunming Changshui International Airport 47,088,140
7 Xi’an Xianyang International Airport 44,653,311
8 Shanghai Hongqiao International Airport 43,628,004
9 Chongqing Jiangbei International Airport 41,595,887
10 Hangzhou Xiaoshan International Airport 38,241,630

Full chart.

Over 100 million passengers flew in or out of Bejing in 2018, or a mean average of 8.3 million per month.

Even the smallest airport in the top 100 by passenger volume, Nanyang Jiangying Airport, saw over 907,000 passengers through its doors in 2018.

Wuhan Tianhe International Airport had 24.5 million in 2019, or about 2 million per month — about the same amount of time before travel restrictions came into place and the virus was widely reported.

How many flights depart from Wuhan each month?

I could not find specific flight data for Wuhan, so let’s get creative.

Given most travellers are domestic, let’s use one of the most popular short/medium range aircraft, the Boeing 737 (ignoring the ongoing MAX 8 problems).

The 737 MAX 8 typically holds around 178 in a 2 class seat configuration.

Assuming only the 737 Max flew from Wuhan, that would mean over 11,235 flights landed / departed. Given there will be larger planes in operation, let’s assume 10,000 plane movements per month.

Divide that by two, to only consider departures, gives 5,000 plane departures per month.

And this is one city alone.

Summary

According to this same calculation using the amount of 737 seats to estimate number of flights would result in the 4.8 billion passengers who flew in 2018 to have done it on about 60 million flights or 5 million each month!

And that’s just air travel.

Without a total ban on travel, I cannot see how COVID-19 will be contained.

To finish, it is important I note this is not meant to be a post designed to scare.  Remember, even if you contract the virus, it is very likely you will survive.

Improvements

These stats are clearly not accurate model of the spread of COVID-19. The post is designed to highlight how interconnected the modern world is.

I’m very interested to see the models that researchers develop as our understanding of this virus increases. I am no where near skilled enough to do this.

tl;dr

With an estimated 5 million flights taking off around the world each month, stopping viruses penetrating borders is an impossible task.

Footnotes

  1. Data sources + data used in this post.

The Best Cities for Travelling via Public Transportation

Despite the increase of Uber and similar on-demand ride services in most major cities, public transportation still serves millions of locals and tourists every year.

Living in London it is easy to take public transportation for granted. With tube lines snaking across the city, bus stops on almost every road, and a cycle hire scheme for those rare sunny days it is easy to venture across the UK’s capital without wondering how you’ll get to a destination.

Though in other larger cities, even though public transportation systems do exist, they are often sparse and serve a seemingly small population. For me this is especially true of cities in America. In San Francisco I find it generally easier and sometimes cheaper to take an Uber over the BART.

Having recently returned from the America, this thought sparked an idea for a post. Which cities have the best public transportation networks?

Methodology

Wikipedia users have curated a list of metro systems around the world. The data includes; date the system was first opened, the size of the network, the number of stations and yearly ridership figures. In total 177 metro systems are listed.

I also collected population data for a cities metro area from Wikipedia to compare against yearly ridership figures.

Results

The oldest networks

Age rank City Country Name Year opened
1 London United Kingdom London Underground 1890
2 Budapest Hungary Budapest Metro 1896
2 Glasgow United Kingdom Glasgow Subway 1896
4 Chicago United States Chicago “L” 1897
5 Paris France Paris Métro 1900

Full ranking.

The UK has the oldest network dating back to 1890 (128 years old). This is significantly below the average age of 1984 (38 years old) for all 177 networks considered. The average age figure is significantly boosted by new metro systems appearing across growing Chinese and Indian cities. In a previous post I covered how the rapid growth of Chinese high-speed rail links.

The age of each network got me thinking; will older networks be larger than their smaller counterparts? It seems logical given the additional time they have had to expand.

Age versus size

System Length Versus Age

Download chart.

Perhaps surprisingly the average metro system is just 70 kilometers. Again, this figure is slightly skewed due to the new metros emerging in cities.

System Length Rank City Country Name System Length kilometers
1 Shanghai China Shanghai Metro 637
2 Beijing China Beijing Subway 599.4
3 London United Kingdom London Underground 402
4 Guangzhou China Guangzhou Metro 390.7
5 New York City United States New York City Subway 380.2

Full ranking.

China is home to three of the five largest metro networks with Shanghai’s system the largest at a total length of 637 kilometers. First opened in 1993 it spans 200 more kilometers than the London Underground opened over 100 years earlier.

However, length isn’t everything. As a passenger it is important that many destinations are served allowing maximum freedom to roam a city.

Length versus number of stations

System Length Versus Number of Stations

Download chart.

There is a linear correlation between number of stations and system length. As the size of a network grows as does the number of stations. This suggests that most metros follow a similar pattern in spacing out their stations. This is to be expected given the way cities grow and the fact that a small number of companies design and build these networks.

Rank num stations City Country Name Stations
1 New York City United States New York City Subway 424
2 Shanghai China Shanghai Metro 324
3 Seoul South Korea Seoul Subway (lines 1-9) 307
4 Beijing China Beijing Subway 306
5 Paris France Paris Métro 302

Full ranking.

Despite being only the fifth largest network by length, New York boasts the highest number of stations serving passengers — 100 more than Shanghai even though it is over 250 kilometers shorter in length.

This highlights another important question; how well are travelling passengers served by stations?

Stations versus passengers

Stations Versus Passengers

Download chart.

A linear correlation between stations and passengers is seen for the 161 metros there is ridership data for. 16 systems carry more than 1 billion passengers each year.

Ridership rank City Country Name Ridership Ave pax per station
1 Beijing China Beijing Subway 3,660,000,000 11,960,784
2 Shanghai China Shanghai Metro 3,401,000,000 10,496,914
3 Seoul South Korea Seoul Subway (lines 1-9) 2,856,500,000 6,056,790
4 Guangzhou China Guangzhou Metro 2,800,000,000 14,070,352
5 Tokyo Japan Tokyo Metro 2,642,100,000 9,184,906

Full ranking.

Each of the top five metro systems each serve over 2.5 billion journeys a year. Beijing’s metro serves a staggering 3.66 billion journeys each year!

Despite having the most stations, 424, New York Cities Subway network carries just 1.75 billion passengers each year. Compare that to Beijing, 306 stations and over 3.6 billion passengers a year — that’s an average of 11 million people travelling through each station on the network each year.

In recent years there has been some talk of so called “white elephant” infrastructure projects being undertaken in China. Perhaps there is some truth in this for metro networks?

Passengers versus population

Passengers vs Metro Population

Download chart.

There is a slight correlation comparing passengers to population.

Rank pax/pop City Country Name Metro Population Ridership (year) % Pop vs ridership
1 Milan Italy Milan Metro 1,365,156 468,300,000 34303.77%
2 Taipei Taiwan Taipei Metro 2,704,974 746,000,000 27578.82%
3 Munich Germany Munich U-Bahn 1,464,301 398,000,000 27180.20%
4 Hong Kong China MTR 7,409,800 1,767,100,000 23848.15%
5 New York City United States New York City Subway 8,175,133 1,756,800,000 21489.56%

Full ranking.

Milan’s metro system serves¬†34,304% more passengers each year than the metro areas total population! 22 systems see yearly ridership greater than 10,000% higher than the cities population too.

At the other end of the spectrum the Hefei Metro system in China carries around 18% of the cities population each year over its 52.4 kilometre network.

tl;dr

Despite being the 5th largest network by size, the New York City Subway comprising of 424 stations offers the most flexibility for passengers. Compared to other similarly large networks the New York Subway also has comparatively lower passenger numbers meaning less crowding onboard.

Get the Data

Get all the data used in this blog post on Google Sheets.

Concorde Would Still Beat Hyperloop

High-speed trains now compete with airlines for short-haul routes.

The fastest high-speed train in operation is limited by a maximum speed of 431 km/h, although trains have reached speeds of 603 km/h in testing.

Hyperloop is a proposed new mode of passenger and freight transportation that propels a pod-like vehicle through a near-vacuum tube at airline speeds. Top speeds of 1200 km/h have been touted. To put that in perspective a Boeing 777-200-LR, one of the fastest commercial jets in operation has a maximum operation speed of 1037 km/h.

What could a world with interconnected Hyperloop routes look like?

Methodology

Given Hyperloop is still in development phase, with some limited testing completed, much of this post takes into account assumptions about what is possible (speeds, geography, engineering limitations etc).

Many of the routes I’ve considered¬†are already in operation — served either by plane or train. Distances for these routes via Hyperloop are very rough given I have not considered geography in any detail. That said,¬†because of the physics involved Hyperloop needs a very straight track.

The aim of this post is to provide a basic overview of what could be made possible with Hyperloop.

Analysis

Maximum operating speeds

Max operating speed by mode of transport

Download chart.

The fastest of each mode of transport currently in operation are considered: Hyperloop, Boeing 777-200LR (plane), Shanghai Maglev (train).

Hyperloop is most similar to the existing train networks. The difference between current high-speed trains and Hyperloop is massive (1200 km/h vs. 431 km/h — almost 3 times faster.).

Average Speeds for Popular Air / Rail Routes

Plane vs train vs Hyperloop

Download chart.

For the above 5 routes, all of which show actual times for plane and train journeys, the train is always the slowest (almost twice as slow as a plane for most routes).

Travelling by plane would be¬†the second quickest option, with Hyperloop coming a clear first for all routes (in some cases Hyperloop is almost 4 times quicker on some routes — Paris to Lyon, and twice as quick as taking the plane on others).

It is important to note these figures only consider time spent moving, excluding check-in time, etc.

There are a number of cities currently considering implementing Hyperloop. LA to San Francisco, Dubai to Abu Dhabi, Helsinki to Stockholm, etc. Most of these routes are well below 1000km, as are the routes considered above with the exception of Beijing to Shanghai at just over 1000km apart. At an estimated $19.1 million USD per kilometre, cost is the biggest inhibiting factor to a longer Hyperloop routes at present.

Did you know? It will cost an estimated $2.65 billion USD ($19.1MM USD x 139km) to build a Hyperloop route between Abu Dhabi and Dubai.

Hyperloop potential

Concorde once circumnavigated the globe in 32 hours 49 minutes and 3 seconds, starting and ending its journey in New York, via Toulouse, Dubai, Bangkok, Guam, Honolulu, Acapulco, to refuel and then back to New York JFK (a total of 36,787.6 km).

Assuming Hyperloop took the same journey at its average estimated speed of 970 km/h (which unlike a plane would not need to account for refuelling time), it would take almost 38 hours (36787.6 km/970 km/h). That’s 5 hours slower than Concorde.

Hyperloop round the world

Full map.

The circumference of Earth at the equator is about 24,874 miles (40,030 km), but from pole-to-pole ‚ÄĒ the meridional circumference ‚ÄĒ Earth is only 24,860 miles (40,008 km) around. This shape, caused by the flattening at the poles, is called an oblate spheroid. A “true” round-the-world journey circling¬†the equator would take over 41.3 hours (40030 km¬†/970 km/h) at a cost to build of $764.6 billion USD (40030 km¬†* $19.1 million USD per km).

Improvements

As discussed, many of these calculations are very rough. When Hyperloop gets closer to a production version where better estimates are available the calculations used can be improved.

tl;dr

Concorde once circumnavigated the globe in 32 hours 49 minutes and 3 seconds, the same route on Hyperloop would take almost 38 hours.

Get the Data

 

It’s Cheaper To Take An Uber To The Airport, And Back

Uber has been fighting continuing battles with major cities and their traditional licensed cab drivers for years now. In some cities this has resulted in legal rulings banning Uber drivers from operating completely.

However it’s not just cities taking aim at¬†the taxi app. Airports are also drawing battle lines with the company¬†in an effort to protect revenues.

The ease of catching a cab is often paramount for weary travellers, but many are still very price-sensitive even after the longest of flights.

But are the costs of Uber rides on fiercely competitive (and regulated) airport journeys cheaper then traditional taxis (as they are in cities)?

Methodology

I was able to calculate fares for Uber journey costs from airports on their website (UK & US) (April 2016).

These could then be compared with traditional taxi fare data available from TfL (UK) and Taxi Fare Finder (UK & US) (April 2016).

The Black Taxi

The Black Taxi an icon of London. In recent years Black Taxi drivers have been losing market share of passengers travelling overground to minicabs (often operated by taxi apps, like Uber). The situation is so bad, TfL (who operate Black Cabs), are struggling to recruit new Black Cab drivers.

Queue a fierce fight between TfL and Uber.

One that almost saw Uber being banned from the British capital altogether in 2015 but for a small technicality about wether the app constituted a taximeter. But technicalities aside; why are consumers switching to taxi apps?

Analysis

Uber Services

Uber vs Black Taxi Journey Costs in London

Download chart.

Service Ave Fare (1 mile) GBP Ave Fare (2 mi) GBP Ave Fare (4 mi) GBP Ave Fare (6 mi) GBP
UberX 5.18 7.25 10.95 15.10
UberX (1.9x surge) 9.83 13.78 20.81 28.69
UberX (2.5x surge) 12.94 18.13 27.38 37.75
UberEXEC 9.25 13.30 20.50 28.60
UberLUX 13.78 20.35 31.85 45.00
UberXL 10.53 13.45 18.85 24.70
Black Taxi (Mon-fri (06:00-20:00)) 7.20 11.20 18.50 26.00
Black Cab (Mon-fri (20:00-22:00), Sat-sun (06:00-22:00)) 7.30 11.50 19.00 30.00

Full table.

UberX rides are generally cheaper than traditional Black Cabs, unless the dreaded surge pricing is in effect.

For example, the more expensive Black Taxi Fare would cost ¬£7.30 GBP to go 1 mile. An UberX would cost only ¬£5.18 GBP. However, during periods of high demand where 1.9x and 2.5x surge pricing is in effect the same journey would cost you ¬£9.83 GBP and ¬£12.94 GBP respectively through Uber — potentially well over twice the price of a Black Taxi.

Uber Services To / From London airports

Registered taxis in London (Hackney Carriages, not necessarily Black Taxis) have strictly regulated fares from its many airport into the city.

Uber vs Taxi cost from airport to Paddington Station, London

Download chart.

London Heathrow (T5) – Paddington St GBP (ave) Gatwick South¬†— Paddington St GBP (ave) Stansted – Paddington St GBP (ave) Luton- Paddington St GBP (ave) City- Paddington St GBP (ave)
uberX 37.00 96.00 74.50 46.50 28.50
uberX (1.9x surge) 55.50 144.00 111.75 69.75 42.75
uberX (2.5x surge) 92.50 240.00 186.25 116.25 71.25
uberXL 56.00 149.50 114.00 89.50 40.00
UberEXEC 71.00 135.50 104.50 82.50 55.00
UberLUX 109.50 283.50 221.50 171.00 88.50
Taxi (Mon-fri (06:00-20:00)) 67.40 125.60 78.80 59.90 40.60

Full table.

Many registered taxis and private hire cars pay surcharges to airports (often passed on to passengers directly). Uber drivers are also liable for such charges.

Even so, UberX rides are considerably cheaper from all London airports — almost ¬£50 GBP cheaper from Gatwick South Terminal to Paddington Station.¬†Only recently did Uber scrap flat fees from London airports which could¬†have potentially made these journeys even cheaper.

Registered taxis sit somewhere between UberX and UberXL services, and are the second cheapest option when compared to the other services Uber offers.

Uber To / From US airports

Lets change colours to the bright yellow of the second most iconic taxi, the New York City Cab (and the less iconic taxis of San Francisco and Chicago). And yes, both these cities have had their disagreements (like the London Black Cabs), to put it lightly, with Uber.

Uber vs Taxi cost from US airport to major landmark

Download chart.

JFK – Grand Central (ave) USD SFO – Market St (low) USD ORD – Chicago Union (ave) USD
uberX 59.00 25.00 33.00
uberXL 88.50 37.00 56.50
UberBLACK 116.50 68.00 88.00
UberSUV 145.50 84.00 111.50
Taxi (Mon-fri (06:00-20:00)) 64.73 64.69 52.45

Full table.

Like London, the UberX service comes out much cheaper than choosing a taxi in all cities. In San Francisco and New York the fares are staggeringly cheaper. In New York UberX services are about $20 USD cheaper, and in San Francisco almost $40 cheaper!

Improvements

Uber operates in 81 countries (Oct 2016). In the US and UK cities considered, UberX is cheaper than traditional taxis to and from airports. I would be interested to learn if the same was true in the 79 other countries.

tl;dr

UberX services from airports are cheaper than registered taxis from airports.

Footnotes

  1. Data sources + data used in this post.

High-Speed Rail Is Killing Short-Haul Air Travel

Planes dominate our skies. According to some estimates there are about 100,000 flights, everyday (about 25% of which are flown by low-cost airlines). Where trains were once the best ways to travel 700km, many are now choosing to fly instead.

But long-distance train travel has seen a recent explosion in many countries with the growth of high-speed rail networks. Bullet trains were once only associated with Japan (they have been operating in the country since 1964!). Though today high-speed trains are much common across the world.

As high-speed rail networks grow, as does their viability to compete with equivalent plane routes.¬†China Southern Airlines, China’s largest airline, expects the construction of China’s high-speed railway network, the largest in the world by a large margin, to impact (through increased competition and falling revenues) 25% of its route network in the coming years.

But how do the two forms of transport currently stack up around the world?

Methodology

Train and plane routes used for comparison were selected to test the following statement on the “High-speed rail” Wikipedia page:

“High-speed rail (HSR) is best suited for journeys of 1 to 4¬Ĺ hours (about 150‚Äď900¬†km or 93‚Äď559¬†mi)”.¬†For trips under about 700¬†km (430¬†mi), the process of checking in and going through airport security, as well as traveling to and from the airport, makes the total air journey time equal to or slower than HSR.¬†European authorities treat HSR as competitive with passenger air for HSR trips under 4¬Ĺ hours.”

I chose 5 well trafficked routes of varying distances above and below 700km from places where both air and rail transport is common: Beijing – Shanghai (1318km), Madrid to Barcelona (621km), London to Paris (492km), Tokyo to Osaka (515km), and Paris to Lyon (409km).

Note, the absence of any US cities. Whilst I did some brief research on potential routes in the country, I could find none that posed a current threat to the massive low-cost airline market. Things are changing though.

For train ticket data for journey costs and time I used Seat61 (25/05/2016).

Using data from Skyscanner I collected flight schedules for each route to obtain the fastest flying time between airports. I also used Skyscanner to search for the cheapest possible fares available during the month of September for comparison (25/05/2016).

To estimate carbon emissions for journeys on each mode of transport I used data from the Aviation Environmental Federation.

Train and Plane Route Distance Jun 2016

Download graph.

In order to calculate emissions I needed actual journey distance. Train track length is freely available for the routes I selected. Unfortunately, planes never take an exact route due to weather conditions, delays, air traffic, etc. Because of this I used point-to-point distance between airports. As a result, figures used are less than average actual distances flown.

Results

Speed overview

Model Type Max operating speed (train) / cruise ground speed (plane) km/h
Boeing 777-200LR Plane 1037
Boeing 707-320 Plane 963
Lockheed L-1011-1 Plane 963
Sukhoi Superjet 100-95 Plane 954
Tupolev Tu-414A Plane 950
Boeing 777-200ER Plane 948
Airbus A350-1000 XWB Plane 945
Airbus A350-900 XWB Plane 945
Airbus A350-800 XWB Plane 945
Boeing 777-300 Plane 945
Airbus A380-800 Plane 945
Airbus A350-1000 Plane 945
Shanghai Maglev Train 431
Harmony CRH 380A, Train 380
AGV Italo Train 360
Siemens Velaro E/AVS 103 Train 350
Talgo 350 Train 350
E5 Series Shinkansen Hayabusa Train 320
Alstom Euroduplex Train 320
SNCF TGV Duplex Train 320
ETR 500 Frecciarossa Train Train 300
THSR 700T Train 300

Full table.

This table shows the fastest planes and trains currently operating commercially.

Most modern aircraft cruising speeds are just below 1,000 km per hour. The Boeing 777-200LR is the only commercial airliner that has a cruising speed above this figure at 1,037 km/h Рabout 200 km/h slower than the speed of sound (MACH 1 = 1,236 km/h). The 777s speed is probably one of the reasons it is used on many of the longest non-stop commercial routes.

Did you know? Concorde had an average cruising speed of 2,140 km/h.

The world record for the fastest train belongs to Mitsubishi L0 Series maglev. A speed of 603 km/h was achieved on a test track in April 2015. Whilst world records help push innovation, the sad fact is actual operating speeds for the travelling public are much slower.

The Shanghai Maglev, the fastest train in operation today reaches a max speed of 431 km/h¬†in just over 3 mins 20 secs¬†(it has reached a non-commercial speed of 501 km/h). You won’t be travelling that fast for long, the track is only 30km long¬†and the fastest scheduled journey time to cover the distance is 7 mins 20 secs.¬†The¬†line was designed to connect Shanghai’s Pudong airport to the¬†outskirts¬†of the city so that passengers can quickly connect to¬†metro lines.

The¬†9 other fastest trains in operation¬†travel at speeds between 300 – 380 km/h. The 10th fastest airliner has a cruise speed of¬†945 km/h — over 3 times faster than the slowest train.

Journey Times

Train and Plane Journey Time Jun 2016

Download graph.

Whilst planes can take more direct routes, stations are usually much more time efficient for passengers. Stations are often found in the middle of cities with great transport links. Take Paris, for example. Charles Du Gaulle Airport is a 29km car journey to the centre of the city. Gare Du Nord is walkable to many hotels in the city.

Check-in times are also much more lenient at stations. Whilst you might need to arrive at the airport 2-3 hours before departure, some international train journeys advise you to arrive 45 mins before departure¬†— a potential time saving of over 2 hours.

Given this, lets assume additional travel time for air travel is +150 minutes and  +60 minutes for rail.

Train and Plane Complete Journey Time Adjusted Jun 2016

Download graph.

Even taking these factors into consideration, travel times, in most cases, are tipped in favour of the plane for three of these five journeys. Notably all three journeys are below the 900 km / 270 min estimates described as the maximum distance high-speed trains could compete well with air travel.

In China, travelling by plane between Beijing and Shanghai (1077 km by plane, 1318 km by train) could save you over one hour in travel time compared to the train (285 min by plane, 348 min by train).

There is no difference in time (225 min) between each mode of transport on the Madrid to Barcelona route (483 km by plane, 621 km by train) suggesting the optimum distance for high-speed train travel versus plane might be slightly less than 900km / 270 min.

Of course these are very crude assumptions as my adjustments do not take into account accurate door-to-door journey time for each city which can vary enormously.

Did you know? You’ll save the most time taking the Eurostar from London to Paris saving 50 minutes compared to flying.

Average Speeds

Journey Average Speeds

Download chart.

It is interesting to compare average speed to top speed of the train serving the route.

The Beijing РShanghai route average speed for the journey is only 25 km/h slower than the top speed (275 km/h vs. 300 km/h). The trains are operating close to their allowed safe maximum throughout the route (all 1077 km!).

On other routes the difference between average and maximum is¬†greater. The Eurostar Paris – London route is the second best for speed efficiency (300 km/h vs. 256 km/h). Compare that to the Paris to Lyon route 300 km/h vs. 187.33 km/h — average being almost 40% slower.

Journey Cost

Train and Plane Journey Cost Jun 2016

Download graph.

I didn’t realise just how expensive riding the bullet train was in Japan. A fare between Tokyo and Osaka will cost you at best, $124 USD. The cheapest airfare between the two cities is 63% cheaper (about $80 USD — $124 – $46) — a difference that would make most leisure travellers think twice. Interestingly it is only the third longest¬†journey of those analysed (463 km by plane / 515 km by train), resulting in a high-cost per km.

For the other four routes there was very little in cost difference between the cheapest tickets. The second greatest ticket cost difference was on the Paris to Lyon route where air travellers could make savings of about $25 USD.

It’s worth noting you’ll also avoid additional charges often imposed by budget carriers on short-haul routes by choosing the train.¬†The train journeys analysed have no additional charges for personal luggage.

Journey Emissions

Train and Plane CO2 Emissions Per Passenger Jun 2016

Download graph.

According to the Aviation Environmental Federation trains emit 2.5 times less CO2 per passenger than planes (based on optimum fuel efficiency for both modes of transport).

Planes use the most fuel, and produce the most harmful emissions, during takeoff. On short flights, as much as 25 percent of the total fuel consumed is used at this time. The most fuel-efficient route length for airlines is 4,300 km, roughly a flight from Europe to the U.S. East Coast. About 45 percent of all flights in the European Union (and 80% of the routes analysed in this post) cover less than 500 km.

Based on these figures for the journeys analysed it is clear how much more polluting planes are per passenger when compared to trains.

Did you know? Between Beijing and China 100kg more CO2 per passenger will be emitted on plane journeys when compared to a train journey (plane = 188.80 kg/CO2, train = 79.34 kg/CO2).

Summary

Route Plane Distance (km) Train Distance (km) Is train faster? Is train cheaper? Plane cost / time Train time / cost
Beijing-Shanghai 1077 1318 -18.10% 4.94% 0.30 0.23
Madrid-Barcelona 483 621 0.00% 27.78% 0.20 0.16
London-Paris 380 492 28.57% 38.10% 0.26 0.24
Tokyo-Osaka 463 515 12.68% -62.90% 0.19 0.58
Paris-Lyon 392 409 9.95% -31.25% 0.26 0.42

See calculations.

Improvements

Clearly more routes could be used for comparison. 5 routes is too small of a sample size to provide any conclusive findings.

I’d also be interested to obtain passenger load statistics (number of seats occupied for each journey) to improve emissions figures. Although emissions are far less of a concern to consumers, if at all, compared to journey price and time.

Comfort is also a big consideration for travellers. Being able to obtain quantative data-points on passenger comfort would add another variable to base a conclusion on.

tl;dr

  • Trains are faster for journeys up to 620km (track distance).¬†You’ll save the most time taking the Eurostar from London to Paris saving 50 minutes compared to flying.
  • Costs are very sensitive to locale (not distance) for train travel, less so for air travel.¬†In some geographies, for example Tokyo to Osaka (515km by train), you can save over $80 USD when taking the train. Whereas taking the Eurostar from London to Paris you’ll pay $16 USD more for the privilidge. Though these figures do not take into account additional baggage charges often imposed on air travellers.

Acknowledgements

  • Axlegeeks,¬†who provide lots of interesting statistics¬†about aircraft — I used crusing ground¬†speeds.
  • Wikipedia for a wealth of¬†data on high-speed train travel.
  • Skyscanner, where I collected¬†plane journey data from.
  • Seat61, a site¬†that¬†provided train journey data.
  • and finally…¬†The Aviation Environmental Federation for emissions data.

Get the Data

Airports That Are Unreasonably Far From City Centres

On a journey home from the airport last year I shared a cab with a fellow passenger heading in the same direction. In the awkward small talk between two strangers he asked if I had heard of London Oxford Airport (OXF). “Oxford, in the UK?”, I asked in clarification. “Yes”, he chuckled.

For those not familiar with UK geography, Oxford is over 100km outside of London. A bold, if a little questionable, effort by the airports marketing team.

It would appear Oxford is not alone. See: Paris Vatry (XCR)¬†– 161km to Paris, Oslo Torp (TRF) – 119km to Oslo, or¬†Munich West (FMM) – 115km to Munich, as examples. Arguably none of these are considered their cities “main” airport, however.

Such extreme journeys got me thinking; where are the shortest and longest transfer distances from airports to the cities they serve?

Methodology

To obtain a list of cities for analysis I used¬†Euromonitors’ “Top 100 City Destinations Ranking“. I chose such a list¬†because passengers will expect good infrastructure from heavily travelled destinations and many will be places readers¬†have, or are intending to, visit.

Using EuroMonitors list of cities I then found the closest major airport using a Google search. If there are multiple main airports in a city, airports with the highest yearly passenger volume were selected.

I then used the reported co-ordinates of the geographical centre of each city using a WorldMaps dataset. Similarly, I used the geographical centre of a cities corresponding airport from the OpenFlights dataset to calculate the shortest route by distance on Google Maps using a car from airport to city.

Regional Distribution of Euromonitor Top 100 Cities 2014

Download graph.

It is important to note we are only considering the Top 100 tourist destinations as reported by Euromonitor. This does add some bias to the analysis with 70% of the cities considered located in Europe and Asia Pacific.

Analysis

All airports by distance to their cities

km to City Centre by Road Euromonitor Top 100 2014

Download graph.

km to City Centre by Road Euromonitor 100

Download map.

km to City Centre by Road Euromonitor Top 100 2014 histogram

Download graph.

Did you know? The large majority of Top 100 cities are served by airports within 50km by road to their closest main airport.

Based on the shortest road journey by distance (not time) you will need to travel a median distance of 20km from a main airport to its city (50 of the 100 cities are within 20km of their main airports).

Not all¬†cities¬†on this top 100 list had their own¬†airport.¬†The two cities with the longest connection between airport and city — Andorra la Vella, Andorra (TLS) (194km) and Edirne, Turkey (BOJ) (177km) — are both served by airports¬†from other cities. These two destinations received¬†2.4MM, and 3.1MM tourists respectively, with visitors¬†having to endure connections of around 3-4 hours by car (assuming a 50km/h average speed)!

Best airports by distance to their cities

km to centre (road) rank City Main Airport km to centre (road)
1 Jeju CJU 2.5
2 Heraklion HER 2.9
3 Mexico City MEX 5.3
4 Lisbon LIS 5.9
5 Mumbai BOM 6.5
5 Punta Cana PUJ 6.5
7 Ho Chi Minh City SGN 7.4
7 Nice NCE 7.4
9 Sofia SOF 7.5
10 Florence FLR 8.5

Full table.

Did you know? The airports closest to the main cities they serve are: 3. MEX (5.3 km to Mexico City), 2. HER (2.9km to Heraklion), 1. CJU (2.5km to Jeju).

Worst airports by distance to their cities

km to centre (road) rank City Main Airport km to centre (road)
91 Seoul ICN 48.8
92 Milan MXP 50.3
93 Jerusalem TLV 54.2
94 Kuala Lumpur KUL 66.4
95 Halong HPH 69.8
96 Artvin BUS 77.4
97 Mugla DLM 94.4
98 Makkah JED 103
99 Edirne BOJ 177
100 Andorra la Vella TLS 194

Full table.

Did you know? The airports furthest away from the main cities they serve are: 3. JED (103km to Mekkah), 2. BOJ (177km to Edirne), TLS (194km to Andorra la Vella).

Best regions for airports by distance to their cities

Min, max, median km from main airport to city by region Euromonitor Top 100 Cities 2014

Download graph.

The median distance from a main airport to the city its serves for each world region is less than 40km. If a short car journey is important, a destination in Australia and New Zealand, Central and South America, or Africa would be a good choice (median distance = 15km from main airport to city).

Secondary ‘airports’

Many budget airlines also use secondary airports, London Luton (LUT) or London Oxford Airport (OXF) around London for example. Secondary airports offer airlines lower operational costs compared to more central, main, airports thus ensuring they can maintain the cheapest fare. Consequently they are typically even further away from the city increasing travel times to and from the airport.

Improvements

To get a wider view I downloaded a dataset from OpenFlights listing 8108 airports and their latitude and longitude. I cross reference the cities listed in this dataset for each airport with cities listed in the WorldMaps dataset used earlier. 2231 cities matched between the two datasets.

I then looked at the point-to-point (as the crow flies) distance between two co-ordinates (city and airport) using a Haversine calculation for these 2231 cities and airports. Obviously the resulting distance from this calculation will be much less than actual road distance.

From these 2231 cities and airports I¬†get a median point-to-point distance¬†of¬†just over 7.1km from airport to city — about 13km less than the median distance by road for the Top 100 cities.

There are a number of different factors that could explain for this, in addition to the difference in calculation. Smaller cities often have more centrally located airports because of low building density, for example.

tl;dr

  • 3 airports are over 100km by car from the cities they serve: 3.¬†JED (103km to Mekkah), 2. BOJ (177km to Edirne), TLS (194km to Andorra la Vella).
  • If you’re visiting a top 100 destination, the median¬†distance from a cities main airport to the city centre is about 20km.
  • A¬†destination in Australia and New Zealand, Central and South America, or Africa would be a good choice if a short journey to and from the airport is important (median distance =¬†15km¬†from main airport to city).

Acknowledgements

Get the Data