The biggest rail transport flows, measured in passenger-km

[Nottingham59]

I was looking at @RailAleFan 's extremely useful Flow Statistics website, which lists the top 100 station-to-station passenger flows, as reported by ORR. I see the biggest flow is Liverpool St to Tottenham Court Rd at 8000 Pax/d, but those stations are only a kilometre or so apart. Which got me thinking, what are the biggest flows when measured by Passenger-km?

From the website, I think the following are the biggest, in P-km/day

1. London-Edinburgh (540km): 2730 x 540 = 1.5M P-km/d
2. London-Manchester (260km): 4000 x 260 = 1.04M
3. London-Newcastle (400km): 2000 x 400 = 0.8M
4. London-Leeds (270km): 2540 x 270 = 0.68M
5. London-Birmingham (190km): 3500 x 190 = 0.65M
6. London-Liverpool (290km): 1950 x 290 = 0.56M
7. London-York (280km): 1900 x 280 = 0.53M
8. London-Glasgow (555km): 750 x 555 = 0.42M
9. London-Stansted (50km): 7000 x 50 = 0.35M
10. London-Bristol (165km): 1900 x 165 = 0.31M
11. London-Reading (55km): 5600 x 55 = 0.31M
12. London-Cambridge (75km): 3450 x 75 = 0.26M
13. London-Milton Keynes (65km): 4000 x 65 = 0.26M

Are there any others bigger than these?

I think the biggest non-London flow might be Glasgow-Edinburgh (64km): 3000 x 64 = 0.19M. This is twice the size of Leeds-Victoria (57km): 1400 x 57 = 0.08M, which did surprise me. Hopefully TRU will fix that. Is this data available anyway in a sortable table form?

The biggest surprise (to me at least) was London-Stansted. Is this being neglected in network planning? What surprises you?

(I've measured straight line distance on Google maps. I've not tried to combine stations like Manchesters Victoria, Piccadilly and Airport to get a better city-to-city estimate, as I doubt it would make much difference to the overall rankings. Nor have I tried to include intermediate stations, which would give a better indication of the flows on a particular corridor.)

Website here: https://railalefan.co.uk/labs/flowstats/NR00/

[Edited to add Liverpool and Glasgow; Leeds and York; Cambridge; Reading]

 

[sir_gummerz]

I have nothing to add, just wanted to say thank you for pulling this together, quite an interesting rabbit hole

 

[Nottingham59]

I have nothing to add, just wanted to say thank you for pulling this together, quite an interesting rabbit hole

You're welcome.

Other entries in the top ten are London to Liverpool and to Glasgow. I'll edit for list above

 

[A S Leib]

I've measured straight line distance on Google maps

There's signal.eu.org/osm as a free option (in km) or rail miles as a paid route length checker, although for e.g. London to Birmingham, Northampton means route lengths aren't always exactly the same.

== Doublepost prevention - post automatically merged: 2 Feb 2025 ==

1. London-Edinburgh (540km): 2730 x 540 = 1.5M P-km/d
2. London-Manchester (260km): 4000 x 260 = 1.04M
3. London-Birmingham (190km): 3500 x 190 = 0.65M
4. London-Liverpool (290km): 1950 x 290 = 0.56M
5. London-Glasgow (555km): 750 x 555 = 0.42M
6. London-Stansted (50km): 7000 x 50 = 0.35M
7. London-Bristol (165km): 1900 x 165 = 0.31M
8. London-Milton Keynes (65km): 4000 x 65 = 0.26M

1. London King's Cross – Edinburgh Waverley 2730 × 632 (if all passengers go via York) = 1.73 mn P-km/d / 1.07 mn P-mi/d
2. London Euston – Manchester Piccadilly (0.67 × 295 × 4000) (via Stoke) + (0.33 × 304 × 4000) (via Crewe) = 1.19 mn P-km/d / 0.74 mn P-mi/d
3. London Kings Cross – Newcastle 2000 × 432 = 0.86 mn P-km/d / 0.54 mn P-mi/d
4. London Kings Cross – Leeds 2540 × 299 = 0.76 mn P-km/d / 0.47 mn P-mi/d
5. London Euston – Birmingham New Street 3500 × 183 (avoiding Northampton is 181 km, via 185 km) = 0.64 mn P-km/d / 0.40 mn P-mi/d
6. London Euston – Liverpool Lime Street 1950 × 311 = 0.61 mn P-km/d / 0.38 mn P-mi/d
7. London Kings Cross – York 1900 × 303 = 0.58 mn P-km/d / 0.36 mn P-mi/d
8. London Euston – Glasgow Central 750 × 649 (via the Trent Valley's 646 km and via Birmingham's 658, so that's approximately average) = 0.49 mn P-km/d / 0.30 mn P-mi/d
9. London Liverpool Street – Stansted Airport 7000 × 60 = 0.42 mn P-km/d / 0.26 mn P-mi/d
10. London Paddington – Bristol Temple Meads 1900 × 190 = 0.36 mn P-km/d / 0.22 mn P-mi/d
11. London Euston – Milton Keynes Central 4000 × 80 = 0.32 mn P-km/d / 0.20 mn P-mi/d
12. London Kings Cross – Cambridge 3444 × 93 = 0.32 mn P-km/d / 0.20 mn P-mi/d

 

[Class 170101]

The biggest surprise (to me at least) London-Stansted. Is this being neglected in network planning? What surprises you?

Why do you say its neglected?

4tph Liverpool Street to Stansted
Stansted to Norwich and Stansted to Birmingham services

Not sure what else you are expecting?

 

[Nottingham59]

Why do you say its neglected?

4tph Liverpool Street to Stansted
Stansted to Norwich and Stansted to Birmingham services

Not sure what else you are expecting?

How about Lea Valley four tracking; high-speed trains; more direct route to London instead of a single track tunnel via Stansted Mountfitchet? An hour to cover just 50km on one of the top ten routes in the country is a bit feeble really. Even just longer trains on Birmingham services which are notoriously over-crowded would help.

I'm just struck by the contrast with Manchester Airport, and all the noise about providing a second railway station on a high-speed line.

 

[A S Leib]

London Kings Cross – Cambridge (93 km, 3444 passengers per day) works out at the same to tens of thousands of P-km/d as Euston – Milton Keynes Central.

 

[Mcr Warrior]

What would the figures be for London Victoria + East Croydon?

 

[A S Leib]

What would the figures be for London Victoria + East Croydon?

17 × 4482 = 0.076 mn P-km/d / 0.047 mn P-mi/d

 

[Class 170101]

How about Lea Valley four tracking; high-speed trains; more direct route to London instead of a single track tunnel via Stansted Mountfitchet? An hour to cover just 50km on one of the top ten routes in the country is a bit feeble really.

47 to 48 minutes with 2 or 3 intermediate stops for main part of the day rather than an hour. Not convinced 4 tracking for Lea Valley will happen anytime soon and certainly not due to Stansted. It will be a London led project most likely.

Even just longer trains on Birmingham services which are notoriously over-crowded would help.

Yes but can only be four cars

I'm just struck by the contrast with Manchester Airport, and all the noise about providing a second railway station on a high-speed line.

But this is a two track line from both Manchester Piccadilly and from Wilmslow directions and built at different times? Stansted has one Bi-Di Tunnel to enter and exit it.

 

[A S Leib]

Not convinced 4 tracking for Lea Valley will happen anytime soon and certainly not due to Stansted. It will be a London led project most likely.

Was 4-tracking at least to Cheshunt / Broxbourne supposed to happen as part of Crossrail 2?

 

[87015]

47 to 48 minutes with 2 or 3 intermediate stops for main part of the day rather than an hour. Not convinced 4 tracking for Lea Valley will happen anytime soon and certainly not due to Stansted. It will be a London led project most likely.

Yes but can only be four cars

But this is a two track line from both Manchester Piccadilly and from Wilmslow directions and built at different times? Stansted has one Bi-Di Tunnel to enter and exit it.

47 to 48 mins plus about ten to get through the RPIs and enforced longer exit routes on arrival because its easier to not use some exits and staff up enough for the loadings. Not to mention providing totally unsuitable stock for some diagrams because of they can't get Stadler to meet the required reliability.

Its an awful experience - far worse than Euston - at outrageous cost.

 

[Bald Rick]

How about Lea Valley four tracking; high-speed trains; more direct route to London instead of a single track tunnel via Stansted Mountfitchet? An hour to cover just 50km on one of the top ten routes in the country is a bit feeble really.

47 mins approx for 59km including 2-3 stops, sometimes 4.

How many extra trains would 4 tracking provide? If your answer is more than “0”, I’ll need to see your workings!

Was 4-tracking at least to Cheshunt / Broxbourne supposed to happen as part of Crossrail 2?

Yes, to Broxbourne, from Coppermill Junction.

 

[ruaival]

Insightful piece of analysis - thanks @Nottingham59 and @RailAleFan both - the item that caught my eye was only one major airport and i wonder if Gatwick was a candidate with around 0.37 P-Km/d ?

helpful links to the ORR data on the Flow data source site; answered my first question of does EDB-LKX include EDB-NCL and its does not, ORR says point to point ticket sales are the starting data they use.

 

[Magdalia]

Website here: https://railalefan.co.uk/labs/flowstats/NR00/

London Liverpool Street – Stansted Airport 7000 × 60 = 0.42 mn P-km/d / 0.26 mn P-mi/d

London Kings Cross – Cambridge 3444 × 93 = 0.32 mn P-km/d / 0.20 mn P-mi/d

Some clarification would be helpful here please.

Are the flows numbers 2022/23 or 2023/24? I can only find the former on the link provided.

Does Stansted Airport include passengers to/from Tottenham Hale?

Does Cambridge include Thameslink stations, or Cambridge North?

 

[A S Leib]

Are the flows numbers 2022/23 or 2023/24? I can only find the former on the link provided.

Does Stansted Airport include passengers to/from Tottenham Hale?

Does Cambridge include Thameslink stations?

I used 2022-3 as that's what railalefan.co.uk currently uses; https://www.railwaydata.co.uk/odm/gbr/ has 2023-4. As the OP said in post 1 that they were just counting station to station rather than city to city, so I haven't counted e.g. Kings Cross to Cambridge North, Cambridge to City Thameslink or Tottenham Hale to Stansted either.

 

[Nottingham59]

Does Stansted Airport include passengers to/from Tottenham Hale?

Does Cambridge include Thameslink stations?

The ORR data is just single station to single station.

So for example, Gatwick to Victoria is 6000 passengers per day. While Gatwick to all the biggest stations in Central London total around 12000 per day. But even so, 12000 x 39km = 0.47 P-km/d, so a bit bigger than London-Stansted but still much smaller than London-Edinburgh.

So my simple table tends to break down when there are multiple stations in the cities at either end of the flow. It looks like flows could be out by up to a factor of 2.

EDIT: It looks like Stansted is surprising because there is only one destination station in central London. Unlike Gatwick with connections to Clapham Jn; Victoria; St Pancras; Farringdon; Blackfriars; London Br; or Heathrow with direct service to everywhere on the Elizabeth Line. (Note I'm happy to discuss London-Stansted, but preferably in a separate thread.)

 

[Magdalia]

I used 2022-3 as that's what railalefan.co.uk currently uses; https://www.railwaydata.co.uk/odm/gbr/ has 2023-4.

It is preferable to use the most recent data, except where there is a good statistical reason not to.

Stansted is surprising because there is only one destination station in central London.

In practice Tottenham Hale is also a Central London destination for Stansted, because airport passengers use Tottenham Hale and the Victoria Line to get to/from the West End. This predates the airport, it has been a feature of Lea Valley services ever since the Victoria Line opened. For Stansted Airport, if Liverpool Street and Tottenham Hale are taken together, the split is roughly 75/25 between Liverpool Street and Tottenham Hale. That isn't very different from the Victoria/London Bridge split for Gatwick Airport.

 

[GRALISTAIR]

I have nothing to add, just wanted to say thank you for pulling this together, quite an interesting rabbit hole

+1 from me too. Fascinating.

 

[deltic]

Recalculated on the 2023-24 data set and using all city to city flows - the numbers are annual passenger miles - some of the distances may be a bit suspect as they come from AI. All the flows are to and from London

City of Edinburgh	371,226,860
Manchester	290,274,223
Newcastle upon Tyne	207,600,027
Gatwick Airport	186,348,825
Brighton and Hove	175,686,720
Birmingham	174,792,822
Leeds	165,199,540
Glasgow City	161,943,449
Liverpool	154,764,770
York	142,619,244
Stansted Airport	128,535,627
Cambridge	117,713,680
Milton Keynes	97,641,164
Bristol, City of	86,863,844
Heathrow	86,052,825
Sheffield	81,337,440
Peterborough	81,277,974

 

[Magdalia]

Recalculated on the 2023-24 data set and using all city to city flows

Thanks, that was impressively speedy!

The most significant addition is Brighton and Hove. It is interesting that Blackfriars has much bigger flows to/from Brighton since the pandemic: I'm wondering if that's a real thing or a change in the methodology?

 

[deltic]

Edinburgh to Glasgow is 83.7m pass miles, Manchester to Liverpool is 42m and to Leeds is 30m.

 

[Nottingham59]

Recalculated on the 2023-24 data set and using all city to city flows - the numbers are annual passenger miles - some of the distances may be a bit suspect as they come from AI. All the flows are to and from London

Thank you. That's very helpful. City to city flows are more informative, I think.

Interesting to note that total GB passenger miles were 10 Billion in 3Q2024, according to ORR, which is 40Bn per year. So even these top 17 city to city flows account for only a small fraction (leas than10%) of the total Passenger miles delivered.

 

[Topological]

An interesting element of this is how many of the flows are WCML.

Almost like that would be the best route for High Speed

 

[Nottingham59]

An interesting element of this is how many of the flows are WCML.

Almost like that would be the best route for High Speed

Especially if that high speed route included a spur off to Leeds ....

And of course, Edinburgh will be fastest via the West Coast, if/when they build phase 2a.

 

[Bald Rick]

Recalculated on the 2023-24 data set and using all city to city flows - the numbers are annual passenger miles - some of the distances may be a bit suspect as they come from AI. All the flows are to and from London

City of Edinburgh	371,226,860
Manchester	290,274,223
Newcastle upon Tyne	207,600,027
Gatwick Airport	186,348,825
Brighton and Hove	175,686,720
Birmingham	174,792,822
Leeds	165,199,540
Glasgow City	161,943,449
Liverpool	154,764,770
York	142,619,244
Stansted Airport	128,535,627
Cambridge	117,713,680
Milton Keynes	97,641,164
Bristol, City of	86,863,844
Heathrow	86,052,825
Sheffield	81,337,440
Peterborough	81,277,974

Good info.

I’m a bit surprised St Albans isn‘t on that list. There’s 6.3m passengers a year, and I reckon 4.5-5m will be to London, at 20-23 miles each.

 

[Topological]

Especially if that high speed route included a spur off to Leeds ....

And of course, Edinburgh will be fastest via the West Coast, if/when they build phase 2a.

Such a spur would then hoover up York and Newcastle.

It makes you think though, that all of this would be delivered on a 2-track line with a 10-platform Euston. Efficiency of land use is an understatement.

 

[Magdalia]

Good info.

I’m a bit surprised St Albans isn‘t on that list. There’s 6.3m passengers a year, and I reckon 4.5-5m will be to London, at 20-23 miles each.

If I'm doing my sums correctly, all of the core stations St Pancras-London Bridge add up to just under 4 million.

Another that I think comes into play is Chelmsford, especially if Stratford passengers are included. Chelmsford is almost 30 miles from Liverpool Street and has more than 2.8 million passengers per year between the two, plus another 1 million Chelmsford-Stratford.

These are the ORR two way flows.

 

[Verulamius]

If I'm doing my sums correctly, all of the core stations St Pancras-London Bridge add up to just under 4 million.

Another that I think comes into play is Chelmsford, especially if Stratford passengers are included. Chelmsford is almost 30 miles from Liverpool Street and has more than 2.8 million passengers between the two, plus another 1 million Chelmsford-Stratford.

Might need to include West Hampstead for a change for the Jubilee to the West End.

 

[Krokodil]

I'm just struck by the contrast with Manchester Airport

Manchester Airport largely acts as a very long turnback siding. It's somewhere you send trains to, to keep them out of the way of P13/14