Case Study Calculations (WIP)

[dodedic]

@michaelweinold @arebe337 I have made some calculations for the case study based on literature found online. Have a look at them and let me know if you think some of my assumptions are ill-considered.

My basic thought process behind the case study is the following:

1. Identify a city pair with strong air transport demand and potential for HSR (done)
2. "Verify" route against exisiting/proposed projects (done)
3. Look at existing HSR projects for average construction and maintenance costs for a Life Cycle Cost estimation. (below)
4. Estimate current and future demand of air travel route between both cities. (below)
5. Estimate fuel and CO2 numbers based on aircraft types and calculate associated CO2 cost. (below)
6. Look at exisiting competing HSR & Air travel routes for numbers on market share between the two travel modes. (below)

3. Look at existing HSR projects for average construction and maintenance costs for a Life Cycle Cost estimation.

Here I found two very interesting papers. One is this this internal report from the Australian Government where they assess funding experiences from other countries and projects for their own. Here they found the following costs of counstruciton per km for different projects:

The UK project stands out since it involved a lot of tunneling, which makes the cost per km a lot higher. If we take an average cost of the other eleven construction projects we got to around 30Mio EUR/km, when we adjust for inflation from 2004 to now we get to around 47.7Mio EUR/km.

From the Life Cycle Assessment of the Chinese Beijing-Shanghai HSR we could take some data from this paper which outlines that over a 100 year life cycle just the construction cost make up around 80% of the total cost of the project, while around 19% are operation and maintenance cost, the other 1% are conception and disposal.

For our example of Lagos to Abuja we would then take the distance from Bastian's model of 546 km and multiply it by the cost per km to give us a total construction cost of: 26'044 Mio EUR = 26 Bn EUR

This estimation of 26 Bn EUR sits right around the costs of the Beijing-Shanghai project of 34.7 Bn (for 1300km) and the LA-San Francisco initial budget of 40Bn (now rising). I did not multiply by Bastian's cost factor per km of 1.79 since the averaged out price of the other railway projects already include cost variables like terrain in the per km price, which is what Bastian's cost factor is supposed to do. I took this approach since I couldn't get a price for "HSR tracks on flat ground per km" which I could then multiply with Bastian's cost factor.

Taking the percentage estimate over a 100 year life cycle from above we get a total operations & maintenance cost of around 65M per year. ((26Bn / 4)/100)

4. Estimate current and future demand of air travel route between both cities.

Since our friends at ADSBExchange are taking their time answering my E-Mail I took a slightly different approach to estimating the PAX numbers on the flights between Abuja and Lagos.

From Sabre/OAG (this website only links to the OAG site, but they are gatekeeping the file with the actual data) we know that in 2023 the route between Lagos-Abuja had 1'265'267 PAX (PAX*504km=637'694'568 RPK). From our AeroDataBox API we know that the route Lagos->Abuja has around 15-16 daily flights, and the route Abuja->Lagos also has around 15-16 daily flights. Meaning a total of around 30-32 flights per day on this route. If we now divide the number of yearly PAX by the number of yearly flights we get: 1'265'267 PAX / 10950 flights = 115 PAX/flight

Now from FlightRadar24 I looked at all the types of aircraft that are flying this route regularly and looked at their capacities to make sure if these numbers could accurately represent the actual flights.

They have the following capacities: A320 ~150 PAX B733 ~135 PAX B737 ~140 PAX BCS3 ~130 PAX CRJ9 ~ 90 PAX E145 ~ 50 PAX If we assume more or less the same distribution of flights across these aircraft we actually get an average seats available of 115. So I would say this estimation is in the right ball park, no?

Forecast on PAX based on GDP We take the data from the IMF which we have until 2028 (actually from an Excel Table but here visually) and if we assume constant growth after that of around 3% we would get the following numbers for PAX scaling. Simply multiplying the growth factor of 3% with every years PAX numbers. Assuming a 2:1 correlation as stated by IATA as the median over the last 20 years:

The Nigerian government itself is targeting a growth rate of 7%.

5. Estimate fuel and CO2 numbers based on aircraft types and calculate associated CO2 cost.

Done on paper, transfer onto Github Thursday

6. Look at exisiting competing HSR & Air travel routes for numbers on market share between the two travel modes.

This study from China between 2011-2016 found a 45% drop of flight frequency on routes where HSR was made available on a distance between 300-500km (Lagos-Abuja = 504km). China is a mature HSR market with strong city pair connections. From this paper a clear substitution effect is evident.

Another study relates market share to travel time across multiple countries in Europe:

[arebe337]

Looks really good to me! Just one question concerning the data from our AeroDataBox API: Did you take the data from only one month or did you look at it for the whole year? And what is your approach for the GDP scaling since I wanted to start with it now for the global map?

[dodedic]

Looks really good to me! Just one question concerning the data from our AeroDataBox API: Did you take the data from only one month or did you look at it for the whole year? And what is your approach for the GDP scaling since I wanted to start with it now for the global map?

I averaged across the year for the flights to make it more representative. And for now I have just scaled with the yearly GDP growth rate forecast numbers we have of the map with all the countries until 2029, the next step is then to scale it with a more general number we have for Africa unti 2050, correct?

[arebe337]

I averaged across the year for the flights to make it more representative. And for now I have just scaled with the yearly GDP growth rate forecast numbers we have of the map with all the countries until 2029, the next step is then to scale it with a more general number we have for Africa unti 2050, correct?

Exactly, that's the concept. However, what I was getting at is how we can effectively scale the GDP with the flights. We previously discussed the need to address this issue because we understand that GDP correlates with passenger numbers but not necessarily directly with the number of flights.

[dodedic]

I averaged across the year for the flights to make it more representative. And for now I have just scaled with the yearly GDP growth rate forecast numbers we have of the map with all the countries until 2029, the next step is then to scale it with a more general number we have for Africa unti 2050, correct?

Exactly, that's the concept. However, what I was getting at is how we can effectively scale the GDP with the flights. We previously discussed the need to address this issue because we understand that GDP correlates with passenger numbers but not necessarily directly with the number of flights.

Ahh I see the confusion, but in this case right here I do have the PAX numbers though, so I just scaled those.

[michaelweinold]

Estimate fuel and CO2 numbers based on aircraft types and calculate associated CO2 cost. Done on paper, transfer onto Github Thursday https://github.com/sustainableaviation/demandmap/issues/30#issue-2298572973

How are you planning to compute fuel consumption ($\sim CO_2$) for different aircraft at scale? Perhaps it would be best to use the data I calculated with my last master's student. See the column EU (MJ/ASK): https://github.com/sustainableaviation/Aircraft-Performance/blob/main/Databank.xlsx

TBC: Forecast on PAX based on GDP Done on paper, transfer onto Github Thursday https://github.com/sustainableaviation/demandmap/issues/30#issue-2298572973

Can you show me the equations?

[dodedic]

Can you show me the equations?

Added to the initial comment! Rest I am still working on.

[michaelweinold]

Simply multiplying the growth factor of 3% with every years PAX numbers. https://github.com/sustainableaviation/demandmap/issues/30#issue-2298572973

So you're using a 1:1 correlation?

[dodedic]

So you're using a 1:1 correlation?

So sorry, I was a bit too eager there hahaha! Actually a 2:1 correlation would be the 20-year median as stated by IATA

Also this graph from 2017:

[michaelweinold]

This is for global traffic? Was the idea from your excellent literature review Excel table not that for developing countries, the correlation is higher?

[dodedic]

I think the main takeaway from that literature review was the causality aspect of the GDP/RPK connection, not the actual correlation factor between the two. So I thought it might be wise to stick to a broader more global perspective.

If we want to zoom in on Nigeria though I can cite this paper from said literature review which states the following on causality again:

The calculations are in the paper itself. Unfortunately no stipulation on correlation factor is made.

[michaelweinold]

Following up on our catch-up call today:

Data Quality Check on Passenger Numbers

Estimate current and future demand of air travel route between both cities. https://github.com/sustainableaviation/demandmap/issues/30#issue-2298572973

This seems like a reasonable approach. Although in the original comment above you had not yet used the new API calls to get aircraft types and seats. Ultimately, we want to compare our numbers on pax/year on a specific route to other data sources. That could be FlightRadar24, or Sabre/OAG, etc. data.

Estimate of Carbon Emissions

As discussed, you can use a metric like (average) Energy Usage (EU[MJ/ASK]), but this calculation can be made almost arbitrarily complicated. We could consider higher fuel burn during takeoff/landing cycles, etc. In this case, fuel burn per ASK would be higher on short routes... remember the figure from my lecture?

For our poster, it might indeed make sense to use a simple approximation like (average) EU.

Please try out the ICAO carbon emissions calculator and compare it against EU.

[dodedic]

I think for our purposes as you said @michaelweinold a simple calculation will do. With the ICAO model being a bit more sophisticated (and perhaps more up to date) than the figure, that's what I chose going forward.

When calculating with a distance of around 510 km from Abuja to Lagos and an average available seats of 140 as per our matrix (we take seats as PAX since for the actual amount of CO2 it doesnt play a huge difference in their model), if we take the ICAO carbon emissions calculator we get an emission of around 85 kg CO2/Seat/leg or 11'900 kg = 11.9 tons of CO2 per leg.

With an average of 16 daily flights from DNMM-DNAA and another 16 flights back, we get around 32 daily flights on this route from this API call.. Over one year we then have 32 flights/day*'365 days = 11'680 flights.

Giving us a total carbon emission of 138'992 tons of CO2 for this city pair per year.

If we want to subsidize the running of the HSR network in Abuja which we would estimate at a running cost of 65M EUR per year (as seen in my inital comment in this issue) with solely this air transport route, this would put us at a price of 467 EUR/ton of CO2.

Now it would make a lot more sense to consider all flights from the major city airports in Nigeria for this calculation, so let's say Abuja and Lagos as the biggest airports. We could tax all flights leaving those airports to get an estimate of a lower, more realistic price per ton of CO2. This I can do tomorrow.