WASHINGTON (AP) -- The city of light is also the city of flight, says a new study listing Paris as the most important connecting point for international air travel.

Anchorage, Alaska, places a surprising second on the list, followed by London, Singapore and New York. The ranking reflects each cities' importance in the world's air network - which is not the same as the busiest cities for connections.

The connections among 3,883 communities with airports around the world were analyzed by a team of researchers led by Luis Amaral of Northwestern University. The results appear in Tuesday's issue of Proceedings of the National Academy of Science.

The air transportation network is like the Internet, the study concluded, with networks and hubs funneling traffic around the world.

The findings are important in understanding the flow of travelers and in studying the potential movement of new diseases, Amaral said.

In addition, the analysis could help regulators determine airports where more competition is needed. And study of the network could even shed light on the functions of biological networks within the human body, according to Amaral, an associate professor in the Department of Chemical and Biological Engineering.

A traveler can get from any of the cities to any other with an average of 4.4 flights, and more than half the communities are connected with four flights or fewer, the researchers found.

The most difficult air route? Getting from Mount Pleasant in the Falkland Islands to Wasu, Papua-New Guinea, requires 15 flights.

The researchers found the busiest locations are not always the most important.

Anchorage, for example, has nonstop flights to 39 other cities, far fewer than the 242 cities connected to London nonstop.

But Anchorage edges London in a vital measure called centrality -- a rating of the shortest paths connecting any two cities that involve a transfer at a particular city.

It's centrality that boosts the importance of cities such as Anchorage and Port Moresby, Papua-New Guinea, which serve as connection hubs between many other airports and international connections, the researchers explained.

Alaska, for example, has many airports, but most connect only to other Alaskan airports. Only a few connect to the "lower 48" states. There are political constraints on flights directly between most of Alaska and Canada, even to cities close to the border.

Thus, getting from most of Alaska to somewhere outside the state often involves going through Anchorage, boosting that city's centrality rating.

Similarly, many Pacific islands are connected by air, and Port Moresby is the hub that links lots of them to the outside, placing that community seventh on the worldwide list. That ranks behind Los Angeles but ahead of such busy places as Frankfurt, Tokyo and Moscow.

Paris and London benefit from their nation's colonial pasts, with many flights from Africa and Asia going to those cities, where travelers transfer to other planes to go on.

Indeed, they are the top cities in the world for nonstop flights to other places. Paris leads with flights to 250 other cities, followed by London, 242; Frankfurt, 237; Amsterdam, 192 and Moscow, 186.

The two busiest airports in the U.S. are in Chicago and Atlanta. The study ranked Chicago 13th on the worldwide centrality list and 6th for nonstop flights, with connections to 184 cities. Atlanta ranked 29th for centrality and 8th in connections, with flights to 172 cities.

The study analyzed 531,574 flights operated by 800 airlines worldwide from November. 1 to November. 7, 2000. While the data are four years old, the researchers say the current worldwide airport network is virtually identical to the one at that time.

The research was funded by the National Institutes of Health.

Amsterdam has more connections than Chicago.

I read just yesterday that HK currently has 140, and aiming firmly at 180 by 2010.

Very very interesting.

Interesting.

Kind of weird for O'Hare and Hartsfield to be so low on the list.

Kind of weird for O'Hare and Hartsfield to be so low on the list.

Not really. The United States is a huge domestic market. European cities benefit from having a ton of countries in a similar sized area.

Paris?..Is that guy who lives in that Airport for more than 14 years still there?..:laugh:

Very interesting when I read the report in my local papers. Went ahead to get the article from Proceedings of the National Academy of Science (http://www.pnas.org) and got this centrality table:
Table 2. The 25 most central cities in the worldwide air
transportation network

1 Paris 58.8
2 Anchorage 55.2
3 London 54.7
4 Singapore 47.5
5 New York 47.2
6 Los Angeles 44.8
7 Port Moresby 43.4
8 Frankfurt 41.5
9 Tokyo 39.1
10 Moscow 34.5
11 Seattle 34.3
12 Hong Kong 30.8
13 Chicago 28.8
14 Toronto 27.1
15 Buenos Aires 26.9
16 Sao Paulo 26.5
17 Amsterdam 25.9
18 Melbourne 25.5
19 Johannesburg 25.4
20 Manila 24.4
21 Seoul 24.3
22 Sydney 23.1
23 Bangkok 22.9
24 Honolulu 21.1
25 Miami 20.1Due to copyright issues, I cannot reproduce the entire text here, but I should be able to quote parts of it if you guys wish to know more! :D

Notice the different reasons why a lot of these cities make it on the list. It makes sense that Paris and London feature prominently because of their colonial past. I saw a lot of different varieties of aircraft at both Heathrow and Charles de Gaulle. Frankfurt was also up there, which is quite interesting since their colonial reach was nowhere to the extent of the British and French empires.

Anchorage is a major stopover point for transpacific traffic between North America and Asia. Many cargo carriers stop there, while I've spotted Korean Air and Cathay Pacific at Anchorage Airport before.

Many American cities are major domestic hubs, which seem to be overlooked in the scope of this study. The aviation market in the US is much different from the rest of the world. Transferring flights at hub cities is quite common in the US whereas in Asia, for example, direct traffic is more popular.

Hmm.....the elaboration on Anchorage's high ratings in the actual report seems to be concentrated on factors quite different from what was speculated above.

Meanwhile, any explainations on Singapore's rankings, especially in comparison to other Asian cities?

Anchorage's situation is similar to neighboring Canada. Much of the international traffic out of Canada departs from Vancouver and Toronto. Hence people living in between will have a higher likelihood of taking a connecting flight to one of these hubs. There are limited point-to-point services between smaller cities and international destinations.

This pattern is quite similar in the United States as well. Some airlines were fuming that the recent China - US bilateral air services agreement favoured more flights out of New York than building new connectivities from cities previously not served with a direct transpacific route.

There is a whole branch of mathematics that deals with connectivity and how airlines route and schedule their flights through various hubs.

However, connectivity is not only related to passengers, but also cargo. Many cargo carriers, such as UPS and Fedex, stop at Anchorage. Cargo traffic didn't seem to be included in this study, which is quite incomplete to fully assess why certain airports are very well connected. Anchorage has a significant cargo operation.

There is also another caveat in analyzing these studies. Notice not many of the top connectivity airports are the busiest, and even those at the top are not necessarily well connected at all. For example, Anchorage is far less connected internationally than New York. In Asia, the busiest airports are Tokyo Narita, Bangkok, and Hong Kong.

Anchorage's situation is similar to neighboring Canada. Much of the international traffic out of Canada departs from Vancouver and Toronto. Hence people living in between will have a higher likelihood of taking a connecting flight to one of these hubs. There are limited point-to-point services between smaller cities and international destinations.

This pattern is quite similar in the United States as well. Some airlines were fuming that the recent China - US bilateral air services agreement favoured more flights out of New York than building new connectivities from cities previously not served with a direct transpacific route.

There is a whole branch of mathematics that deals with connectivity and how airlines route and schedule their flights through various hubs.

However, connectivity is not only related to passengers, but also cargo. Many cargo carriers, such as UPS and Fedex, stop at Anchorage. Cargo traffic didn't seem to be included in this study, which is quite incomplete to fully assess why certain airports are very well connected. Anchorage has a significant cargo operation.

There is also another caveat in analyzing these studies. Notice not many of the top connectivity airports are the busiest, and even those at the top are not necessarily well connected at all. For example, Anchorage is far less connected internationally than New York. In Asia, the busiest airports are Tokyo Narita, Bangkok, and Hong Kong.
A few questions:

1. Have you read the actual report?

2. If "Anchorage's situation is similar to neighboring Canada", then can you explain why Toronto is much lower down the list, and Vancouver is nowhere in sight?

3. The above report was indeed based on mathematical tools, if that was not obvious by now. I am just wondering if, given your claims of your mathematical background, if you may give us a analysis of the above even without the aid of the actual report?

4. In what ways are cargo connectivity comparable to passenger connectivity, and in what ways are they dissimilar? To what extend do this have an impact on overall airport connectivity study?

5. You pointed out that many of the world's busiest airports (only in terms of passengers, I presume) are not high on "connectivity". This was, of coz, not entirely original, although it looks a tad misunderstood. This study tries to measure "importance" in terms of "centrality", and seeks to demonstrate and explain, if any, disparaties between "centrality", "connectivity" and total passenger volumes, all distinct and seperate definitions which are not actually "re-defined" by this study. So they main point of interest was not actually about connectivity being disproportionate to passenger volumes. Its more to do with centrality's dispararity with connectivity or total passenger volumes. And since we are on this, could you offer us an explaination or your views on this disparatiy?

6. You have not answered my question related to Singapore, have you?

Due to copyright issues, I cannot reproduce the entire text here, but I should be able to quote parts of it if you guys wish to know more! :D

Is there a definition for centrality?? :?

Meanwhile, any explainations on Singapore's rankings, especially in comparison to other Asian cities?

Singaporean aviation officials are intelligent and understand the concept of being an aviation hub is simply more than funneling as many people down the runway as possible.
Ive said it once and ill say it again, Singapore is the world's greatest economic experiment and we all stand to learn a great deal form it, economically, not socially unfortunately.

The most difficult air route? Getting from Mount Pleasant in the Falkland Islands to Wasu, Papua-New Guinea, requires 15 flights.

lol!
I wish they could give details of the flights, and how long it takes!

A few questions:

1. Have you read the actual report?

2. If "Anchorage's situation is similar to neighboring Canada", then can you explain why Toronto is much lower down the list, and Vancouver is nowhere in sight?

There is a lot of direct point-to-point traffic between Canadian cities. For example, residents of Calgary don't need to connect in Vancouver to reach Toronto, but they do if they want to fly to Hong Kong or Tokyo. In fact, Canada's large cities are well-connected via direct domestic flights, whereas international flights require a hub connection system due to economic feasibility. Alaska's unique location makes Anchorage's situation quite interesting. Many people heading to the lower 48 (domestic flights) will need to connect through Anchorage while a lot of point-to-point international traffic require technical stops in Anchorage. Toronto and Vancouver are far larger economic centers than Anchorage with much bigger populations, so it's not surprising they have more proportionate end-to-end traffic. That's a major difference between the hub and connecting traffic between Anchorage, Toronto, and Vancouver. Although the three are major connecting hubs, Anchorage is far more dependent on connecting flights than either Toronto or Vancouver - hence the ranking.

Understanding the context of each city's market is important in assessing what the survey actually means rather than just merely reading the report and not knowing the forces that drive the results.

3. The above report was indeed based on mathematical tools, if that was not obvious by now. I am just wondering if, given your claims of your mathematical background, if you may give us a analysis of the above even without the aid of the actual report?

It's part of network flow theory. Although it is unlikely a flight from city X to city Y may go through the most optimal route, airlines schedule their flights using complex mathematical formulas that consider cost equations and revenue streams. However, a lot of the data is highly secretive and heavily guarded by airlines, so I was quite surprised that this organization was able to assemble enough to make an analysis. In fact, some airlines have gone to the courts for industrial espionage over their prized traffic and scheduling data.

There are complications because cities X and Y may not be connected through one airline only. For example, there are no direct flights between Canada and Mongolia. However, it is possible to connect through a third country with another airline. So the problem becomes how to reconcile the optimal paths of two (or perhaps more) airlines with two different geographical scopes and network flows. Winnipeg and Ulan Bataar might have an optimal flight path, but it will not be taken for sure due to economic and political limitations.

4. In what ways are cargo connectivity comparable to passenger connectivity, and in what ways are they dissimilar? To what extend do this have an impact on overall airport connectivity study?

Cargo moves in a much different way than passengers. Many cargo flights route through multiple cities before reaching their final destination. Some of the airports cargo flights land in may not be big passenger hubs either. However, movement of goods has become a contentious issue especially as globalization becomes more widespread. It is no longer sufficient to analyze the flow of people since billions of goods are imported and exported daily.

For example, during SARS, cargo traffic was not affected at all in Hong Kong even though passenger traffic fell drastically.

Anchorage is a major cargo hub. UPS and Fedex serve the airport at high frequencies on a daily basis. This is quite interesting given Anchorage is not a major point-to-point origin traffic centre due to its small population. Air cargo traffic patterns deviate from typical passenger traffic with the world's busiest cargo airport being Memphis, home of Fedex. Anchorage is 4th in the world in terms of cargo traffic, confirming my explanation of Anchorage's cargo hub role.

5. You pointed out that many of the world's busiest airports (only in terms of passengers, I presume) are not high on "connectivity". This was, of coz, not entirely original, although it looks a tad misunderstood. This study tries to measure "importance" in terms of "centrality", and seeks to demonstrate and explain, if any, disparaties between "centrality", "connectivity" and total passenger volumes, all distinct and seperate definitions which are not actually "re-defined" by this study. So they main point of interest was not actually about connectivity being disproportionate to passenger volumes. Its more to do with centrality's dispararity with connectivity or total passenger volumes. And since we are on this, could you offer us an explaination or your views on this disparatiy?

Centrality has a lot to do with network flow theories, such as the shortest di-path problem. Interestingly, some cargo carriers think of their routes more as flows (circulation paths) than paths like passenger traffic. I don't expect cities with high connecting traffic to necessarily be busy hubs or well-connected hubs. Atlanta is a good example. It is a major domestic US hub with a lot of traffic, but it is far less connected internationally than New York. This disparity has a lot to do with solving the shortest di-path problem using economic constraints.

However, it's very important to analyze the purposes of connecting traffic in order to assess whether it's a good thing or not. If two cities are not able to sustain direct flights economically , then a stopover may naturally be a good thing because you can connect three cities that otherwise may not gain that route. In this case, disparity is a good thing.

Although this trend is decreasing, some flights require stopovers and connections because of technical constraints with the modern airplane. The kangaroo route is a prime example. As Airbus and Boeing introduce new longer-range aircraft, technical stopovers may be less prevalent now than a decade ago. An example is Air Canada's new Hong Kong - Toronto nonstop using the A340. In the past, a nonstop route was only possible on Cathay Pacific heading eastbound in the winter months due to technical reasons. A technical stop does very little economic benefit to the connection city, since passengers cannot get out of the airport to shop, eat, and sightsee.

Interestingly, Vancouver is both a major connecting hub in Western Canada and also a direct point-to-point airport. Many Asian airlines only fly into Vancouver and no other Canadian city, such as Air China and EVA, and passengers transfer to other carriers to reach other Canadian cities. At the same time, other carriers fly directly point-to-point to Vancouver, such as Cathay Pacific to Hong Kong and British Airways to London. It is also a technical stop for other North American services.

Politics may also influence these rankings. As the report mentioned about Anchorage's role being a connecting city even if direct routes are shorter, political agreements have a very powerful impact on how network flow theories don't always work perfectly in real life. There was quite an uproar when the US government forced transit passengers to get US-entry visas even for technical stops at Anchorage, and routes were changed to adapt to the new regulation.

Hence it's not sufficient to merely read a report and quote some sentences. There is no value if the researcher cannot tie it back to market dynamics, how airlines operate, and why they do so. This ignorance can easily be cured with academic course work, personal interest in the field, and an open mind to find out more.

So instead of asking me whether I've read the report, I now ask you how capable you are to interpret the results? By the way, I have access to the works of the Proceedings of the National Academy of Science through my research memberships.