Hidden Vulnerability Discovered in the World's Airline Network

I think the paper is incorrect. According to the summary:

"An example is St Petersburg airport in Tampa Bay, Florida. This has 24 connections and only 24 flights. Close this airport and the airports it connects to are all cut out of the global network."

St. Pete/Clearwater (PIE) airport is a destination airport for low-cost carriers. http://en.wikipedia.org/wiki/St._Petersburg%E2%80%93Clearwat... lists some of the places it connects to. I count 34 destinations on the map.

#1 with 27,000 passengers is Gulfport-Biloxi International (GPT). According to http://en.wikipedia.org/wiki/Gulfport-Biloxi_International_A... , 147,910 fly from GPT to Atlanta.

#10 with 18,000 passengers is Blue Grass Airport (LEX). According to http://en.wikipedia.org/wiki/Blue_Grass_Airport, there are 139,460 passengers who fly from LEX to Atlanta.

In fact, all of #1-#10 also fly to Atlanta, except Tri-State Airport in West Virginia. That one has flights toDetroit Metro. I spot checked a couple of others not on the top-10 list to verify that there aren't 24 disconnected airports in the list of destinations.

Thus, I don't see 24 connections which would be cut-off should PIE be closed.

To check if the summary made a mistake, the preprint says:

> Most airports in the first category have really few connections but some of them are busy airports having a significant number of connections to and from cut-off regions, and hence we call them peripheral hubs. For instance, St. Petersburg Airport, Tampa Bay, Florida, is a huge tourist destination consisting of 24 connections and no clustering.

Tampa Bay is a huge tourist destination, agreed. But most people fly into Tampa airport, which is 16 km to the east. Tampa deals with some 17 million passengers per year, while PIE is around 900,000.

I just don't see how the St Petersburg airport is so important.