Case Study

In space there is no definite location, only in relation to everyone else in the space — you are located by who you are connected to.

September 11, 2001 event – Uncloaking Terrorist Networks

Source: Valdis Krebs (http://orgnet.com)

Data Gathering – Within one week of the attack, information from the investigation started to become public. We soon knew there were 19 hijackers, which planes they were on, and which nation’s passports they had used to get into America. As more information about the hijackers’ past was uncovered the investigators decided to map links of three strengths (and corresponding thickness). The tie strength would largely be governed by the amount of time together by a pair of terrorists. Those living together or attending the same school or the same classes/training would have the strongest ties.

Those traveling together and participating in meetings together would have ties of moderate strength and medium thickness. Finally, those who were recorded as having a single transaction together, or an occasional meeting, and no other ties, they classified as weak ties that were shown with the thinnest links in the network.

The investigators started their mapping project upon seeing several summaries of data about the hijackers in major newspapers (Sydney Morning Herald, 2001; Washington Post, 2001). These data collections contained information about the nodes/hijackers and their links/relationships.

From two to six weeks after the event, it appeared that a new relationship or node was added to the network on a daily basis. Several false stories appeared about a cell in Detroit. These stories, originally reported with great fanfare, were proven false within one week. This made the investigators very cautious about adding a link or a node to the network.

The network was created iteratively as data became available. Everyday the investigators checked the major news sources for updated information. Figure “1” shows their computer screen during this process. The browser window shows the news story, the other window shows the network mapping and measuring software. They would add nodes and links to the map as they read the news accounts. Figure “1” shows a link being added between one of the hijackers and an accomplice.

Figure 1

By the middle of October enough data was available to start seeing patterns in the hijacker network. Initially, the investigators examined the prior trusted contacts – those ties formed long ago through living and learning together. The network self-organized (via a network layout algorithm) into the shape of a serpent – how appropriate, they thought.

Figure 2

The investigators were amazed at how sparse the network was and how distant many of the hijackers on the same team were from each other. Many pairs of team members were beyond the horizon of observability from each other – many on the same flight were more than two steps away from each other. A strategy for keeping cell members distant from each other, and from other cells, minimizes damage to the network if a cell member is captured or otherwise compromised. Usama bin Laden even described this plan in his infamous videotape, which was found in Afghanistan. In the transcript (U.S. Department of Defense, 2001) Usama bin Laden mentions: “Those who were trained to fly didn’t know the others. One group of people did not know the other group.”

The network metrics for the network in Figure “2” are found in Table “1”. For a small network of less than 20 nodes, we see a long average path length of 4.75 steps. Several of the hijackers are separated by more than 6 steps. From this metric and bin Laden’s comments above we see that covert networks trade efficiency for secrecy.

Table 1: Small-World Network Metrics

Clustering Coefficient Average Path Length
Contacts 0.41 4.75
Contacts + Shortcuts 0.42 2.79

Yet, work has to be done, plans have to be executed. How does a covert network accomplish its goals? Through the judicious use of transitory shortcuts in the network. Meetings were held that connected distant parts of the network to coordinate tasks and report progress. After coordination was accomplished, the cross-ties went dormant. One well documented meeting of the hijacker network took place in Las Vegas. The ties from this and other meetings are shown in gold in Figure “3”.

Figure 3

Six (6) shortcuts were added to the network temporarily in order to collaborate and coordinate. These shortcuts reduced the average path length in the network by over 40% thus improving the information flow in the network – see Table “1”. When the network is brought closer together by these shortcuts, all of the pilots ended up in a small clique – the perfect structure to efficiently coordinate tasks and activities. There is a constant dynamic between keeping the network hidden and actively using it to accomplish objectives.

The 19 hijackers did not work alone. They had other accomplices that did not get on the planes. These co-conspirators were conduits for money and also provided needed skills and knowledge. Figure “4” shows the hijackers and their network neighborhood – their direct and indirect associates.

Figure “4”

After one month of investigation it was ‘common knowledge’ that Mohamed Atta was the ring leader of this conspiracy. Again, bin Laden verified Atta’s leadership role in the video tape (U.S. Department of Defense, 2001). Looking at the diagram he has the most connections. In Table “2” we see that Atta scores the highest on all network centrality metrics – Degrees, Closeness, and Betweenness.

The network metric Degrees reveals Atta’s activity in the network. Closeness measures his ability to access others in the network and monitor what is happening. Betweenness shows his control over the flow in the network – he plays the role of a broker in the network. These metrics support his leader status.

Table 2: Hijackers Network Neighborhood

* possible false ID
Betweenness Closeness
0.361 Mohamed Atta 0.588 Mohamed Atta 0.587 Mohamed Atta
0.295 Marwan Al-Shehhi 0.252 Essid Sami Ben Khemais 0.466 Marwan Al-Shehhi
0.213 Hani Hanjour 0.232 Zacarias Moussaoui 0.445 Hani Hanjour
0.180 Essid Sami Ben Khemais 0.154 Nawaf Alhazmi 0.442 Nawaf Alhazmi
0.180 Nawaf Alhazmi 0.126 Hani Hanjour 0.436 Ramzi Bin al-Shibh
0.164 Ramzi Bin al-Shibh 0.105 Djamal Beghal 0.436 Zacarias Moussaoui
0.164 Ziad Jarrah 0.088 Marwan Al-Shehhi 0.433 Essid Sami Ben Khemais
0.148 Abdul Aziz Al-Omari* 0.050 Satam Suqami 0.424 Abdul Aziz Al-Omari*
0.131 Djamal Beghal 0.048 Ramzi Bin al-Shibh 0.424 Ziad Jarrah
0.131 Fayez Ahmed 0.043 Abu Qatada 0.409 Imad Eddin Barakat Yarkas
0.131 Salem Alhazmi* 0.034 Tarek Maaroufi 0.409 Satam Suqami
0.131 Satam Suqami 0.033 Mamoun Darkazanli 0.407 Fayez Ahmed
0.131 Zacarias Moussaoui 0.029 Imad Eddin Barakat Yarkas 0.404 Lotfi Raissi
0.115 Hamza Alghamdi 0.026 Fayez Ahmed 0.401 Wail Alshehri
0.115 Said Bahaji 0.023 Abdul Aziz Al-Omari* 0.399 Ahmed Al Haznawi
0.098 Khalid Al-Mihdhar 0.022 Hamza Alghamdi 0.399 Said Bahaji
0.098 Saeed Alghamdi* 0.017 Ziad Jarrah 0.391 Agus Budiman
0.098 Tarek Maaroufi 0.015 Ahmed Al Haznawi 0.391 Zakariya Essabar
0.098 Wail Alshehri 0.013 Salem Alhazmi* 0.389 Mamoun Darkazanli
0.098 Wail Alshehri 0.013 Salem Alhazmi* 0.389 Mamoun Darkazanli
0.098 Waleed Alshehri 0.012 Lotfi Raissi 0.389 Mounir El Motassadeq
0.082 Abu Qatada 0.012 Saeed Alghamdi* 0.389 Mustafa Ahmed al-Hisawi
0.082 Agus Budiman 0.011 Agus Budiman 0.372 Abdelghani Mzoudi
0.082 Ahmed Alghamdi 0.007 Ahmed Alghamdi 0.372 Ahmed Khalil Al-Ani
0.082 Lotfi Raissi 0.007 Ahmed Ressam 0.365 Salem Alhazmi*
0.082 Zakariya Essabar 0.007 Haydar Abu Doha 0.361 Hamza Alghamdi
0.066 Ahmed Al Haznawi 0.006 Kamel Daoudi 0.343 Abu Qatada
0.066 Imad Eddin Barakat Yarkas 0.006 Khalid Al-Mihdhar 0.343 Tarek Maaroufi
0.066 Jerome Courtaillier 0.004 Mohamed Bensakhria 0.339 Ahmed Alghamdi
0.066 Kamel Daoudi 0.003 Nabil al-Marabh 0.335 Waleed Alshehri
0.066 Majed Moqed 0.002 Jerome Courtaillier 0.332 Djamal Beghal
0.066 Mamoun Darkazanli 0.002 Mustafa Ahmed al-Hisawi 0.332 Khalid Al-Mihdhar
0.066 Mohamed Bensakhria 0.002 Said Bahaji 0.332 Saeed Alghamdi*
0.066 Mounir El Motassadeq 0.002 Wail Alshehri 0.328 Majed Moqed
0.066 Mustafa Ahmed al-Hisawi 0.001 Abu Walid 0.324 Ahmed Ressam
0.066 Nabil al-Marabh 0.001 Mehdi Khammoun 0.323 Ahmed Alnami
0.066 Rayed Mohammed Abdullah 0.001 Mohand Alshehri* 0.323 Nabil al-Marabh
0.049 Abdussattar Shaikh 0.001 Raed Hijazi 0.321 Haydar Abu Doha
0.049 Abu Walid 0.001 Rayed Mohammed Abdullah 0.319 Mohamed Bensakhria
0.049 Ahmed Alnami 0.001 Waleed Alshehri 0.316 Essoussi Laaroussi
0.049 Haydar Abu Doha 0.000 Abdelghani Mzoudi 0.316 Jerome Courtaillier
0.049 Mehdi Khammoun 0.000 Abdussattar Shaikh 0.316 Kamel Daoudi
0.049 Osama Awadallah 0.000 Abu Zubeida 0.316 Seifallah ben Hassine
0.049 Raed Hijazi 0.000 Ahmed Alnami 0.314 Rayed Mohammed Abdullah
0.033 Ahmed Ressam 0.000 Ahmed Khalil Al-Ani 0.313 Raed Hijazi
0.033 Bandar Alhazmi 0.000 Bandar Alhazmi 0.311 Abdussattar Shaikh
0.033 David Courtaillier 0.000 David Courtaillier 0.311 Bandar Alhazmi
0.033 Essoussi Laaroussi 0.000 Essoussi Laaroussi 0.311 Faisal Al Salmi
0.033 Faisal Al Salmi 0.000 Faisal Al Salmi 0.311 Mohand Alshehri*
0.033 Lased Ben Heni 0.000 Faisal Al Salmi 0.311 Osama Awadallah
0.033 Mohammed Belfas 0.000 Jean-Marc Grandvisir 0.308 Mehdi Khammoun
0.033 Mohand Alshehri* 0.000 Lased Ben Heni 0.308 Mohamed Abdi
0.033 Seifallah ben Hassine 0.000 Madjid Sahoune 0.307 David Courtaillier
0.016 Abdelghani Mzoudi 0.000 Majed Moqed 0.307 Mohammed Belfas
0.016 Abu Zubeida 0.000 Mamduh Mahmud Salim 0.305 Lased Ben Heni
0.016 Ahmed Khalil Al-Ani 0.000 Mohamed Abdi 0.303 Fahid al Shakri
0.016 Fahid al Shakri 0.000 Mohammed Belfas 0.303 Madjid Sahoune
0.016 Jean-Marc Grandvisir 0.000 Mounir El Motassadeq 0.303 Samir Kishk
0.016 Madjid Sahoune 0.000 Nizar Trabelsi 0.281 Mamduh Mahmud Salim
0.016 Mamduh Mahmud Salim 0.000 Osama Awadallah 0.264 Abu Walid
0.016 Mohamed Abdi 0.000 Samir Kishk 0.250 Abu Zubeida
0.016 Nizar Trabelsi 0.000 Seifallah ben Hassine 0.250 Jean-Marc Grandvisir
0.016 Samir Kishk 0.000 Zakariya Essabar 0.250 Nizar Trabelsi
0.081 Average 0.032 Average 0.052 Average
0.289 Centralization 0.565 Centralization 0.482 Centralization

Yet, we are obviously missing nodes and ties in this network. Centrality measures are very sensitive to minor changes in network connectivity. A discovery of a new conspirator or two, or the uncovering of new ties amongst existing nodes can alter who comes out on top in the centrality measures. We must be wary of incomplete data.



To draw an accurate picture of a covert network, we need to identify task and trust ties between the conspirators. The same four relationships we often map in many business organizations would tell us much about illegal organizations. This data is occasionally difficult to unearth with cooperating clients. With covert criminals, the task is enormous, and may be impossible to complete. Table “3” below lists multiple project networks and possible data sources about covert collaborators.

Table 3: Networks to Map

Relationship/Network Data Sources
1. Trust Prior contacts in family, neighborhood, school, military, club or organization. Public and court records. Data may only be available in suspect’s native country.
2. Task Logs and records of phone calls, electronic mail, chat rooms, instant messages, Web site visits. Travel records. Human intelligence: observation of meetings and attendance at common events.
3. Money & Resources Bank account and money transfer records. Pattern and location of credit card use. Prior court records. Human intelligence: observation of visits to alternate banking resources such as Hawala.
4. Strategy & Goals Web sites. Videos and encrypted disks delivered by courier. Travel records. Human intelligence: observation of meetings and attendance at common events.

Of course, the common network researcher will not have access to many of these sources. The researcher’s best sources may be public court proceedings, which contain much of this data (U.S. Department of Justice, 2001).

The best solution for network disruption may be to discover possible suspects and then, via snowball sampling, map their individual personal networks – see whom else they lead to, and where they overlap. To find these suspects it appears that the best method is for diverse intelligence agencies to aggregate their individual information into a larger emergent map. By sharing information and knowledge, a more complete picture of possible danger can be drawn. In our data search we came across many news accounts where one agency, or country, had data that another would have found very useful.

To win this fight against terrorism it appears that the good guys have to build a better information and knowledge sharing network than the bad guys.


Comments are closed.

%d bloggers like this: