Dinorah Frutos, Thanarerk Thanakijsombat, Philip V. Fellman, & Pard Teekasap

Uncovering the Linkages in Maritime Piracy Networks: Developmental Dynamics and

Visualization of Network Data

Dinorah Frutos, Thanarerk Thanakijsombat,

Philip V. Fellman, &Pard Teekasap

Southern New Hampshire UniversityInternational Business Modeling

Laboratory

Nature and Severity of the Threat Pirate Attacks are Largely Confined to 4 Major Areas

Piracy is Increasing• In 2009, maritime piracy reached its highest level since the IMB's

Piracy Reporting Center began tracking piracy incidents in 1992, surpassing levels from 2008, which was the previous record year. Of the forty-nine successful hijackings in 2008, forty-two occurred off the coast of Somalia, including the capture of an oil supertanker, the Sirius Star.

• In 2009, that number rose to forty-seven, despite reduced global shipping and a multi-country naval military force aimed at stemming piracy in the region.

• Overall attacks off the coast of Somalia represented about half of all maritime piracy globally according to the IMB. Though Somalia gets the lion's share of attention on global piracy, the IMB says incidents off the coast of Nigeria, which largely involve robberies rather than hijackings, were more likely to include violence against the crew than anywhere else in the world in 2009. In other areas of the world, including Indonesia, piracy dropped.

Historical Attacks on Vessels Underway

Attack Detail 1998-2008

Methodology: Initial Approach - Social Network Analysis

• Data Visualization based on arrays of network nodes and vertices

• Differentiation between small world and random networks

• “Cleaning up” IMO raw data to eliminate data points below the threshold of interest.

• Development of a seven point ordinal scale for severity of attacks.

Total Density = 0.6458

Two-Mode Network: Time and Location of Maritime Piracy attacks in 1998

Two-Mode Network: Time and Location of Maritime Piracy attacks in 2003

Total Density = 0.6759

Two-Mode Network: Time and Location of Maritime Piracy attacks for 2008 (October)

Total Density = 0.7000

Maritime Piracy Severity Index

Small World Networks: Formal Definitions (Clustering Coefficients)

From Some Considerations on Six Degrees of Separation from A Theoretical Point of ViewNorihito Toyota arXiv.org > physics > arXiv:0803.2399

Output of Network Analysis for the Connection Between Location and Intensity of Piracy Attacks

Output of Network Analysis for the Connection Between Location and Weapon Type Employed in Piracy Attacks

Output of Network Analysis for the Connection Between Location and Number of Pirates Engaged in the Attack

Limitations of the Study

• Difficult to determine whether piracy in fact connects significantly with terrorism outside a formal legal definition (Maritime Terrorism – Risk and Liability, Rand Corp. 2006) http://www.rand.org/pubs/monographs/2006/RAND_MG520.pdf

• Increase In Piracy And Terrorism At Sea; Little Evidence Supports Fear That The Two Crimes Are Merging: The Maritime Dimension of International Security Terrorism, Piracy, and Challenges for the United States (Rand Corp. 2008 http://www.rand.org/pubs/monographs/MG697/

• Piracy Is Terrorism, Published: Dec December 5, 2008

• AT FIRST GLANCE, THE CORRELATION BETWEEN PIRACY AND TERRORISM seems a stretch. Yet much of the basis of this skepticism can be traced to romantic and inaccurate notions about piracy. An examination of the actual history of the crime reveals startling, even astonishing, parallels to contemporary international terrorism. Viewed in its proper historical context, piracy emerges as a clear and powerful precedent.

Ideal Comparative Methodology:The Developmental Dynamics of Terrorist Organizations

Aaron Clauset and Kristian Skrede GleditschThe Santa Fe Institute

• This study employs open source data on frequency and severity of terrorist attacks

• Worldwide database of 3,143 fatal attacks• Calendar Date, Severity and Name of Organization

if known (341 Organizations Identified)• The study quantifies the frequency and severity of

a group’s attacks over its lifetime using a normalized development curve.

Terrorist Attacks – Developmental Curves for Individual Organizations: Cumulative Frequency Increases

Land-Based Terrorist Events: Expected Intensity Does Not Increase over time -- Cumulative Damage is only a product of the overall

number of incidents, not of increasing severity

Observed vs. Expected Severity Cumulative Severity of Events

Comparative Findings Between Patterns of Maritime Piracy and Land Based Terrorism

• Because we don’t have clear group identifiers as do Clauset and Gleditsch we segmented our data into the four main clusters of activity.

• While they observed a developmental curve in the activity of land-based terrorist organizations with respect to frequency of attacks, our analysis of the data did not find a similar curve for frequency and both groups failed to exhibit such a curve for intensity of attacks.

Intensity of Attacks (Number of Attacks)

Based on Intensity of Attack Index

Somalia (1998-2008)

Cumulative Number of Attacks on Vessels Underway

∆t (days)

0

100

200

300

400

500

600

700

0 10 20 30 40 50 60

CUMEVENTSOMALIA

DELT

ATIM

ESOMALIA

DELTATIMESOMALIA vs. CUMEVENTSOMALIA

Malacca Straits (1998-2008)

-200

-100

0

100

200

300

400

500

600

0 20 40 60 80 100

CUMEVENTS

DELTATIM

E1

DELTATIME1 vs. CUMEVENTS

Cumulative Number of Attacks on Vessels Underway

∆t (days)

Indonesia (1998-2010)

0

40

80

120

160

200

0 40 80 120 160 200

CUMULATIVEVENTSINDO

DELTATIM

EIN

DO

DELTATIMEINDO vs. CUMULATIVEVENTSINDO

∆t (days)

Cumulative Number of Attacks on Vessels Underway

Nigeria (1998-2008)

Cumulative Number of Attacks on Vessels Underway

∆t (days)

0

100

200

300

400

500

600

0 10 20 30 40 50

CUMULATIVENIGERIA

DELT

ATIM

ENIG

ERIA

DELTATIMENIGERIA vs. CUMULATIVENIGERIA

Piracy Follows Seasonal Patterns

• However, for each region, there is a seasonality pattern based on the frequency of the attacks

• The following charts show the longitudinal data for the seasonal frequency of attacks for Somalia, Malacca Straits, Indonesia, and Nigeria.

Somalia

Malacca Straits

Indonesia

Nigeria

Conclusion

• It is difficult to make strong inferences about the behavior of piracy organizations, even using advanced tools of statistical physics similar to those used to model sunspot activity.

• The observed behavioral regularities still require additional research.