Spatial And Temporal Evolution And Risk Analysis Of Global Terrorist Attack

Terrorism, which is a complex and ancient social phenomenon, has become the world’s widespread threat. Terrorism in China has also spread from Southern Xinjiang to the whole country in recent years. To obtain a comprehensive understanding of terrorism, this study investigates four aspects of terrorism based on the global terrorism database:the power law of frequency-fatalities, relationships between fatality levels and influencing factors, spatial aggregation and spatial autocorrelation, and risk assessment. This study aims to verify the rules and characteristics of the spatial and temporal evolutions of terrorist attacks and establish the risk index of countries.Terrorist attacks often result in a large number of casualties. By using Pareto distribution analysis, the frequency-fatality can be fitted by a power law distribution in log-log coordinates. The power law is a common phenomenon in terrorism and has spatial and temporal invariances. Four factors, which influence power law, are discussed:months, weapons, attack types and attack targets. By considering these factors, scaling exponents can be established. Smaller scaling exponents correspond to higher probabilities of attack incidents with high fatality levels. The results show that: the months of March, April, and November, chemical and vehicular weapons, hijacking, unarmed assault, and hostage taking (barricade incident), tourist and utility sites are associated with high fatality levels. Under certain conditions, frequency fatality can be well fitted by the double power law, which is different from previous conclusions. The double power law presents a scale invariance, and most double power laws are robust under the influence of different factors. Studies have observed that the ratio of α2/α1is approximately equal to two and that a strong correlation exists between the two exponents and critical points of the double power law. The top-heavy distribution model is built based on scaling exponents, which can evaluate how many of the fatalities lie in the certain proportion of incidents. A study on spatial factors shows that North America and Eastern Europe are different from other regions, that is, these regions have smaller scaling exponents in the second stage of power law than other regions. China is similar to the United States in having a higher probability of high fatality level than Iraq, India, and Colombia. The results of the power law are helpful for predicting the probability of extreme fatality incidents by identifying the probability of low fatality incidents.After investigating the distribution of fatalities and influence of the aforementioned factors (i.e., region, country, weapon, attack type, target, and month), correspondence analysis (CA) is used to explore the relationships between fatality levels and factors. To the best of our knowledge, this study is the first to use CA on terrorism. The results provide the key factors associated with the high fatality levels: the developing countries and regions (i.e., Central America&Caribbean and the Sub-Saharan Africa) tend to be related with fatalities levels of10-29and≥30; hijacking, barricade incident and facility/infrastructure attack have strong relationships with high fatality levels; incendiary, explosives/bombs and BCRN (biological, chemical, radiological and nuclear weapon) are associated with fatality levels of10-29and≥30; military, airports&aircrafts and maritime, February and December tend to be related with level of≥30. The results have been found to have sufficient importance for the prevention of terrorist attacks. Therefore, policymakers should focus on these key factors when generating strategies for terrorist attacks.The spatial and temporal of frequency and fatality of terrorist attacks are investigated, and the spatial inequality test model based on the Gini coefficient and Lorenz curve is proposed, which can obtain the hot spots and spatial aggregation of the terrorist attacks. Larger Gini coefficients correspond to higher spatial aggregation. By considering aforementioned influencing factors,"country" has the largest Gini coefficient. The characteristics of terrorist attacks in different countries influenced by the factors can be identified, for example, the results show that nearly100%of the abortion-related attacks occur in five countries:the Unite States, Canada, West Germany, Argentina and Northern Ireland. Thus, these countries should strengthen their defenses against abortion-related attacks. The global Moran’ I of frequency, fatality, firearm weapons, explosive weapons and incendiary are positive values and their z-scores are greater than1.96, which indicates that these factors have spatial autocorrelation and are significant at the5%level. A larger positive value of Anselin’s LISA statistics indicate the high-high and low-low cluster patterns, while a smaller negative value can show the high-low and low-high cluster patterns. Finally, the significance maps and cluster maps of terrorist attacks are given.A comprehensive risk assessment model based on factor analysis is proposed to conduct risk assessment for the top100countries with the most serious terrorist attacks. The assessment includes seven indicators that are derived from two aspects of terrorism:comprehensive risk factor and the vulnerability factors. The risk ranking of countries is established on the basis of the two factors; this ranking is helpful for improving the self-protection consciousness of citizens and for assisting in the development and implementation of counterterrorism measures. The evolution of terrorist attack indicates that they are becoming increasingly serious. First, terrorist attacks present a diffusion trend, that is, more and more countries are suffering terrorist attacks. Second, terrorist attacks present a centralized trend, that is, a large number of incidents and fatalities are concentrated in a few countries. Iraq, Pakistan and Afghanistan have become the present terrorism hot spots, whereas Syria, Yemen and Somalia are emerging as new terrorism hot spots. Finally, by using GIS technology, the spatial and temporal distribution maps of global terrorist attacks can be established.