INTRODUCTION
One of the areas of terrorism that has not been adequately examined is that of the use of fire as a terrorist weapon in general and forest pyro-terrorism (PT) in particular. Relying on Clark and Newman’s understanding that terrorism is a form of crime (Clarke and Newman Reference Clarke and Newman2006), criminological theories and practices, which focus on regular crime, greatly contribute to the study of terrorism and assist in developing effective crime prevention tactics and counterterrorism measures (Dugan, LaFree, and Piquero Reference Dugan, LaFree and Piquero2005; Perry Reference Perry2020; Perry, Hasisi, and Perry Reference Perry, Hasisi and Perry2019). Thus, based on the “crime and place” field in criminology, this paper studies the geographic concentration of such forest PT attacks in Israel.
The study of crime and place is among the most rapidly developing fields in criminology, specifically in terrorism research (Brantingham Reference Brantingham2011), partly because many challenges are encountered when seeking to identify the unique profiles of terror offenders (Silke Reference Silke and Silke2003). Clarke and Newman (Reference Clarke and Newman2006) argued that a greater focus on place and opportunities rather than on the offender could generate significant benefits. To date, dozens of high-quality studies have identified that crime is concentrated in selected places (Eck et al. Reference Eck, Lee and Martinez.2017; Lee et al. Reference Lee, Eck and Martinez.2017). Similarly, new studies have demonstrated that terrorism shares many spatial characteristics with ordinary crime (Hasisi et al. Reference Hasisi, Perry, Ilan and Wolfowicz2020; Perry Reference Perry2020; Perry et al. Reference Perry, Hasisi and Perry2019). Based on this premise, new strategies, such as hot spot policing and focused deterrence tactics, have been successfully deployed to prevent and reduce terrorism. Some of these studies explore the geographical distributions of terrorism at the macro-level units of analysis: internationally, regionally, nationally and even locally (e.g. Braithwaite and Li Reference Braithwaite and Li2007; LaFree et al. Reference LaFree, Morris, Dugan and Fahey2006; LaFree, Miller, and Yang Reference LaFree, Miller and Yang2013). Perry’s study (2020) revealed that similar to regular crime, there is a higher frequency of terror attacks concentrated at the micro-level in specific street segment hot spots and that they were stable over time. Hasisi et al. (Reference Hasisi, Perry, Ilan and Wolfowicz2020) examined the spatial characteristics of vehicular terror attacks’ “hot routes” and concluded they were highly concentrated.
Besenyő (Reference Besenyő2019) argued that even though forest fires as a form of terrorism are a worldwide phenomenon, they have not received enough research attention. By examining the spatial characteristics of forest PT attacks, this study follows Besenyő’s (Reference Besenyő2019) recommendation and investigates forest PT in Israel. The study aims to help develop environmental and situational prevention counter-terrorist approaches to combat this form of terror.
The next (second) section will review what we know about the spatial concentration of terror attacks. Every act of crime/terror is determined by the immediate circumstances of the situation (Clarke Reference Clarke1997), which, to a large extent, are shaped by geographic spatial characteristics. Therefore, according to Clarke and Newman (Reference Clarke and Newman2006), the opportunity related to an attack’s circumstances should be analysed. The circumstances related to opportunity will be presented in the third section and divided into four elements, which Clarke and Newman (Reference Clarke and Newman2006) referred to as the “four pillars of opportunity”: target; weapon; tools/training; and facilitating conditions. This section focuses on these “four pillars of opportunity” concerning forest PT attacks. The fourth section will present the applicability of crime concentration to forest PT, while the fifth section will discuss the design of the current study, followed by the presentation of study findings and discussion.
THE SPATIAL CONCENTRATION OF TERROR ATTACKS
According to “rational choice” theory, criminal acts are an outcome of a series of choices in which offenders balance opportunities, risks and potential rewards when deciding whether and where to offend since various places offer different opportunities in terms of suitable targets and the level of guardianship (Cornish and Clarke Reference Cornish and Clarke1987). Similarly, the “routine activities” theory claims that in addition to a motivated offender, a suitable victim/target and lack of appropriate guardianship are necessary for a crime to occur (Felson Reference Felson2017).
The perspective that specific places offer unique opportunities over others is substantiated in crime pattern theory, which suggests that offenders feel more confident to offend in environments that they are more familiar with (Levine Reference Levine2009; Turner Reference Turner, Canter and Youngs2017). Groff et al. (Reference Groff, Taylor, Elesh, McGovern and Johnson2014) contended that offenders become familiar with potential opportunities through their routine activities, enabling them to become familiar with the settings and the different elements of the environment. Individuals operate daily in what is referred to as an activity space, which typically includes “nodes”, “pathways” and “edges”. The nodes are specific locations a person routinely visits, such as home, work, “hang out” locations, stores, schools, etc. Commuting between nodes, the individual uses “pathways”, which often cross the activity space. The boundaries of the individual’s activity space are equivalent to edges in an urban environment, which are unique to each individual’s activity space. These “edges” may be represented by pathways, such as major highways (Hasisi et al. Reference Hasisi, Perry, Ilan and Wolfowicz2020). Brantingham and Brantingham (Reference Brantingham and Brantingham1995) claimed that crime will be more likely to occur when a motivated offender’s activity space overlaps with the activity space of a potential victim/target.
Based on the above understanding, many studies have discovered that a large percentage of the total number of crimes in various cities is concentrated in a small number of micro-geographic places (Braga, Hureau, and Papachristos Reference Braga, Hureau and Papachristos2011; Brantingham and Brantingham Reference Brantingham and Brantingham1999; Pierce, Spaar, and Briggs Reference Pierce, Spaar and Briggs1988; Roncek Reference Roncek, Goldsmith, McGuire, Mollenkopf and Ross2000; Sherman, Gartin, and Buerger Reference Sherman, Gartin and Buerger1989; Weisburd and Amram Reference Weisburd and Amram2014; Weisburd and Telep Reference Weisburd and Telep2012). In a comparison study, Weisburd (Reference Weisburd2015) found that in larger cities, approximately 50% of crime incidents were concentrated between 4.2 and 6% of street segments, whereas 25% were concentrated between 0.8 and 1.6%. In addition, not only are large proportions of crime concentrated in a small number of places, but there is also a spatial–temporal stability whereby the concentrations remain relatively stable over time (Braga, Andresen, and Lawton Reference Braga, Andresen and Lawton2017; Gill, Horgan, and Corner Reference Gill, Horgan and Corner2019). Song et al. (Reference Song, Andresen, Brantingham and Spicer2017) explained that the spatial–temporal stability of crime concentration is mostly the result of the spatial characteristics of criminal offences, such as proximity to high-risk areas, land use, urban density and location accessibility. Moreover, studies recognize that there is a significant overlap between ordinary criminals and terrorist offenders, especially as to their rational choice decisions (Clarke and Newman Reference Clarke and Newman2006; Dugan et al. Reference Dugan, LaFree and Piquero2005; Newman and Hsu Reference Newman, Hsu, Kumar and Mandal2012; Perry and Hasisi Reference Perry and Hasisi2015).
Terrorist actions, similarly to regular criminal activities, are constrained by the same types of geographical restrictions (Clarke and Newman Reference Clarke and Newman2006; Hasisi et al. Reference Hasisi, Perry, Ilan and Wolfowicz2020; Perry Reference Perry2020). The geographical and spatial characteristics of a particular target play an essential role in determining its appeal as a potential location for a terrorist attack (Gill et al. Reference Gill, Horgan and Corner2019; Griffiths, Johnson, and Chetty Reference Griffiths, Johnson and Chetty2017; Newman and Hsu Reference Newman, Hsu, Kumar and Mandal2012; Perry et al. Reference Perry, Hasisi and Perry2019; Rossmo and Harries Reference Rossmo and Harries2011). As mentioned above, when selecting a target, terrorists consider a range of circumstances such as accessibility, distance and the target area, as well as escape routes, the complexity of the journey, opportunities for bringing a weapon to the target, and the terrorist’s familiarity with the routes to and from the target (Clarke and Newman Reference Clarke and Newman2006; Gill et al. Reference Gill, Horgan and Corner2019; Perry et al. Reference Perry, Hasisi and Perry2019; Spaaij Reference Spaaij2010, Reference Spaaij2012).
According to Clarke and Newman (Reference Clarke and Newman2006), the opportunity that relates to the circumstances of an attack can be divided into four elements, referred to as the “four pillars of opportunity”: weapon; target; tools/training; and facilitating conditions. These four elements are essential when carrying out a terrorist attack (Clarke and Newman Reference Clarke and Newman2006; Fahey et al. Reference Fahey, LaFree, Dugan and Piquero2012; Perry et al. Reference Perry, Hasisi and Perry2019). The following section presents forest PT’s four pillars of opportunity, suggesting that engaging in this type of terrorism is a rational choice.
THE “FOUR PILLARS OF OPPORTUNITY” OF FOREST PYRO-TERRORISM ATTACKS
PT in general and forest PT, in particular, could be defined as the use of arson as a method of violence by non-state individuals or organizations to psychologically impact, terrorize and intimidate in order to pressure both the government and the civilian population and to advance political, religious, or any ideological or social motivations (Baird Reference Baird2006; Bendle Reference Bendle2008; Besenyő Reference Besenyő2019; Deshpande Reference Deshpande2009). Forest fires are a fairly new form of terrorism, identified in the United States, Israel, France, Spain and Greece (Besenyő Reference Besenyő2019; Deshpande Reference Deshpande2009). This form of terrorism must alarm us, especially as global terrorist organizations continue to modify their strategies (Bendle Reference Bendle2008).
Thousands of forest PT incidents have been identified in Israel, causing injury to people, loss of property and natural treasures, as well as having a psychological and moral impact on the general population. There are also financial costs resulting from the significant expense of extinguishing the fire, covering the damage caused, and the cost of rehabilitation. A cost analysis conducted in Israel by Levi et al. (Reference Levi, Disney, Shechter and Naman2019) concluded that the costs of extinguishing fires caused by suspected forest PT (compared to other causes of fires) are the highest due to the number and severity of fires in this arson category. Analysing forest PT’s four pillars of opportunity can shed light on and provide insights into this phenomenon, emphasizing the ability to deal with it more effectively and place-focused.
Weapon
The terrorist’s weapon and target selection process are guided by carefully considering resources versus benefits (Clarke and Newman Reference Clarke and Newman2006). Therefore, the choice of arson as a weapon calculates the ease with which one can ignite a forest fire on the one hand and the immense potential damage it could cause on the other. Fire and resulting smoke are powerful weapons that threaten civilians and firefighters and have the potential to destroy vast areas (Besenyő Reference Besenyő2019). Terrorists choose this form of attack as it is very destructive and constitutes a fairly low risk of being apprehended and convicted. Therefore, PT can be perceived as providing an opportunity whereby minimal effort can lead to extreme and disastrous results (Baird Reference Baird2006; Bendle Reference Bendle2008; Besenyő Reference Besenyő2019; Deshpande Reference Deshpande2009; Marsden, Marino, and Ramsay Reference Marsden, Marino and Ramsay2014).
Marsden et al. (Reference Marsden, Marino and Ramsay2014) quoted Inspire, a jihadist magazine, to demonstrate the attackers’ point of view regarding the cost suffered by the “enemy” as a result of PT. The jihadist author describes the fires in New South Wales in the south of Australia in December 2002 as the worst event of wildfires in 30 years. A firestorm released heat energy equal to that of 23 nuclear bombs, causing tremendous damage in a horrifying event. Besenyő (Reference Besenyő2019) quoted the Inspire magazine from 2010, which advocated forest fires as part of jihad. In 2012 (issue no. 9), the magazine elaborated on how igniting forest fires in America, Australia or Europe could greatly support Al-Qaeda.
The Cybersecurity and Infrastructure Security Agency (2021) has prepared an action guide to provide information to all first responders, security professionals and the general public in the United States. The guide’s goal was to mitigate the use of fire, including deliberately setting forest fires as a weapon of terror. The National Counterterrorism Center, Department of Homeland Security, and Federal Bureau of Investigation (2022) posted a warning about wild-land arson in the United States by foreign terrorist organizations, especially in the West, due to its perceived simplicity and the major human, property, infrastructure, environmental and economic losses such cases of arson can cause. This publication mentions an English-language video released by the Islamic State of Iraq and Syria (ISIS) in July 2020, titled “Incite the Believers”, that urged supporters to conduct arson attacks in forests, fields, cities and villages, and to dispose of evidence safely to hinder law enforcement investigations. The video advises to “monitor well for a place where you can set a fire without drawing attention” and “consider that the fire will be so great that efforts made to extinguish it will cost your enemies greatly, and perhaps they will not be able to put it out”. A target map at the end of the video recommends areas for forest fires in the western United States (National Counterterrorism Center, Department of Homeland Security, and Federal Bureau of Investigation 2022).
Target
Theoretically, there are unlimited potential targets that can be selected. However, several important characteristics make certain targets more attractive for terrorists (Clarke and Newman Reference Clarke and Newman2006; Newman and Hsu Reference Newman, Hsu, Kumar and Mandal2012). Targets chosen by terrorists are mainly divided into two categories: “hard” and “soft” targets (Asal et al. Reference Asal, Rathemeyer, Anderson, Stein, Rizzo and Rozea2009; Newman and Hsu Reference Newman, Hsu, Kumar and Mandal2012; Perry et al. Reference Perry, Hasisi and Perry2019). Hard targets have a greater effect yet are more difficult to attack, as they are better defended (e.g. government buildings and military bases). Soft targets, such as accessible areas, roads and public transport, shopping centres, outdoor public events and forests, are more exposed and, thus, facilitate easier attacks.
As mentioned in the previous paragraph, an additional key consideration in the target selection is the accessibility of the target to the attacker, primarily the acquaintance of the attacker with the route and the target area, the distance to the target, potential escape routes, the complexity of the journey and the possibility of bringing a weapon to the target (Clarke and Newman Reference Clarke and Newman2006; Hasisi et al. Reference Hasisi, Perry, Ilan and Wolfowicz2020; Spaaij Reference Spaaij2010, Reference Spaaij2012). Repeated studies have revealed that lone terrorists, who have limited resources (Bakker and De Graaf Reference Bakker and De Graaf2011; McCauley, Moskalenko, and Van Son Reference McCauley, Moskalenko and Van Son2013), are primarily those who tend to choose targets in familiar settings, close to their place of residence, or part of their daily routine (Becker Reference Becker2014; Brantingham and Brantingham Reference Brantingham and Brantingham1993; Gill, Horgan, and Deckert Reference Gill, Horgan and Deckert2014; Hasisi et al. Reference Hasisi, Perry, Ilan and Wolfowicz2020; Newman and Hsu Reference Newman, Hsu, Kumar and Mandal2012; Perry et al. Reference Perry, Hasisi and Perry2019).
Tools/Training
The third pillar of opportunity – “tools/training” – refers to items, equipment or training needed when preparing for and executing the terrorist attack. Such objects are usually associated with one’s daily life (Bakker and De Graaf Reference Bakker and De Graaf2011; Clarke and Newman Reference Clarke and Newman2006; Meyer Reference Meyer2013). Setting fire to a forest does not require special training; necessary tools (combustible material, matches and a vehicle to reach the specific location) are easily accessible.
Facilitating Conditions
The fourth “pillar of opportunity” – “facilitating conditions” – includes, in addition to the societal and institutional circumstances, the environmental circumstances that enhance the opportunity of the terrorists and increase the likelihood that an attack will succeed (Abadie 2004; Piazza Reference Piazza2006). The right environmental conditions, such as weather conditions, are directly related to the speed of fire spread and are one example of facilitating conditions (Kutiel and Kutiel Reference Kutiel and Kutiel1991). These environmental conditions depend on the availability of combustible material, the strength of the wind, and the atmospheric conditions, with an emphasis on low humidity and high temperatures.
Another important facilitating condition is imitating recent successful attacks (“copycat”), which may help offenders overcome obstacles and realize their will to attack. This is even more important for lone actors because of their limited resources. Following and imitating others can be used as a “shortcut” to effective attacks (Gill Reference Gill2015; Perry, Hasisi, and Perry Reference Perry, Hasisi and Perry2018; Perry et al. Reference Perry, Hasisi and Perry2019), especially when in comparison to members of terror organizations (that have infrastructure), lone attackers lack resources and knowledge. For these reasons, lone attackers repeatedly choose soft targets and relatively uncomplicated means of attack (Becker Reference Becker2014; Borum Reference Borum2013; Gill et al. Reference Gill, Horgan and Deckert2014; Spaaij Reference Spaaij2010, Reference Spaaij2012).
Following the “four pillars of opportunity”, it is clear that forest PT is a simple, low-cost means of terror that does not require extensive planning or resources under suitable atmospheric conditions and exploiting the right environmental conditions. Consequently, forest PT can be carried out by a lone attacker. For example, a potential arsonist can drive along a path, stop at certain points, light a piece of paper, throw it out of the car window and continue his journey without being caught. The weather and terrain conditions will play their part in fuelling the fire (Shelef Reference Shelef2016). Both in light of the ease of implementation and the increasing difficulty of achieving significant damage with more conventional means, PT has been identified as one of the “weapons” that a potential terrorist has. Indeed, studies have proven the connection between forest arson and terrorism (Baird Reference Baird2006; Besenyő Reference Besenyő2019; Fighel Reference Fighel2009; Pfeifer Reference Pfeifer2013).
“HOT FORESTS”: THE APPLICABILITY OF CRIME CONCENTRATION TO FOREST PYRO-TERRORISM
The appeal of a forest as a potential terrorist attack target is likely to be determined by its geographical and spatial characteristics, similar to the selection of other terror attack targets (Gill et al. Reference Gill, Horgan and Corner2019; Griffiths et al. Reference Griffiths, Johnson and Chetty2017; Newman and Hsu Reference Newman, Hsu, Kumar and Mandal2012; Perry et al. Reference Perry, Hasisi and Perry2019; Rossmo and Harries Reference Rossmo and Harries2011). One can assume that pyro-terrorists, similar to ordinary criminals and terrorists, will consider a range of geographical and spatial characteristics when selecting a forest to increase their chances of success. These might include accessibility, distance and the complexity of the journey, as well as their familiarity with the routes to and from the target, including escape routes (Clarke and Newman Reference Clarke and Newman2006; Gill et al. Reference Gill, Horgan and Corner2019; Perry et al. Reference Perry, Hasisi and Perry2019; Spaaij Reference Spaaij2010, Reference Spaaij2012). These geographical and spatial characteristics, which play a role in selecting a forest for a PT attack, are expected to result in a concentration of PT attacks in a small number of forests, similar to the concentration of regular crime and terror in a few places.
Previous studies on terrorism, in general, have attempted to understand why terrorists repeatedly choose certain locations to carry out their attacks. While it was reasonable to assume that these particular hot spots provide the attacker with the opportunity to cause maximum harm, Perry (Reference Perry2020) found that the attacks that were committed in the various hot spot street segments were about half as lethal as the other attacks that did not occur in hot spots. These findings suggest that while causing maximum harm is likely to be an important objective, it might not be the attackers’ primary consideration when choosing a specific place to carry out their attack (Perry Reference Perry2020; Perry et al. Reference Perry, Apel, Newman and Clarke2017). This is likely to be because the preference of many criminals and terrorists is to operate in familiar locations (Farrell and Pease Reference Farrell and Pease2001; Perry et al. Reference Perry, Apel, Newman and Clarke2017, Reference Perry, Hasisi and Perry2019; Weisburd and Telep Reference Weisburd and Telep2012).
Terrorists are expected to minimize the distance to their selected attack site to increase their attack’s utility (Bernasco and Block Reference Bernasco and Block2009; Clarke and Newman Reference Clarke and Newman2006; Gill and Corner Reference Gill and Corner2016; Gruenewald, Allison-Gruenewald, and Klein Reference Gruenewald, Allison-Gruenewald and Klein2015). Several studies examined terror distance decay patterns and found that they were similar to those of ordinary crime (Becker Reference Becker2014; Gill et al. Reference Gill, Horgan and Corner2019; Griffiths et al. Reference Griffiths, Johnson and Chetty2017; Rossmo and Harries Reference Rossmo and Harries2011; Smith et al. Reference Smith, Cothren, Roberts and Damphousse2008). As offenders feel more confident attacking in settings familiar to them, they are more likely to commit arson in proximity to either their place of residence (Brantingham and Brantingham Reference Brantingham and Brantingham1995, Reference Brantingham and Brantingham2003) or activity space.
THE CURRENT STUDY
As mentioned, previous studies have focused on the concentration of terrorism from the global level down to the level of street segments (Perry Reference Perry2020). There have also been a few studies that focused on forest PT (Baird Reference Baird2006; Bendle Reference Bendle2008; Besenyő Reference Besenyő2019; Pfeifer Reference Pfeifer2013), yet we are not aware of any research that has dealt with the concentration of forest PT: “hot forests” in the literal sense.
Based on criminological findings on the concentration of crime and terror, this study aims to examine whether similar to the “law of crime concentration”, forest PT is concentrated in a few selected forests. Consequently, this study examines the entire population of suspected PT in Israeli forests between 2008 and 2022 (except for 2016), totalling 2,297 attacks.
Research Questions
Based on the literature about crime concentration in general and terror specifically, this study aims to examine whether the “law of crime concentration”, which applies to terror attacks at the street-segment level (Perry Reference Perry2020), is also applicable to forest PT This study examines PT in KKL-JNF Footnote 1 forests as a case study in order to answer the following research questions:
-
(1) Do a significant number of PT attacks cluster in only a few forests?
-
(2) Have these “hot forests” been a scene/location for a significant proportion of PT attacks over time?
-
(3) Did the PT attacks that took place in these “hot forests” cause less damage (in terms of the size of the burnt area) than PT attacks that occurred in forests that are not “hot forests”, given the preference of many criminals and terrorists to work with what they know and operate in familiar locations?
-
(4) Do these particular “hot forests” have distinct spatial characteristics that make them more favourable for PT attacks?
Study Population and Research Design
This study focuses on the entire population of KKL-JNF’s 274 forests in Israel and explores whether PT attacks occurred in specific “hot forests”. It utilizes two KKL-JNF databases collected about forest fires, the first between 2008 and 2015 and the second between 2017 and 2022. In November 2016 there was a large wave of fires in Israel. The person responsible for investigating this wave of fires (at the National Fire and Rescue Authority) claimed that close to 90% of the large, significant fires were attributed to PT (Besenyő Reference Besenyő2019; Shelef Reference Shelef2016). That year (2016), a new database was under construction, resulting in data not being collected in an orderly manner for this period. Due to the poor data collection, we chose not to include the 2016 data in our analysis. The two databases included in this study contain information about every forest fire (multiple fire foci in one forest at the same time were counted as one event), dates of the fire, name of forest where the fire occurred, KKL-JNF region, size of the burnt area, and cause of the fire. The cause of the fires is divided into 13 different categories: vacationer negligence; trash burning; forestry work; agricultural work; nature (lightning); mechanical tools (e.g. welding); electrical; criminal; military activity; enemy fire; unknown; suspected terrorist arson; and others.
It is generally accepted to consider a crime as being solved only in cases where a suspect is arrested and when there is enough evidence to convict. Due to the nature of forest PT crime, there are very few incidents in which arsonists are apprehended. Moreover, there are very few cases where there is sufficient evidence against the suspect, proving beyond a reasonable doubt that the fire was maliciously set due to nationalist motives. Consequently, even with a strong suspicion of PT arson, it could not be definitively categorized as a PT attack. Therefore, the term “suspected (PT) arson” is used if two conditions are met:
-
(1) A fire with special characteristics compatible with PT (for example, several ignition spots near each other).
-
(2) All other fire causes (vacationer negligence, trash burning, forestry work, agricultural work, lightning, welding, electrical installations, criminal activity, military activity, enemy fire, etc.) were ruled out.
In addition to these categories, there is an “unknown” category in which the investigator did not reach any conclusive conclusions regarding the cause. Of the 10,583 forest fires in the study period, 4,207 were classified as having “unknown” causes (39.8% of all fires). Most likely, some of these instances were, in fact, PT attacks. Due to the uncertainty, fires classified with an unknown cause were not counted as suspected PT attacks.
Of the 10,583 forest fires in the study period, 2,297 (21.7 %) were suspected PT attacks. The 2,297 fires examined in this case study are the population of suspected PT attacks. The smallest geographic location unit in most of KKL-JNF’s data is the specific forest rather than an exact location within that area, such as a landmark. Therefore, this study refers to forests as the basic unit of analysis. The total number of forests under the supervision of KKL-JNF during this study is 274.
FINDINGS
Did a Significant Number of PT Attacks Cluster in Only a Few Forests?
The 2,297 PT attacks (forests set on fire in one or more sites) took place in 125 out of KKL-JNF’s 274 forests. Thus, 100% of all attacks occurred in only 45.6% of the 274 forests. In all, 33 forests endured only one attack, 34 forests between two to five attacks, 15 were set on fire six to 10 times, and 43 forests more than 11. “Moradot Nazeret” Forest had the largest number of PT attacks, totalling 205 events (8.92% of the total PT attacks). “Kfar HaHoresh” had 163 PT attacks (7.1%), “Ben Shemen” Forest had 137 (5.96%), and 134 attacks (5.83%) occurred at “Ezyon” Forest (see Table 1). A total of 27.82% (n = 639) of the PT attacks occurred in just four “hot forests”, representing 1.46% of the 274 KKL-JNF’s forests, while 52.5% (n = 1,206) of the attacks occurred in just 10 “hot forests”, representing 3.65% of the 274 KKL-JNF’s forests.
Table 1. Distribution of Pyro-terrorism (PT) Attacks in Forests Where More Than 10 Events Occurred During the Study Period
Av. No., average number; Sq M, square metres.
The Gini coefficient is a conventional measure of concentration that varies from 0 to 1. Within spatial concentration, a Gini of 0 would indicate that all events have a perfectly equal distribution, whereas a Gini of 1 would indicate that all events were limited to a single place. The Gini coefficient (G’) for the concentration of all PT attacks in the 274 KKL-JNF’s forests was 0.86 (see Figure 1), representing an extremely high concentration level. It indicates that the concentration is more than double that of pure randomness.
Figure 1. Gini coefficient (G’) and Lorenz curve (––) for pyro-terrorism (PT) forest attacks in 2008–2022. Data for 2016 are not included. Cum, cumulative.
One might suspect that a positive relationship exists between the size of the forest and the number of PT attacks, whereby the larger the forest, the greater the chance of a PT attack. Figure 2, which presents a scatterplot of all forests and the number of PT attacks they have experienced, suggests that this is unlikely to be the case. It reveals that the larger forests are not characterized by having a large number of PT attacks. A high number of attacks occur in medium-sized forests. To illustrate, the largest “Massu’a” Forest is in 13th place in terms of the number of PT attacks; the second largest forest, the “Yatir” Forest, is only in 73rd place (only three); and “Duda’im”, the third largest forest, is in 53rd place. At the same time, the “Moradot Nazeret” Forest, which has the highest number of attacks, is in 23rd place in terms of its size, and the “Kfar HaHoresh” Forest, with the second highest number of attacks, is in 33rd place.
Figure 2. Scatterplot of the number of pyro-terrorism (PT) attacks by forest size.
To strengthen the finding regarding the clustering of PT attacks, we performed a sensitivity analysis designed to check if the concentration of PT in certain “hot PT forests” is not accidental but directly related to opportunities to commit PT. Accordingly, we examined whether fires caused by factors other than PT were also concentrated in PT “hot forests”. We examined the other three major causes of fire (vacationer negligence, trash burning and agricultural work). Footnote 2 The findings revealed that the four “hottest PT forests” had a relatively low number of fires caused by these three causes mentioned above (see Table 2).
Table 2. Comparison of the Concentration of Pyro-terrorism to the Concentration of Fires from Another Three Causes
Moreover, when examining forests with the highest concentrations of fires resulting from the three other causes of fires, we found that the Shukada Forest had the highest concentration of fires resulting from agricultural work (7.3%) but was without a single incident of PT; the Menashe Forest had the highest concentration of fires resulting from trash burning (14.2%), while its share in PT incidents was only 1.61%; and the Yerushalayim (Jerusalem) Forest had the highest concentration of fires due to vacationer negligence (12.4%), while the forest’s share of PT attacks was only 5.05%.
These findings reinforce the proposition that PT hot forests, when compared to forests suffering from fires due to vacationer negligence, trash burning and agricultural work, do not share the same characteristics of PT forests and constitute a different set of opportunities related specifically to PT.
Was There a Significant Proportion of Repeated PT Attacks in These “Hot Forests” Over Time?
When looking at the “hot forests” that endured 11 or more PT attacks during the 14-year observation period (Table 3), we see that the attacks repeatedly occurred in the same “hot forests” over the years studied. Upon studying the top eight “hottest forests”, we found that two forests had at least one attack every year over each of the 14 years, one forest had attacks every year in 13 out of the 14 years, one forest had attacks every year in 12 out of the 14 years, two had attacks every year in 11 out of the 14 years, one had at least one fire in 10 out of the 14 years, and one forest experienced an attack in nine out of the 14 years. This finding reveals that PT’s attacks were concentrated in a limited number of forests over the years. This concentration was stable over time and not a sporadic episode (see Table 3).
Table 3. Number of Pyro-terrorism Attacks per Year
Were PT Attacks in “Hot Forests” Less Harmful than PT Attacks in Other Forests?
In order to examine whether arsonists chose the “hot forests” because they provided opportunities to cause maximum damage, we examined whether the PT attacks in “hot forests” caused more (or less) damage than attacks in forests that had a single PT event. We compared the damage, i.e. the size of the forest burned in forests with a single PT attack, to “hot forests”, where most PT events occurred. As mentioned, there were 33 forests with only one PT, meaning they do not meet the definition of a “hot forest”. In nine of these forests, there was no information about the size of the damage – the burned area. In one forest that endured a single PT attack, the area burned was atypical (too large), and there was most likely an error in the information provided (15,000 dunams, where a dunam is equal to 1,000 square metres). Therefore, this particular event was not included in the analysis.
The total area burned in the remaining 23 forests where there was one PT was 1,905.42 dunams. This translates into an average of 82.84 dunams per fire. These forests were compared with the 23 “hottest” forests, where most PT attacks occurred. An examination of the 23 forests with the highest concentration of PT revealed that the area burned in all 1,694 fires occurring in them was 19,210.34 dunams. Therefore, the average area burned by PT was only 11.34 dunams. As previously mentioned, Perry’s (Reference Perry2020) finding concluded that causing maximum damage was not the main consideration when choosing hot spots for terror attacks in Jerusalem. This is also the case when selecting a forest to carry out the PT.
Do These “Hot Forests” Have Distinct Spatial Characteristics Favourable to PT Attacks?
As mentioned, between the years 2008 and 2022, 27.82% of the PT attacks occurred in just the following four “hot forests”: “Moradot Nazareth”, “Kfar HaHoresh”, “Ben Shemen” and “Ezyon”. Geographically, several characteristics can be identified for these forests: “Moradot Nazeret” Forest and “Kfar HaHoresh” Forest are part of the Nazareth district and are characterized by old trees planted over a century ago. The forests are surrounded by two large urban areas – Nazareth and Nof HaGalil. The city of Nazareth is the economic, political and media centre of Israeli Arabs. Nazareth and the Arab urban settlements around – Iksal, Yafia, Kfar Kana, Mashhad, Ein Mahal, Ilot and Raina – numbering about 200,000 Israeli Arab residents (i.e. more than a tenth of all Arabs in Israel; see the website of the Israeli Central Bureau of Statistics), is an urban continuum that is adjacent to these forests. Road number 60, the longitudinal path from north to south, ends in Nazareth and crosses the “Moradot Nazeret” Forest. This path is connected to the “Green Line”. Footnote 3 Moreover, the Palestinian city of Jenin is a central crossing point for the passage of legal and illegal Palestinian workers.
The “Ben Shemen” Forest is one of the first forests planted in the 1920s. This forested area is close to the large population concentrations of the Dan region and the larger Tel Aviv metropolitan area. The forest is located along Road 1, between Tel Aviv and the Jordan Valley, and Road 443, which connects the centre of Israel and northern Jerusalem and the Palestinian settlements around Ramallah. Road 443 is the main daily commuting route for Palestinian labourers and the passage of goods to the centre of Israel from northern Jerusalem and the Palestinian communities located north of Jerusalem.
The “Ezyon” Forest is located adjacent to the Green Line, with its eastern sector located beyond the Green Line, between the Palestinian town of Zurif and the Israeli settlement of Rosh Zurim and the villages of Mata and Neve Michael, south of the city of Beit Shemesh. Road 367 crosses the forest from east to west, functioning as a transit route between Gush Etzion and Beit Shemesh. Palestinian legal and illegal workers also use it.
The high concentration of PT attacks in these four “hot forests” could be classified by two main characteristics related to situational opportunities. The first characteristic is their accessibility and proximity to the forests via main travel artery pathways, providing easy access to potentially radically motivated Arabs. Second, these forests are characterized by old and dry pine trees, ideal for fires. When combined with certain weather conditions, this dry fuel material results in ideal conditions for arson and is perfect for PT attacks. Despite what is claimed in these two hypotheses, it is important to emphasize that there is a need for future in-depth research to examine the unique characteristics of the PT “hot forests” and to assess what makes them more attractive for PT attacks.
DISCUSSION
This research is one of few that has studied forest PT and the first to examine the concentration of PT forest attacks. The KKL-JNF’s databases of the 2,297 suspected PT attacks in Israel’s 274 forests over 14 years presented a unique opportunity to study the phenomenon. As seen from the Lorenz curve (Figure 1) and Table 1, the 2,297 PT attacks occurred in 125 forests – 45.6% of the 274 KKL-JNF forests. Over a quarter (27.82%; n = 639) of all PT attacks occurred in just four “hot forests”, representing 1.46% of the 274 KKL-JNF forests. Lastly, more than half (52.5%; n = 1,206) of the attacks occurred in just 10 “hot forests”, representing 3.65% of the 274 KKL-JNF forests. The observed concentrations at both the 25% and 50% cut-offs demonstrate great consistency with Weisburd’s (Reference Weisburd2015) findings regarding the distributions of crime in street segments in major US cities.
The advantage of the Gini coefficient is that it enables us to compare concentrations between different types of crime or terror attacks. Therefore, the standardization of the Gini coefficient enables us to compare the concentration of forest PT attacks with the concentration of other types of offences/terror in other places. The Gini coefficient of 0.86 for the concentration of all PT attacks in the 274 KKL-JNF forests represents an exceptionally high concentration level, especially for specific types of crime/terror. This finding is quite significant when compared to findings in other studies. Examples of this can be found in a G’ = 0.61 for terror attacks in Jerusalem (Perry Reference Perry2020). The Gini coefficient of 0.86 is also higher than non-aggregated individual robberies in the Hague (G’ = 0.72) (Bernasco and Steenbeek Reference Bernasco and Steenbeek2017). De Moor, Vandeviver, and Vander Beken (Reference De Moor, Vandeviver and Vander Beken2018) found Gini coefficients of 0.43, 0.54, 0.55 and 0.70 for individual crime types. In another study, Vandeviver and Steenbeek (Reference Vandeviver and Steenbeek2019) found Gini coefficients ranging from 0.50 to 0.65 when limiting their analysis to residential burglaries in Antwerp, Belgium, between 2005 and 2016. Studies of aggregated general crime have found Gini coefficients ranging from 0.70 to 0.80 (Bernasco and Steenbeek Reference Bernasco and Steenbeek2017; Eck et al. Reference Eck, Lee and Martinez.2017; Favarin Reference Favarin2018; Hardyns, Snaphaan, and Pauwels Reference Hardyns, Snaphaan and Pauwels2019; Schnell, Braga, and Piza Reference Schnell, Braga and Piza2017; Steenbeek and Weisburd Reference Steenbeek and Weisburd2016).
Examining stability over time reveals that a significant proportion of forest PT attacks occurred repeatedly in the same small number of hot forests. Similar to previous crime studies, it was hypothesized that the concentration of PT attacks in a small number of “hot forests” would be relatively stable over time and not randomly distributed, which was the case.
The high level of the forest PT attack concentration and its stability over time is explained well by the rational choice perspectives of terror, which find there to be no fundamental differences between the decision-making processes and basic motivations of regular criminals and terrorists (Clarke and Newman Reference Clarke and Newman2006; Perry Reference Perry2020; Perry and Hasisi Reference Perry and Hasisi2015; Perry et al. Reference Perry, Hasisi and Perry2019). The preferable spatial settings of particular forests increase the probability that they will be chosen as targets. Like regular criminals, terrorists evaluate the opportunity structures of specific places, such as forests, weighing the circumstances in light of the expected outcomes and benefits.
While the 274 KKL-JNF forests include many potential targets, various essential characteristics increase the opportunity for “successful” attacks at specific forests. We have identified two possible main situational opportunity characteristics of the “hot forests”: proximity and accessibility of the forests via pathways giving easy access and possible escape routes to potential radically motivated attackers, together with the fact that these “hot forests” consist of highly flammable old trees, resulting in high risk of ignition, especially when combined with dry weather conditions. These two factors may well have been part of the attackers’ considerations when choosing these particular “hot forests”, making them more attractive for repeated attacks.
This study found that PT’s average area burnt in “hot forests” was only 11.34 dunams, compared to 82.84 dunams per fire in forests with only one attack. This finding aligns with prior studies (Perry Reference Perry2020; Perry et al. Reference Perry, Hasisi and Perry2019), suggesting that while causing maximum harm might be an important objective, it is not the primary consideration when choosing a specific place to carry out an attack. Lone actors are most likely to prefer a target with which they are more familiar and locations that present better opportunities to attack.
Since 2017, there has been a significant decrease in PT attacks. In the conversations we had with officials at KKL-JNF, they suggested two main reasons for this. The first is the large wave of fires that occurred in November 2016, where, as mentioned, it will take years for the huge areas burned to become mature forests once again. This, of course, reduces the opportunity to carry out PT attacks in these places. The second reason is a significant change in the work procedures of the KKL-JNF and the National Fire and Rescue Authority. Over the last three years, advanced technological measures have been introduced to detect, identify and launch forces, including drones, mobile observation poles and advanced thermal scanners that can detect a fire during the day. A new procedure that has been introduced enables firefighting planes to respond immediately upon detection of a fire, even before the arrival of ground forces, in order to prevent the fire from spreading. These factors obviously greatly affect the opportunity to perform successful PT attacks and increase the chance of exposure.
Policy Implications
As certain spatial characteristics of “hot forests” increase the situational opportunities for PT attacks, modifications to “hot forest” characteristics could contribute to preventing and reducing future PT attacks. This aligns with Clarke and Newman’s (Reference Clarke and Newman2006) argument that counterterrorism through situational prevention is one of the most effective preventative measures. There is robust evidence that place-based interventions can be effective without displacing crime to nearby areas (Braga, Papachristos, and Hureau Reference Braga, Papachristos and Hureau2010). In the few studies that have examined the effectiveness of place-based interventions in counterterrorism (in addition to overall effectiveness, e.g. Lum, Kennedy, and Sherley Reference Lum, Kennedy and Sherley2006), no evidence of significant displacement of attacks was found. Situational opportunities in alternative locations, which did not receive place-based interventions, were already perceived as less favourable (Hsu and Apel Reference Hsu and Apel2015; Perry et al. Reference Perry, Apel, Newman and Clarke2017). In line with the above findings, attackers would not be likely to displace their PT attacks to other forests. It should be emphasized that the situational approach does not overlook offenders or their motivations (Braga et al. Reference Braga, Andresen and Lawton2017), and in the long run, it is important to reduce terrorist motivations. However, minimizing the opportunities for attacks is more feasible by implementing situational prevention methods such as hot spot policing (Perry and Hasisi Reference Perry and Hasisi2015; Perry et al. Reference Perry, Apel, Newman and Clarke2017). Integrating different strategies, namely by minimizing opportunities via problem-oriented policing in hot spots, is very effective in crime prevention. Effective situational crime prevention measures of target hardening have also prevented terror attacks in Jerusalem (Perry Reference Perry2020; Perry et al. Reference Perry, Hasisi and Perry2019).
Based on the current study’s findings, implementing preventative measures (target hardening), such as obstacles and electronic surveillance devices when focusing on the four chronic “hot forests”, can make a significant impact on forest PT attacks. Another preventative tactic might be replanting the burned “hot forests” with less flammable trees.
Study Limitations
As previously mentioned, the term “suspected (PT) arson” (rather than “PT arson”) is used since it is difficult to arrest suspects and compile a full body of evidence that proves guilt beyond a reasonable doubt. At the same time, the term “suspected (PT) arson” was used only in cases where the investigation results and the intelligence and the scene’s characteristics unequivocally indicated that this was indeed a case of a forest PT attack. Moreover, the number of cases classified as unknown was large (39.8% of all fires), reinforcing the assumption that investigators were not quick to classify a fire as a PT attack. They classified PT attacks only when they were convinced it was a PT attack. They preferred to classify as unknown in cases where they were unconvinced.
Another limitation is that the KKL-JNF data do not provide information that determines the exact location of the igniting spot, thus resulting in using the entire forest as the geographic location unit. Such information would have greatly improved the geographic analysis of crime and place.
Future Research
To further strengthen the findings of this research, a trajectory analysis should be employed to examine whether the concentration of forest PT incidents in certain forests is statistically significant and not merely a result of stochastic variation. This method can present a more robust, data-driven assessment of the likelihood of repeated PT in the same forests. Future research should also examine the significance of the difference in the size of the area unit between the forests (as a large polygon) and street segments, as well as the difference in the density of road networks in the context of the law of crime concentration. As mentioned, future research should also focus on the unique characteristics of PT “hot forests” to ascertain what makes them more attractive for PT attacks.
CONCLUSIONS
The concentration and stability over time of the PT attacks in the 274 KKL-JNF forests are consistent with the findings of Weisburd (Reference Weisburd2015) regarding the distributions of crime in street segments in major US cities. In fact, the findings demonstrate a tremendously high concentration level, especially for non-aggregated individual crime/terror types.
The phenomena of “hot forests” concentration and its stability over time are well explained by the rational choice perspectives of terror since terrorists, similar to ordinary criminals, evaluate the opportunity structures of specific places, in our case forests, and weigh the circumstances in light of the expected outcomes and benefits. In line with prior studies (Perry Reference Perry2020; Perry et al. Reference Perry, Hasisi and Perry2019) and the rational choice perspective, the findings suggest that the opportunity structures (cost–benefit) of specific places, and not causing maximum harm, is the primary consideration when choosing a site to carry out an attack. Therefore, counterterrorism through situational prevention is one of the most effective preventative measures for terrorism in general and PT in particular.
Acknowledgements
We would like to thank KKL-JNF and especially Dr Michael Sprintsin, a forest engineer with KKL-JNF, who assisted in providing data and consulting.
Simon Perry is an Associate Professor at the Institute of Criminology, Hebrew University of Jerusalem (Israel). His research interests include terror, policing terror, rational choice, crime and place, and evidence-based policing. Professor Perry is a retired officer in the Israeli National Police (INP), where he served for 30 years in a variety of positions, including INP head of European Operations between 1987 and 1991, head of narcotics, head of special operations, commander of the intelligence and operations wing of the INP national unit, as well as Police Attaché to the United States and Canada between 2003 and 2007 at the rank of Brigadier General.
Shai Amram is a researcher at Israel’s Ministry of National Security and an adjunct lecturer at the Hebrew University’s Institute of Criminology in Jerusalem. He holds a PhD from the Institute of Criminology at the Hebrew University, and his research interests include urban crime, crime and place, and geography of crime.
Open access
