Drone use “outside areas of active hostilities”: an examination of the legal paradigms governing us covert remote strikes

This article examines the use by the US of drone strikes in regions described as ‘outside areas of active hostilities’ a phrase that appears to presume the application of international humanitarian law. In response to this, the article examines these regions to assess whether armed conﬂicts can be said to exist, and thereby whether international humanitarian law does in fact apply. Periods of armed conﬂict are identiﬁed, as are periods which cannot be characterized as such. Consequently the relevant paradigms of international law applicable to the strikes are established, belying the presumption that international law applies generally to drone strikes.


Introduction
Extraterritorial drone 1 strikes were first undertaken by the US in 2001 as part of Operation Enduring Freedom in Afghanistan. 2 Since then, their use has increased dramatically and includes operations in conflicts in Afghanistan, Iraq, Syria and Libya, as well as covert operations in Pakistan, Yemen and Somalia. A large number of drone strikes have been carried out ostensibly in secret, in places described by the US government as 'outside areas of active hostilities', 3 the vast majority of which are part of the extensive and on-going operations in Pakistan, Yemen and Somalia.
The phrase 'outside areas of active hostilities' is curious; it does not reflect recognized international law pertaining to uses of force or the conduct of hostilities, but nonetheless implies a temporal and spatial classification of events in a quasi-legal manner. The idea of forcible actions being 'outside' (or, conversely, 'within') implies a link with the notion of armed conflict, a legal category applied to situations bearing certain specific characteristics, and which activities can be described as occurring either 'outside of' or 'within'. When an armed conflict is deemed to exist, the legal paradigm governing the conduct of hostilities changes. It is almost universally recognized that during armed conflict States' international human rights law (IHRL) obligations are to be interpreted through the lens of international humanitarian law (IHL) 4 while, during peacetime, IHRL obligations are accorded their ordinary meaning. The augmentation of IHRL by IHL occurs either by lex specialis, 5 mutual application 6 or 1 Variously referred to as remotely piloted aircraft systems, unmanned aerial vehicles or unmanned combat aerial systems, the term of 'drone' will be adopted throughout. Specifically, this article discusses armed combat drones that are piloted remotely; autonomous systems are outside of the scope of the work. derogation, 7 creating a paradigm of law that is largely less restrictive 8 of the use of force than that outside of armed conflict, meaning that States that are party to armed conflicts are able lawfully to resort to greater uses of force than would otherwise be available. In this manner, the classification of 'armed conflict' acts as a gatekeeper between low levels of force, essentially for purposes of law enforcement, and higher levels of force having a military purpose.
This dichotomy of paradigms between peacetime and armed conflict appears transcended by US use of the phrase 'outside areas of active hostilities'. While certain restrictions, self-imposed by the US, make methods of combat 'outside areas of active hostilities' potentially more restrained than those undertaken in 'areas of active hostilities', 9 the phrase obscures the parameters of the concept of armed conflict, as it is applied to regions in which no attempt has been made to demonstrate that armed conflicts are occurring. Though official statements on targeting 'outside areas of active hostilities' do not explicitly state that IHL rules are applied to situations that are not armed conflict, IHL is nevertheless recognized as the primary corpus of law in these areas, 10 suggesting either that the existence of armed conflict is presumed, or that IHL is being purposefully applied in situations that do not amount to armed conflict.
Similar sentiment permeates the 2010 speech of Harold Koh, Legal Adviser to the US State Department during the Obama administration, in which he argued that 'a state that is engaged in an armed conflict or in legitimate self-defense is not required to provide targets with legal process before the state may use lethal force'. 11 Elsewhere, the US government juxtaposed lethal targeting within an armed conflict with that against individuals 'otherwise targetable in the exercise of national self-defense'. 12 These statements gesture towards an attempt to broaden the circumstances in which lethal targeting is permitted, allowing military force to be used outside of armed conflict when the requirements for the lawful resort to force in self-defence under jus ad bellum are satisfied. This conflation of the jus ad bellum with IHL is contrary to established law, in which the determination of whether a use of force is lawful under jus ad bellum has no bearing on the permissible manner in which hostilities may be conducted. 7 Commentators remain divided as to the nuances of the relationship between IHL and IHRL. See, for instance, Hill-Cawthorne (2015); Sivakumaran (2012), pp. 89-94; Greenwood (1999), pp. 288-289; Milanović (2011), p. 124; Scobbie (2010), pp. 456-457; Garraway (2010), pp. 509-510; Sassòli and Olson (2008), p. 605. 8 On the shift in perception of IHL from a protective body of law to one that is permissive, see Kretzmer (2009), pp. 23-31. 9 For instance, asserting that, 'outside areas of active hostilities', 'absent extraordinary circumstance, direct action will only be taken if there is near certainty that the action can be taken without injuring or killing non-combatants'; Procedures, above n. 3, p. 1. This is more restrictive than the standard IHL proportionality requirement, which prohibits collateral damage 'which would be excessive in relation to the concrete and direct military advantage anticipated'; see Henckaerts and Doswald-Beck (2005), ch. 4. 10 Procedures, above n. 3, p. 1.
The phrase 'outside areas of armed conflict' thus forms part of a perceived trend in US foreign policy towards legitimizing the use of military force outside of armed conflicts, a palpable manifestation of which is the notion of 'naked self-defence', 13 or 'self-defence targeting'. 14 Elsewhere it has been suggested that, if uncontested by other States, US practice, in carrying out military operations in this manner over the previous decade and a half, has the potential to encourage the development of customary international law in support of a broadening of permissible military force outside of armed conflict. 15 In the absence of general practice accepted as law, 16 the present author rejects the idea that such a norm of customary international law is crystallizing, but, nonetheless, it does appear that actions have been presumed by the US to be governed by IHL without a determination that they are carried out during armed conflict.
In order to look behind the implicit presumption that IHL is applicable to operations 'outside areas of active hostilities' and to resist the expansion of lawful uses of military force it is necessary to ask whether drone strikes in regions deemed 'outside areas of active hostilities' have in fact occurred within or outside of armed conflict, thereby confirming the applicable paradigm for assessing their lawfulness. It is with answering this question that the present article is concerned. Despite the important implications of this analysis it is an issue that has thus far been underexplored and so it is anticipated that the present work will greatly aid subsequent research into the lawfulness of drone use. The assessment of the lawfulness of specific strikes is beyond the scope of this inquiry, but it is an issue that remains a pressing concern and on which there is a growing literature. 17 This article will first set out the international law governing when armed conflicts exist, focusing on the criteria for identifying non-international armed conflicts (NIACs) (Sect. 2). It will then apply these criteria to instances in which drones have been deployed 'outside areas of active hostilities', namely the purported 'global NIAC', as well as the discrete campaigns in Pakistan, Yemen and Somalia (Sect. 3). These regions are focused on as they comprise the principal theatres for lethal drone operations and have been classified by the US as being 'outside areas of active hostilities', in contrast with Afghanistan, Iraq, Syria and, reportedly, Libya, 18 which are recognized as 'areas of active hostilities'. 19 It will be concluded that there have been reasonably clear periods of NIAC within each State, as well as periods that cannot be defined as such due to insufficiently intense fighting (Sect. 4). The 13 Anderson (2011), pp. 8-9. 14 Corn (2012 identification of these periods of NIAC allows tentative accounts to be made as to the number of drone strikes that have occurred within and outside of NIAC, confirming the legal paradigms applicable and demonstrating that the US was mistaken in presumptively applying IHL. As a result of this analysis conclusions will also be made as to the (lack of) normative power of the designation of regions as being 'outside areas of active hostilities'.
2 The Legal Nature of Armed Conflicts

Identifying Armed Conflicts
Armed conflicts may be either international or non-international. 20 Under common Article 2 of the Geneva Conventions, an international armed conflict (IAC) occurs in 'all cases of declared war or of any other armed conflict which may arise between two or more of the High Contracting Parties'. Thus, an IAC is a conflict that pits a State (or States) against another. It is unnecessary to consider IACs presently as US uses of force in Pakistan, Yemen and Somalia have only been against non-State actors (NSAs), rather than States. 21 Therefore, if drone strikes have occurred within armed conflicts these can only have been NIACs.
There are gradations of NIAC, under common Article 3 of the Geneva Conventions and under Additional Protocol II (AP II), the latter having a higher threshold before becoming extant. For present purposes only conflicts arising under common Article 3 need be considered, as the US is not a party to AP II, meaning that, even if it intervenes in a NIAC at the request of a State party to AP II, the Protocol will not apply to US actions. It should be noted that US drone strikes will nevertheless be subject to customary IHL (CIHL), which is governed by the same criteria as common Article 3 NIACs. 22 Common Article 3 provides no definition of a NIAC, stating only that it applies to 'armed conflict[s] not of an international character occurring in the territory of one of the High Contracting Parties'. The International Criminal Tribunal for the Former Yugoslavia (ICTY) provided a generally accepted test in the Tadić case, defining a NIAC as 'protracted armed violence between governmental authorities and organized armed groups or between such groups within a State'. 23 This has been interpreted as requiring '[t]he existence of organized armed groups' and 'fighting of some intensity'. 24  'banditry, unorganized and short-lived insurrections, or terrorist activities' 25 which, as long as they are 'isolated and sporadic', may cause 'incalculable human fatalities and/or colossal damage to property' but will nevertheless remain outside the scope of IHL. 26 The intensity threshold is 'dynamic', 27 the notion of 'protracted armed violence' suggesting a durational requirement, while the term 'large-scale' in the same paragraph 28 points to a requirement of magnitude. 29 This balance between duration and magnitude manifested in the La Tablada case, in which a NIAC was identified due to the intensity of fighting, despite it lasting only 30 hours. 30 Ad hoc tribunals have provided indicative factors demonstrative of intensity, including number of deaths, casualties and buildings destroyed, 31 duration, 32 the involvement of bodies like the UN Security Council 33 and the geographical scope of the conflict. 34 Additional factors include mobilization of volunteers, 35 distribution of weapons, 36 the nature of weapons used, 37 and the manner of governmental response 38 (for instance, resort to the armed forces rather than police 39 ). These factors are indicative and need not be cumulatively satisfied: the absence of one will not definitively negate the existence of a NIAC. 40 Indicative criteria have also been provided to evidence requisite organization and, like the intensity requirement, no single characteristic is 'essential to establish whether the ''organization'' criterion is fulfilled'. 41 These include 'the existence of 25 Prosecutor v. Milošević 37 Milošević, above n. 25, para. 31; Limaj, above n. 22, para. 136; Haradinaj, above n. 31, para. 49; Boškoski, above n. 31, para. 177. 38 Boškoski, above n. 31, para. 178. 39 Abella, above n. 30, para. 155. 40 Sivakumaran (2012), para. 168. 41 Haradinaj, above n. 31, para. 60. headquarters, designated zones of operation, the ability to procure, transport and distribute arms', 42 the erection of checkpoints 43 the issuing of public communiqués and use of spokespeople. 44 The need for a 'command structure' has been asserted 45 and interpreted widely to mean that a group is able to speak 'with one voice' 46 and 'formulate […] military tactics'. 47 In the Boškoski judgment, the ICTY held that 'the degree of organization required to engage in ''protracted violence'' is lower than that required to carry out ''sustained and concerted military operations'''. 48 Similarly, the Inter-American Commission on Human Rights (IACHR) held the organization requirement to be satisfied by a group's ability to carefully plan, coordinate and execute a military operation. 49 Thus the jurisprudence paints a picture in which organization need not match that of national armed forces. In subsequent sections, these indicative criteria will be applied to the hostilities in which US drone strikes have occurred.
In terms of the geographic application of IHL within NIAC, in the Tadić case it was held that, during NIAC, IHL applies to 'the whole territory under the control of a party, whether or not actual combat takes place there', 50 though 'some of the provisions are clearly bound up with the hostilities and the geographical scope of those provisions should be so limited'. 51 Therefore IHL will apply to drone strikes against a NSA carried out anywhere in a State where there is a NIAC involving that NSA, even if far from active battlefields, where there is a nexus between that strike and the NIAC. 52

Internationalized NIACs
Though seemingly paradoxical, the possibility of an 'internationalized' NIAC is important in this analysis as all US drone strikes have been extraterritorial, rather than internal. While this article is not the place to continue the debate over internationalized NIACs, a brief overview is necessary.
Key is the wording of common Article 2, defining IAC as being 'between two or more of the High Contracting Parties'. The fundament of IACs is their international nature-they occur between States, recalling the original term 'international' posited by Jeremy Bentham, meaning 'the mutual transactions between sovereigns'. 53 However, the presence of multiple States will not necessarily transform a NIAC into an IAC, as affirmed in jurisprudence 54 and scholarship, 55 the conflict must be between those States. Considering briefly some permutations of these socalled internationalized conflicts, it becomes clear how, in many instances, NIACs can remain non-international.
When a NIAC spills-over into a third State it continues not to be between States and so remains a NIAC despite its new cross-border geography. 56 Likewise, if a third State intervenes in a NIAC at the invitation of a territorial State government, the fighting remains between States and NSA, thereby continuing to be non-international. 57 Of course, if third States intervene on behalf of a NSA against another State, the NIAC will become an IAC as it opposes one State against others. 58 The most controversial iteration of internationalized NIAC is that of a State engaging in a conflict with a NSA within the territory of another State where there is no pre-existing conflict. Despite calls for international law to recognize a new type of conflict, 59 the lex lata is that, where intervention occurs with the consent of the territorial government, by not opposing two States the conflict remains a NIAC, a finding with much support. 60 Where intervention is done without consent, a separate IAC may arise between the intervening and host States, contemporaneous with the NIAC between intervening State and NSA, 61 though if fighting between State and NSA becomes 'inextricably bound up' with that between States, both conflicts may be part of a single IAC, for purposes of identifying which corpus of IHL is applicable. 62 It is therefore submitted that IACs are only ever those that pit State against State, with all others being NIACs. 54 It is implicit in the dictum of the International Court of Justice in the Nicaragua case, asserting the possibility of an armed conflict that is 'international in character alongside an internal armed conflict':

Drone strikes and Armed Conflicts in Pakistan, Yemen and Somalia
Now that the parameters of NIAC have been sketched they can be applied to the factual situations 'outside areas of active hostilities' in which drone strikes have been used by the US. This will enable the assessment of whether these situations were or were not instances of NIAC and therefore whether the US was mistaken in presumptively applying IHL. This section will consider first the notion that these drone strikes are part of a global NIAC, before moving on to consider individually the situations in Pakistan, Yemen and Somalia, to assess whether there are discrete NIACs occurring in those regions.  70 The notion of a global armed conflict does not correlate with existing legal categories to which IHL applies. It is nevertheless possible that the US is in a single internationalized NIAC against a homogeneous NSA present in Pakistan, Yemen and Somalia, which would bring its drone programmes within the scope of IHL. Indeed, it has been argued that 'armed Islamist extremists remain the enemy that carried out the September 11, 2001 attacks, and they remain the belligerents in the ongoing War on Terror'. 71 For a group to be a single non-State party to a NIAC, each geographically separate entity must fit within the overall organization of a single NSA, thereby satisfying the organization 63 The National Security Strategy of the United States of America (2002), preamble. 64 National Strategy for Combating Terrorism (2003), p. 2. 65 Ibid., p. 7. 66 Ibid., p. 17. 67 The National Security Strategy of the United States of America (2010), p. 19. 68 The National Security Strategy of the United States of America (2015), p. 9. 69 Koh, above n. 11. 70 Obama, 'Remarks by the President at the National Defense University' (23 May 2013), The White House, https://obamawhitehouse.archives.gov/the-press-office/2013/05/23/remarks-president-nationaldefense-university. 71 Walsh (2015), p. 352 (emphasis added). threshold. 72 This necessitates, at least, that the group possesses a command structure 73 and speaks 'with one voice'. 74 Arguably, the organization requirement becomes harder to satisfy when a (purported) group straddles international borders. For instance, the ICTY has identified 'the ability to procure, transport and distribute arms' 75 as indicative of organization, a significantly harder task when done across borders.

The Purported Global NIAC
The oft-identified NSA party to the purported global conflict is 'al-Qaeda, […] the Taliban and associated forces'. 76 That these forces are 'associated' immediately implies the opposite of a unified NSA with one voice, as to be 'associated' necessarily requires separateness. The US defines an 'associated force' as '(1) an organized, armed group that has entered the fight alongside al Qaeda, and (2) a co-belligerent with al Qaeda in hostilities against the United States or its coalition partners'. 77 It has been argued that the term 'associated forces' provides the US with scope to use force against 'regional extremist Islamist groups waging an armed conflict against the United States [that] often do not conduct joint operations with al Qaeda' 78 but this appears only to be in terms of US domestic law. 79 The implication that regional NSAs, unaffiliated with al-Qaeda, will automatically be part of the NIAC between al-Qaeda and the US is manifestly incorrect under the NIAC organization threshold.
The first category-NSAs having 'entered the fight alongside al Qaeda'-is prima facie below the organization threshold as it clearly envisages separate groups fighting together, rather than one organizationally unified group. The need for NSAs to speak with one voice 80 within a single command structure 81 requires more than the ideological identification of one group with another. The word 'alongside', in the US definition, like 'associated', implies organizational separation. The second category, 'co-belligerency', has been reasserted by the US as a basis upon which to include multiple groups within one NIAC. 82 Its invocation is problematic because it is a concept that historically applies to IACs and requires the participation in hostilities to a significant extent. 83 Its use within NIAC appears to be a conflation of the laws of IAC and NIAC. The notion that one NSA can be a co-belligerent purely by virtue of engaging in violent acts against the same State as another NSA is contrary to the NIAC organization requirement as it bypasses the need for unity of command structure. Therefore, it is submitted that two NSAs cannot be viewed as a single entity in this manner.
Ryan Vogel has asserted that al-Qaeda, conceived of as a global entity, satisfies the organization requirement as it: 'maintains ''headquarters'' in Pakistan and Yemen from which it coordinates attacks'; 'operates within designated Zones in Afghanistan, Pakistan, Yemen and elsewhere'; and 'has demonstrated a persistent ability to procure, transport, and distribute arms to countries across the globe'. 84 Vogel's conclusions are based on a 2012 Report by the US Department of State 85 and accords with the position of that administration. It is submitted that this view fails to fully reflect the reality of the nature and degree of co-ordination between disparate al-Qaeda franchises: an examination of the facts demonstrates that there cannot be a single NIAC between the US and one homogeneous NSA, encompassing Pakistan, Yemen and Somalia. The US has undertaken drone strikes in Pakistan against members of al-Qaeda, Tehrik-i-Taliban Pakistan (TTP) and the Haqqani Network. 86 In Yemen, drone strikes have targeted al-Qaeda in the Arabian Peninsula (AQAP), 87 while in Somalia they have targeted al-Shabaab. 88 Though these groups share a broad ideology, that alone is insufficient to demonstrate unified organization 89 -they are distinct groups operating in different States. In Pakistan, al-Qaeda and TTP appear sufficiently integrated as they purportedly train and plan attacks together, 90 which accords with the jurisprudential understanding of organization as including the formulation of military tactics 91 sufficient for 'protracted violence'. 92 However, this group does not include those that operate in other States. Al-Qaeda internationally comprises autonomous cells that lack central organization. 93 AQAP in Yemen has been described as 'the most active operational franchise' of al-Qaeda 94 but operating as a mere 'terrorism franchise' is below the 84  requisite level of organization to view multiple groups as homogeneous. 95 Likewise, despite having strong links to al-Qaeda, 96 al-Shabaab in Somalia is an autochthonous group with its own command identity.
In 2014 Stephen Preston, then General Counsel of the Department of Defense, specifically identified AQAP as 'part of, or at least an associated force' of al-Qaeda, and al-Shabaab as 'openly affiliated' with the group. 97 But this in itself does not satisfy the organization requirement as there is no unity of command. Additionally, Preston's assertion demonstrates that the determination as to whether a group is an 'associated force' is one made by the US 98 contrary to the manner of determining armed conflict under international law which requires objective assessment, 99 mirroring the trend since 1949 away from a legal threshold of war in favour of a factual threshold of armed conflict. 100 Thus these distinct groups are not one single NSA for the purposes of the NIAC organization requirement and so the US cannot be in a single multinational NIAC. Therefore, IHL may not be presumed to apply on this basis. Nonetheless, the US may be fighting a series of distinct NIACs in different States. The ICTY has confirmed that the existence of NIACs can only be determined on a case-by-case basis 101 and, as such, the next sections will consider in turn the situations in which drones have been deployed in Pakistan, Yemen and Somalia.

Pakistan
In Pakistan, US drone strikes have targeted members of al-Qaeda, TTP and the Haqqani Network. 102 There are three possible avenues to include these strikes within a NIAC: they may be spill-over from the conflict in Afghanistan; they may be part of a NIAC between the Pakistani government and a non-State party; or they may be part of a NIAC between the US and a non-State party.

Spill-over from Afghanistan
It is inappropriate presently to conduct an analysis of the nature of US hostilities in Afghanistan, though some discussion is necessary as a spill-over conflict self-evidently requires an armed conflict elsewhere. The International  97 Preston, above n. 82, p. 3. 98 Ibid. 99 Abella, above n. 30, para. 153. 100 Greenwood (2006), pp. 91-92. 101 Limaj, above n. 22, para. 90. 102 Ackerman, above n. 86. government in June 2002. 103 This understanding is accurate as the use of force by the US-led coalition against the Taliban government of Afghanistan satisfies the inter-State definition of IAC under common Article 2. Upon the establishment of the transitional government, the conflict continued as three distinct NIACs, between the Taliban and al-Qaeda, and the US, the International Security Assistance Force (ISAF) and the government of Afghanistan respectively. 104 US officials have reiterated that there continues to be an armed conflict in Afghanistan. 105 The fighting in Afghanistan appears to satisfy the intensity and organization requirements of a NIAC. While duration is a less conclusive determinant of intensity than magnitude, it is a factor, 106 and the non-international fighting in Afghanistan has lasted from June 2002 until the present, strongly suggesting satisfaction of the criterion. In terms of casualty and fatality numbers the non-international part of the conflict resulted in the deaths of 23,496 civilians by the end of 2014. 107 In addition, it has been estimated that at least 2853 Afghan National Army soldiers were killed between 2003 and 2013, 108 2771 of which were between 2010 and 2013. 109 Considering available data covering the period up to 2014, deaths among Afghan civilians and armed forces is increasing. 110 Since the beginning of 2003 there have been 2320 US fatalities 111 decreasing yearly after a peak in 2010. 112 Though the US' involvement in Afghanistan is purportedly drawing to a close, it has increased its troop presence 113 and the Obama Administration has stated it will maintain a force of 5500 soldiers until at least 2017. 114 Statistics on the number of Taliban and al-Qaeda fatalities are sparse, though one estimate suggests that 4300 Taliban fighters may have been killed in 2014 alone. 115 In addition, all sides have mobilized armed forces, using military rather than law enforcement weapons, satisfying further criteria for the existence of a NIAC. Taken together, these data provide a compelling case in favour of satisfaction of the NIAC intensity requirement.
In terms of the organization of the non-State party to the conflict, the Taliban in Afghanistan appears to satisfy a number of indicative criteria. The group has generally spoken with one voice under the leadership of Mullah Omar since 1996, until the revelation of his death in 2015. 116 In November 2015 the group split into two rival factions, one lead by Mullah Mansour (successor to Mullah Omar) and a breakaway group lead by Mullah Rasool 117 though it appeared that Mansour's group continued to represent the principle party engaged in fighting. This is supported by that group's continued military effectiveness: in October 2015, Mansour's Taliban took control of the city of Kunduz 118 for 15 days 119 and mounted a similar attempt in the city of Sangin. 120 These actions demonstrate an ability to carefully plan, coordinate and execute military operations, satisfying criteria set out in the La Tablada decision. 121 In addition, Mansour's Taliban had engaged in negotiations with the central government in Afghanistan, 122 another indicator of the requisite level of organization. 123 These factors point to the continued organization of the Taliban that is sufficient to allow continued military activities, therefore satisfying the second element of the Tadić NIAC test. It remains to be seen whether this level of organization will continue after Mansour's death by drone strike in May 2016, though the swift and unanimous appointment of Mullah Haibatullah Akhundzada as his successor suggests it will. 124 Thus there is likely a continuing NIAC in Afghanistan, which may have spilled-over into Pakistan, though this does not mean that all drone strikes in Pakistan will come within that NIAC. Due to the organizational NIAC 115 Crawford, above n. 107 requirement, only members of the non-State party to, or individuals directly participating in, the NIAC in Afghanistan may be targeted in Pakistan as part of that same conflict. Many of those targeted by US drone strikes in Pakistan have been members of NSAs not party to the NIAC in Afghanistan. A lack of clear data on the identities of those killed limits the ability to say with certainty the number of drone strikes linked to the Afghanistan conflict. However, from a list of 314 militants killed by drones whose identities have been confirmed by The Bureau of Investigative Journalism, only 12 were specifically acknowledged as involved in planning and carrying out activities in Afghanistan, while 25 were members of groups generally active in Afghanistan. 125 Therefore this brings only 18 out of 233 strikes 126 within a NIAC and, though it is likely that the number is actually higher, this demonstrates the limited scope of the NIAC in Afghanistan to bring drone strikes in Pakistan within the remit of IHL. Therefore it is necessary to consider other avenues through which drone strikes may be viewed as part of a NIAC.

TTP, Al-Qaeda and the Pakistani Government
One possibility is that a NIAC exists between the Pakistani government and a NSA, into which the US has been invited. Fighting has been ongoing in the Federally Administered Tribal Areas (FATA) region, and so it is necessary to consider whether this reaches the requisite thresholds of intensity and organization to be deemed a NIAC.
Before the formation of TTP in 2007, Pakistani armed forces engaged in sporadic 'half-hearted' skirmishes against tribal militant groups in FATA. 127 In August 2008 the government launched its first major offensive against TTP, carrying out weekly attacks and deploying over 15,000 troops by 2009. 128 In 2013 the Pakistani government sought peace talks with the TTP 129 but in June 2014 launched operation Zarb-e-Azb in response to assaults by TTP including the beheading of 23 soldiers, 130 a rocket and machine gun attack on Karachi airport 131 and the killing of an army general. 132 Operation Zarb-e-Azb has involved airstrikes and the positioning of 30,000 troops in FATA 133 and is ongoing 134 as are actions by TTP, which, in 2016 carried out 276 attacks, killing 640 people. 135 The number of clashes and casualties in Pakistan support a conclusion that the fighting is sufficiently intense. 136 In the Limaj case, the Trial Chamber of the ICTY identified a NIAC based on 'sporadic acts of violence' over the course of a year, 137 ranging from an attack that resulted in 83 deaths to a 20 minute firefight with no casualties, 138 'great devastation to a limited number of buildings' 139 and the use of 'tanks and armoured vehicles, heavy artillery weapons, air defence systems, [armoured personnel carriers], machine guns, and explosives, among other weapons'. 140 The Chamber asserted that these facts demonstrated intensity sufficient for a NIAC. 141 TTP has used military-grade weapons and the government has responded by deploying the army rather than the police, both evidence of NIAClevel intensity. 142 Furthermore, the conflict has been of a significant duration, lasting over eight years. These factors support the contention that the conflict is 'protracted'.
In terms of organization, TTP and al-Qaeda train and plan attacks together. 143 This suggests that both may come within the same command structure (sufficient to plan and execute 'protracted violence' 144 ) and, therefore, in terms of the classification of a NIAC, may be conceived of as a single NSA. Alternately, the Pakistani government may be engaged in two NIACs, one against TTP and another against al-Qaeda. The TTP/al-Qaeda amalgam has training bases and headquarters in South and North Waziristan, where it controls extensive territory. In addition, it has demonstrated an ability to obtain and distribute arms, possesses a command structure capable of coordinating the protracted violence detailed above, and has engaged in negotiations with the government, all indicate requisite organization. 145 As such it appears that this combined group may be a single NSA satisfying the NIAC organization requirement, though the lack of substantial evidence prevents a definitive conclusion. Nonetheless, since the death of Hakimullah Mehsud in 2013, 133 Zahid, 'The Successes and Failures of Pakistan's Operation Zarb-e-Azb', The Jamestown Foundation Terrorism Monitor (10 July 2015), http://www.jamestown.org/programs/tm/single/?tx_ttnews%5Btt_ news%5D=44144&cHash=d4281630e5ad104ab6fbc0bd5f3bbf9f#.VsIEyZOLSRs. 134 'US Defence Intelligence Chief Lauds Zarb-e-Azb, Karachi Operation', Geo TV News (Karachi 11 February 2016), http://www.geo.tv/latest/101030-US-defence-intelligence-chief-lauds-Zarb-e-Azb-Karachi-Operation. 135 Rana and Sial (2017), p. 14. 136 Haradinaj, above n. 31, para. 49. 137 Limaj, above n. 22, para. 135. 138 Ibid., para. 138. 139 Ibid., para. 136. 140 Ibid., para. 166. 141 Ibid., para. 169. 142 Haradinaj, above n. 31, para. 49; Boškoski, above n. 31, paras. 177-178. 143 Brennan, above n. 90. 144 Boškoski, above n. 31, para. 197. 145 See above, nn. 41-49. the group has reportedly splintered into multiple groups that lack central organization 146 meaning that it may no longer satisfy the organization requirement, despite continuing to carry out intense attacks in 2016. 147 The Haqqani Network historically has been closely involved with both but focuses its activities instead in Afghanistan. 148 Indeed, the group has been linked to the Pakistani government and was only banned in 2015 149 and so cannot be seen as part of a single group with TTP and al-Qaeda in a NIAC in Pakistan.
Thus it seems that there has been a NIAC occurring within Pakistan from August 2008 until sometime in 2014, when the TTP and al-Qaeda may have become too fragmented to satisfy the organization requirement. US drone strikes were invited by the Pakistani government from at least 2007 until October 2013 when consent was rescinded. 150 This consent was given validly, by requisite officials of the legitimate government and so satisfies the requirements of international law governing intervention by invitation. 151 Therefore of the 424 drone strikes carried out in Pakistan, 361 occurred during the NIAC prior to consent being retracted. 152

TTP, Al-Qaeda and the US
Lastly, US drone strikes may fall more comprehensively within a NIAC if the US is engaged in its own NIAC against TTP and al-Qaeda in Pakistan.
Force used by the US in Pakistan (excluding that which is part of the spill-over conflict from Afghanistan) has been predominantly carried out with drone strikes. Between 2004 and 2007 there were 11 strikes, but this increased after 2008 to 128 in 2010, causing 1108 deaths. This has receded and in 2015 there were only 13. 153 Blank and Farley have argued that these drone strikes inherently satisfy the intensity requirement. 154 However, for the existence of a NIAC both sides must have used force: the Tadić NIAC definition specifically refers to protracted violence 'between' parties, 155 implicating a need for violence from each side. The fact that the US has resorted to military force, though indicative of the intense violence of a NIAC, is not determinative. 156 It is necessary, therefore, to consider the use of force by the non-State party to the conflict.
In the case of a possible NIAC between the US and a TTP/al-Qaeda amalgam, the intensity and organization requirements are deeply connected-sporadic acts of violence carried out globally may be sufficiently intense when aggregated, if they were undertaken by a single group satisfying the organization requirement. Conversely, if the organization threshold is not met, acts of violence cannot be considered cumulatively and are, therefore, less likely to reach the intensity threshold. 157 The ICTY's reference to 'the degree of organisation required to engage in ''protracted violence''' 158 further demonstrates this interrelatedness of the thresholds. It has been argued that attacks by al-Qaeda globally can be understood cumulatively as beyond 'isolated or sporadic' thereby satisfying the intensity requirement. 159 However, this view is unpersuasive as actions ostensibly carried out by al-Qaeda in Indonesia, Spain, the US, the UK and elsewhere have in fact been perpetrated by franchises. Disparate franchises cannot satisfy the organization requirement and so their actions cannot be considered collectively when establishing intensity. 160 Therefore, in assessing whether the US is in a NIAC in Pakistan, the violence carried out by the specific non-State party must be considered in isolation.
The TTP/al-Qaeda amalgam has carried out very few attacks against the US: an analysis of available news reports reveals that between 2002 and 2016 there were three carried out by the group, resulting in nine deaths, three of which were American. 161 The intensity requirement distinguishes NIACs from 'banditry, unorganized and short-lived insurrections, or terrorist activities' 162 and the actions of the TTP/al-Qaeda against the US are an archetypal example of terrorism. Although terrorism can produce an armed conflict 163 the number and intensity of those carried out against the US are insufficient to enable such a conclusion. 155 Tadić, above n. 23, para. 70 (emphasis added); reiterated in Art. 8(2)(f) of the Rome Statute. 156 Schaller (2015) 162 Milošević, above n. 25, para. 26. 163 Boškoski, above n. 31, para. 184.
A further possible argument is that a NIAC may exist because violence was employed against a foreign State. In the La Tablada case the IACHR specified an 'internal' aspect to violence falling short of NIAC, 164 raising the possibility that the internationalized nature of TTP/al-Qaeda attacks might produce a NIAC. However, the fact that violence is directed against a State other than the host does not have any special quality that creates a NIAC. No such suggestion has been made elsewhere in jurisprudence or commentary. The ICTY's cleaving of NIACs from 'banditry, unorganized and short-lived insurrections, or terrorist activities' in the Milošević case, 165 though implying an internal element with the word 'insurrections', clearly, by the use of 'or', sees terrorism as a separate type of violence, a type the tribunal did not class as necessarily internal. That, coupled with the fact that NIACs themselves need not be internal in terms of their protagonists or geographical scope, supports a conclusion that an internationalizing element does not transform the legal classification of violence. It is therefore submitted that the violence between the US and TTP/al-Qaeda falls short of the intensity characterizing a NIAC.
Based on the above considerations of possible NIACs in Pakistan, those US drone strikes carried out as part of the conflict between the Pakistani government and the TTP/al-Qaeda amalgam (during the period that the Pakistani government consented to US intervention) and the conflict in Afghanistan between the US and the Taliban, al-Qaeda and the Haqqani Network can be considered to have occurred within NIAC, and are therefore to be subject to IHL. At the time of writing, this equates to 381 out of 423 drone strikes in Pakistan. The fact that a majority of strikes have been part of NIAC belies the suggestion that they were carried out 'outside areas of active hostilities' and as such their lawfulness rests with IHL. Conversely, a sizable minority where carried out outside NIAC and must therefore satisfy the restrictive rules of IHRL to have been lawful.

Yemen
The situation in Yemen is similarly complex. Though there is no spill-over conflict, US drone strikes could come under the remit of IHL either as part of a pre-existing NIAC between the Yemeni government and AQAP or as part of a NIAC between the US and AQAP directly. Despite the lack of effective control by the government of Yemen over much of the country, its continued de jure control has allowed it validly to consent, under international law, to US drone operations. 166

Houthi Rebels and the Yemeni Government
When considering NIAC involving the government of Yemen it is necessary to distinguish one possible conflict, with AQAP, from another, involving Houthi rebels. The latter appears to have begun in 2004 with a series of intense clashes between Houthi groups and government forces 167 involving around 2000 troops and causing up to 600 deaths. 168 Sporadic fighting continued until 2014 at which point it escalated and the Houthis captured much of Sana'a, the capital. 169 A coalition of States have intervened at the request of the Yemeni government, 170 involving airstrikes which continue at the time of writing. 171 Constraints of space prevent a detailed consideration of this conflict, but the use of air strikes, deployments of troops, assertion by the Houthi rebels that they will abide by IHL 172 and the agreement of a (failed) ceasefire, 173 suggest the violence is a NIAC. Regardless, the US has targeted its drone strikes in Yemen only against AQAP, Ansar al-Sharia or al-Qaeda, not Houthi groups. 174 The Houthis are opposed to al-Qaeda, 175 described as their 'strongest opponents' 176 and the two groups cannot in any way be connected in the sense of being a single party to a NIAC. Thus it is necessary to determine whether a separate NIAC exists in Yemen, between AQAP and the government.

AQAP and the Yemeni Government
Yemen has been in a state of instability since before unification in 1990. This grew more acute after the 2011 Arab Spring, with multiple insurgent groups acting against the government. The existing literature has broadly focused on the more obvious conflict between the government and the Houthi rebellion, rather than that between the government and AQAP, and detailed analysis of that conflict appears , http://www.independent.co.uk/ news/world/middle-east/yemen-crisis-saudi-arabia-ends-formal-ceasefire-with-houthi-rebels-after-40fighters-are-killed-over-a6795501.html. 174 Based on data from New America Foundation, available at http://securitydata.newamerica.net/drones/ yemen-analysis.html#page1. This data refers to three strikes in which the target was 'unclear' but it is submitted that this is most likely to be a lack of clarity as to which of the three al-Qaeda affiliated groups were targeted, rather than a suggestion that the group might have been the Houthi rebels. 175  almost entirely absent. This is likely in part due to its overshadowing by the Houthi NIAC, but also because the intensity of the AQAP conflict has only emerged in recent years.
What academic discussion there has been is divided as to the nature of hostilities. Ramsden stated in 2011 that 'the combat in Yemen is not even close to being sufficiently protracted or intense so as to amount to a localized NIAC'. 177 Writing at the very early stages of the fighting, this conclusion is unsurprising as there was little evidence to enable a rigorous analysis. Similarly, in 2012, Breau and Aronsson briefly discussed a possible separate NIAC involving AQAP, concluding that the intensity and organization requirements were not satisfied. 178 Conversely, in 2012, the ICRC implicitly recognized a NIAC, calling for all sides to respect IHL. 179 More recent scholarship contains greater agreement that there is in fact a NIAC involving AQAP. 180 Nonetheless, the in-depth doctrinal analysis required to make a conclusive determination as to the classification of the conflict remains outstanding.
Al-Qaeda in Yemen (AQY), the precursor to AQAP, had a semblance of government support post-unification, and Islamic militants returning from jihad against the Soviet Union in Afghanistan were used by President Saleh as a counterbalance to Marxist politicians of the newly absorbed People's Democratic Republic of Yemen. 181 AQY undertook a variety of terrorist attacks against foreign targets, including the 2000 USS Cole bombing, 182 but did not conduct systemic attacks against the Yemeni government. After 9/11 the Saleh government changed its policy towards AQY and actively pursued it 183 though through domestic law enforcement. 184 Nonetheless, until 2009 AQY and then AQAP still primarily targeted non-Yemeni objects, 185 meaning there was no NIAC between them and the government. The government launched an offensive against AQAP in 2009 in response to an attempted attack on Saudi Arabia and the US 186 ; this involved cooperation with the US, though without the use of armed drones. In November 2009 AQAP action against the government began, with the killing of seven Yemen security personnel. 187 Despite this, the number of attacks by AQAP against the government was minimal and sporadic and therefore did not satisfy the NIAC intensity requirement. 177 Ramsden (2011) 182 Ibid., p. 50. 183 Ibid., p. 52. 184 Ibid., p. 55. 185 'Profile: Al-Qaeda in the Arabian Peninsula', BBC (London 16 June 2015), http://www.bbc.co.uk/ news/world-middle-east-11483095. 186 Ibid.
The situation changed significantly in 2011, during the uprising against President Saleh, becoming much more intense. In May and June 2011, AQAP captured multiple towns in the south of the country, controlling the large provinces of Shabwa and Abyan 188 until June 2012, when the government retook them. 189 Nasir al Wuhayshi, then-head of AQAP, stated that 'control of these areas during one year cost us 500 martyrs, 700 wounded, […] and nearly $20 million'. 190 In the Limaj case 169 deaths from incidents between February and May 1998 was sufficiently intense for a NIAC, 191 the same death rate (41-43 per month) as the violence between AQAP and the government in southern Yemen. Reportedly 300-700 AQAP fighters were involved in capturing Zinjibar, the capital of Abyan. 192 This is higher than the levels in the Limaj case, which were generally in the low hundreds, 193 though lower than those of the Haradinaj case, in which some operations involved up to 2000 Serbian soldiers. 194 The government responded to the AQAP offensive with shelling and airstrikes, 195 in what was described as an 'intensive campaign'. 196 In the Limaj case, use of heavy weapons and mortars indicated a NIAC. 197 In addition, the Yemini government utilized the army, air force and navy to recapture Abyan. 198 This further indicates the existence of a NIAC 199 as does the fact that fighting occurred in multiple regions. 200 Combat during 2011 and 2012 led to 'tens of thousands of civilians fleeing their homes 201 ; in the Limaj case, the displacement of 15,000 people supported the classification of that violence as a NIAC. 202 Finally, the fighting has drawn the attention of the UN Security Council, which has been said to indicate the existence of a NIAC. 203 In Resolution 2014, the Security Council called for AQAP to be addressed 'in accordance with […] international law including applicable human rights, refugee and humanitarian law' 204 suggesting implicit recognition of a NIAC. In the Boškoski case, the ICTY held that such language in Resolution 1345, reminding parties of their obligations under IHL, suggested the Security Council viewed that situation as a NIAC. 205 It is submitted therefore that the conflict between the Yemeni government and AQAP reached the level of intensity necessary for classification as a NIAC during the period of 2011-2012. However, US drone strikes have continued since that time so it is necessary to consider whether the combat has remained sufficiently intense.
2013 was far less intense but still involved large-scale violence. AQAP carried out attacks against Yemeni Security forces in August (killing five 206 213 It was also reported to control many parts of Hadramout province, including its capital, Mukalla. 214 In 2016 AQAP captured the city of Aden, in which President Hadi is based, before it was retaken by the government after two days. 215 Additionally, in 2016 the Saudi-led coalition supporting the Yemeni government against the Houthi rebels began airstrikes against AQAP. 216 It is arguable that fighting between the Yemeni government and AQAP may have ceased to be sufficiently intense during the period of 2013 to March 2015. In order to assess the impact of this upon the legal paradigm governing uses of force, we must first consider the point at which a NIAC ceases. There has been very little written on this topic: common Article 3 and Additional Protocol II say nothing, and key monographs discuss it briefly, 217 or not at all. 218 The ICTY has suggested a NIAC lasts 'beyond the cessation of hostilities until a peaceful settlement is achieved', 219 meaning that, regardless that fighting may diminish in intensity below the threshold for the start of a NIAC, the conflict continues, until a 'peaceful settlement' manifests. This has been supported in the Haradinaj case, in which the Tribunal stated that Since according to the Tadić test an internal armed conflict continues until a peaceful settlement is achieved, and since there is no evidence of such a settlement during the indictment period, there is no need for the Trial Chamber to explore the oscillating intensity of the armed conflict in the remainder of the indictment period. 220 This approach bases the end of a NIAC on the words of the parties, rather than the factual situation. Thus, it is contrary to the tenor of the Geneva Conventions which emphasize facts rather than statements in establishing the start of an armed conflict, 221 reflecting sentiment expressed at the 1949 diplomatic conference that 'the Parties to a conflict could not be sole judges as to whether a state of war existed'. 222 It is perhaps because of this distinction that text writers appear en masse 219 Tadić, above n. 23, para. 70 (emphasis added). 220 Haradinaj, above n. 31, para. 100. 221 Though common Art. 2 applies in 'all cases of declared war' (emphasis added), it also covers 'any other armed conflict which may arise', thus removing the important role declarations previously played. In terms of NIACs, common Art. 3 does not require a declaration and the Tadić definition considers only the factual interaction of parties to a conflict. 222 Final Record of the Diplomatic Conference of Geneva (1949) IIB, 10 (UK). It was also asserted that 'even if war was not recognized, the rules concerning the conduct of war should be applied', ibid., 11 (Norway).
to have adopted a different stance. Sivakumaran has stated that 'the applicability of the law of [NIAC] turns on whether or not a [NIAC] continues to exist at the relevant time'. 223 This means 'the lack of a peace agreement cannot be considered determinative' and that a NIAC may persist after a ceasefire has been agreed 'as violence of requisite intensity may continue to exist'. 224 Dinstein has seemingly agreed, stating that a NIAC will end when 'insurgents are roundly beaten', 225 and that less intense violence may continue after the legal end of a NIAC. 226 Bartels has undertaken a more extensive analysis, concluding that once the threshold requirements are no longer satisfied, a NIAC will cease, 227 though he maintains that, rather than stating the law, he is testing a hypothesis. 228 Going further, Milanović has described this approach as a 'general principle', identifying the disapplication of IHL 'once the conditions that trigger its application in the first place no longer exist'. 229 Clearly contrary to the approach of the ICTY, Milanović consciously posits his formulation as '[a]nother option', that is 'more logical from a purely IHL standpoint'. 230 Nonetheless, in describing its application, he uses the conditional tense ('the only legally relevant question would be whether the threshold continues to be satisfied' 231 ) suggesting that, desirable as this principle may be, its existence cannot be confirmed. Jinks asserts a similar 'general rule' based on Article 6 of the fourth Geneva Convention, which proclaims the application of IHL until 'the general close of hostilities'. 232 Clearly the law is unsettled and open to interpretation. One may adopt a broad interpretation that a NIAC will continue until agreement is reached or the NSA party no longer exists, thereby prolonging the application of IHL and the modified IHRL protections of those subject to it. The alternative is a restrictive interpretation that a NIAC ends when hostilities fall below the intensity threshold, limiting the temporal scope of IHL. Regardless, the choice of interpretation has important implications for the analysis of US drone strikes in Yemen (and Somalia) as the fighting has occurred over long periods with oscillating intensity, likely falling below the Tadić threshold, before satisfying it again subsequently.
As stated above, 'terrorist activities' are distinct from NIACs when they do not satisfy the Tadić test. 233 There is no clear State practice to suggest when terrorist activities are intense enough to become a NIAC, though ICTY jurisprudence is informative. Considering the facts presented above, it seems likely that fighting between AQAP and the Yemeni government fell below the intensity threshold between 2013 and March 2015: there are gaps of many months between attacks (there were 7 reported in 27 months) and the number of those killed is significantly less (an average of approximately 5 per month during the period). 234 Nonetheless, it has been suggested elsewhere that six attacks, globally, over a period of six years is sufficiently intense. 235 This suggestion was made in support of the notion that the global acts of al-Qaeda were part of a single NIAC, a notion that has-within this work-been argued to be incorrect under international law. 236 Even if the underlying assertion as to intensity is correct in terms of the time-frame of attacks (one per year), 237 the six attacks referred to 238 produced a combined death-toll of 3525 (an average of 49 per month) ten times that of the relevant period in Yemen, suggesting those in Yemen would still remain far below the intensity threshold.
In the period from March 2015 until the present, the fighting may have crossed the intensity threshold again, though less overtly than in 2011-2012. While the death-toll remains comparatively low, more territory has been captured, government buildings have been destroyed and the response has been military; all indicative of requisite intensity. 239 Further, the Security Council condemned the actions of AQAP, referring to the fighting in terms of IHL, 240 though it subsequently appeared to describe it as terrorism, with no reference to IHL. 241 Therefore, the violence in Yemen arguably intensified enough during this period to satisfy the intensity threshold.
It is now necessary to look at the organization of AQAP. The primary requirements-of a 'command structure', 242 and the ability to speak 'with one voice' 243 -appear satisfied: AQAP was founded with a hierarchy containing defined roles, including leader, deputy leader, military commander 244 and field commander. 245 Upon the death of original leader Nasser al-Wuhayshi in 2015, a new leader, Qassim al-Raimi, took over straightaway. 246 The group has a bimonthly magazine, Sada al-Malahim, 247 an indication of it having a single, organized voice. The ICTY 234 See above, nn. 206-214. 235 Dalton (2006), p. 527. 236 See above, n. 160. 237 This is purely work of the devil's advocate; the present author submits that, even if attributable to a single NSA, six attacks over six years are manifestly isolated and sporadic and below the intensity threshold. has emphasized the ability to 'formulate […] military tactics' 248 which appears satisfied by the military successes of AQAP in capturing various Yemeni cities. Several other indicators of organization proposed by the ICTY also appear satisfied by AQAP: Sada al-Malahim arguably represents the issuing of communiqués 249 as does AQAP's release of public statements. 250 It has erected checkpoints in occupied regions, 251 another indicative factor. 252 Finally, AQAP has engaged in negotiations and mediations with the Yemeni government, 253 further indicating that the group surpasses the organization threshold. 254 As a result of the above analysis, a nuanced conclusion as to the ongoing NIAC in Yemen can be made. It is submitted that a NIAC existed in Yemen between the government and AQAP beginning in May 2011, lasting until June 2012. Based on available data, and in light of the Yemeni government's purported consent to US drone strikes, 255 this would mean at least 64 US drone strikes in Yemen fall within the scope of IHL. 256 It also means that one strike was certainly outside the scope of IHL as it happened in 2002. 257 After June 2012 there is less clarity, both in terms of facts and law. If a restrictive interpretation of the end of NIACs is adopted then it seems that until March 2015 there was likely no NIAC, or at least there is no way of concluding unequivocally that there was one. In this scenario 85 drone strikes most likely occurred outside the scope of IHL. 258 However, if the broad interpretation of the ICTY is adopted, the ongoing, less intense violence and absence of a peace agreement will have sustained the NIAC, bringing those 85 strikes within the remit of IHL. Finally, from March 2015 until the time of writing it seems more likely than not that a NIAC existed under the restrictive interpretation and certain that one existed under the broad one, therefore bringing a further 48 drone strikes into the remit of IHL. 259 248 Limaj, above n. 22, para. 129. 249 Ibid., paras. 101-103. 250 Al-Moshki, 'AQAP Apologizes for Hospital Attack in Ministry of Defense Operation', Yemen Times (Sana'a, Yemen 24 December 2013), http://www.yementimes.com/en/1740/news/3270/AQAPapologizes-for-hospital-attack-in-Ministry-of-Defense-operation.htm. 251 Al-Qalisi, 'Al Qaeda Seizes two Districts in Yemen's Abyan Province', National (Abu Dhabi 3 December 2015), http://www.thenational.ae/world/middle-east/al-qaeda-seizes-two-districts-in-yemensabyan-province. 252  This conclusion means that either there have been 86 drone strikes in Yemen carried out outside of the IHL framework, or just the individual one that occurred in November 2002. In either case, it has been demonstrated that the existence of a NIAC between AQAP and the Yemeni government and the latter's consent to US intervention does not bring the entirety of the US drone programme within the scope of IHL.

AQAP and the US
Of course, all drone strikes will governed by IHL if the US was in its own NIAC with AQAP, and this possibility will now be considered. It has already been proposed that AQAP has the requisite organization, so that part of the analysis is unnecessary. In addition, it was shown above that al-Qaeda cannot be understood to be a globally organized group such that all terrorist acts carried out by regional franchises in its name can be attributed to it. 260 Therefore, in terms of intensity, we must consider only those acts carried out against the US perpetrated by AQAP (or its predecessor, AQY).
Farley has argued succinctly that violence between the US and AQAP is insufficiently intense to be a NIAC, 261 as has Heller, 262 and it is submitted that these arguments are correct: however, further analysis will help reinforce this conclusion. Attacks upon the US by AQAP do not appear to satisfy the intensity requirement of a NIAC, when considered through the lens of the indicative factors that have informed the ICTY. There are three attacks or attempted attacks, targeted against the US, that appear unquestionably attributable to AQAP or AQY: the attack by suicide-boat bomb on the USS Cole in October 2000, in which 17 sailors were killed 263 ; the failed December 2009 bombing of a flight to Detroit 264 ; and the failed October 2010 bombing of US-bound cargo planes. 265 In addition, it has been claimed that AQAP 'radicalized' those responsible for the 2009 shooting at Fort Hood, Texas and the 2010 attempted bombing of Times Square. 266 It is submitted that the 'radicalization' of an individual provides an insufficient connection to the NSA under the organization threshold for their actions to be considered within an analysis of intensity, in the manner mandated in the Boškoski judgment. 267 Radicalized individuals like these are not under the auspices of a NSA's command structure for purposes of planning armed engagements. Finally, Terrill has suggested that the attacks of 11 September 2001 against the US can be attributed to AQAP; he claims a 'subtle Yemeni link' in that some of the hijackers were 'Saudis of Yemeni descent', 268 Osama Bin Laden's fourth wife was Yemeni, 269 and he (Bin Laden) had funded some al-Qaeda activities in the country. 270 It is submitted that this is insufficient to satisfy the organization requirement of a NIAC that would be necessary in order to include this attack within the present intensity analysis.
Simultaneously, of course, the US has carried out drone strikes against AQAP which could be suggested to push the total violence over the NIAC intensity threshold. It has been argued that the definition adopted by the ICTY in the Tadić case, in focusing on 'protracted armed violence between governmental authorities and organized armed groups' 271 emphasizes the need for violence from at least two parties. 272 This understanding is implicit in common Article 3 of the Geneva Conventions, by its application to 'each Party to the conflict'. Therefore, to pass the threshold, intense violence must have occurred as a result of actions on both sides of the conflict which, in this case, has not occurred, the only deaths being 17 US sailors in a 16-year period.
As a result of this analysis, it is clear that there have been significant periods amounting to NIAC, between the government of Yemen and AQAP, into which the US has been invited to intervene. It is difficult to say with certainty how many strikes fall within this NIAC due to the lack of clarity surrounding the end of NIAC. As stated above, a restrictive interpretation of the end of NIAC means that 86 out of 198 strikes (and between 328 and 492 killed out of between 904 and 1308) occurred outside of NIAC. 273 Conversely, under a permissive interpretation all bar one strike and six deaths occurred within NIAC. Nonetheless, despite the ambiguities of the law producing two distinct figures, it remains clear that drone strikes have not been pursued in a legal vacuum and that claims as to the applicable paradigm of international law can be made. While the term 'outside of active hostilities' seems to flatten out the contours of the fight against AQAP such that drone strikes in the region are all presumptively governed by IHL, the reality is that there are strikes that formed part of the NIAC and one or more that did not. Those in the latter category cannot but be considered through the lens of IHRL obligations given their ordinary meanings, without being interpreted through IHL.
strikes have been carried out against al-Shabaab exclusively; 275 beyond sharing an ideology, al-Shabaab is not part of any NSA with which the US is engaged in a NIAC elsewhere, so there is no question of strikes in Somalia being part of a spillover conflict. Thus, it is necessary, as with Pakistan and Yemen, to determine whether there exists a NIAC between the Somali government and al-Shabaab or, alternately, between the US and al-Shabaab. This question has been given insufficient treatment previously; Alasow has asserted the applicability of common Article 3 to the broader conflict in Somalia-not just that involving al-Shabaabbut without a detailed consideration of intensity and organization. 276 It is therefore submitted that it cannot be viewed as conclusive of the issue, necessitating further engagement, which will be carried out presently.

Al-Shabaab and the Somali Government
It is impossible to examine the existence of a NIAC between al-Shabaab and the Somali government without first considering its recent historical context. After the fall of the Barre regime in 1991, Somalia was without central government for many years, hosting ongoing violence between rival militia and warlords. 277 It is beyond the remit of this work to consider whether that situation was a NIAC though it has been described as a 'civil war'. 278 In 2004, the Islamic Courts Union (ICU) and the Transitional Federal Government (TFG) emerged as two major powers in Somalia, polarizing the conflict, the ICU, backed by Eritrea, and the TFG, by Ethiopia. 279 The violence appears to have the characteristics of a NIAC as it involved automatic weapons, bombs, mines, mortars, grenade launchers and antiaircraft guns, 280 and resulted in hundreds of deaths and the displacement of thousands of civilians. 281 Additionally, negotiations between the two groups 282 and the command framework of the ICU 283 suggests the organization threshold was satisfied. Between December 2006 and January 2007 the TFG defeated the ICU, retaking territory the group had captured. 284 The ICU returned to an insurgency style conflict with the TFG, but in September 2007 senior ICU members were key in forming the Alliance for the Re-liberation of Somalia (ARS) 285 which reached a peace agreement with the TFG in Djibouti in June 2008, 286 thereby concluding the NIAC, even under the broad Tadić criteria for cessation. 287 Al-Shabaab germinated within the ICU but cannot be said to have carried on the ICU's NIAC with the TFG. It was at first 'a relatively marginal insurgent group' 288 and was ideologically distinct from the ICU. 289 Thus it is necessary to consider the conflict between al-Shabaab and the government in isolation when determining its classification.
This determination is difficult due to a paucity of data, particularly during the early stages of the conflict, but it is possible to build enough of a picture to make tentative conclusions. It has been asserted that by November 2007, al-Shabaab carried out 55 percent of attacks against the government, 290 though it is unclear what this actually amounts to. The early period was characterized by 'koormeer' techniques, in which al-Shabaab would capture government outposts or undefended cities for two or more days, enforcing Sharia law, before departing. 291 294 These instances suggest that, though labelled a terrorist organization, the actions of al-Shabaab were more than mere 'terrorist activities'. 295 Individual acts of violence lasted for long periods and occurred over a number of years, covering a large geographical area and resulting in numerous casualties. The governmental response involved the deployment of troops rather than police, and, during the initial period of the violence, 7650 Africa Union Mission in Somalia (AMISOM) troops were deployed. 296 As early as May 2008, the UN Security Council referred to the situation in the language of IHL, indicating violence satisfying the intensity threshold. 297 As a result of this, it is submitted that, at some point during 285 298 During the subsequent period al-Shabaab twice attempted to capture Mogadishu, the capital city. In May 2009 the group (in collaboration with another NSA, Hizb al-Islam) engaged in a protracted battle for the city, against TFG and Ugandan troops, 299 involving heavy weapons by both sides, including mortars, 300 tanks and armoured personnel carriers (APCs). 301 The violence resulted in the displacement of 60,000 civilians 302 and up to 200 deaths. 303 A further offensive against the city, in August 2010 involved at least 1800 al-Shabaab fighters and an opposing AMISOM force of infantry and tanks. 304 These examples suggest the fighting between al-Shabaab and the government remained sufficiently intense throughout 2010.
Between August 2011 and the present, al-Shabaab has returned to insurgency tactics, rather than engaging in large battles to capture towns, and the government and AMISOM have recaptured many areas. 305 The group has continued to carry out suicide and bomb attacks, which have often resulted in large numbers of fatalities. 306 The argument could be made that these attacks are now sporadic enough to be terrorist only, falling short of the intensity of a NIAC. However, in responding to al-Shabaab, the AMISOM military force currently comprises 22,126 troops, and its mission profile emphasizes its military Based on this analysis it is submitted that al-Shabaab has the characteristics of a sufficiently organized group, and that the violence between it and the government meets the intensity threshold recognized within international law. Therefore, a NIAC has existed in Somalia between these parties since 2007 or 2008 and is ongoing at present. As the present author has asserted elsewhere, 321 the US was invited to join the NIAC by the Somali government in 2007, rendering it a party to the NIAC. Nevertheless, at least since 2013 consent given has been specifically restricted to drone strikes against non-Somali fighters. 322 Prior to this, six drone strikes were carried out that are, as such, quite likely within the NIAC. Out of 21 drone strikes since 2013, there have been three that directly targeted fighters whose nationality was certainly Somali. 323 Therefore these strikes may be outside of the NIAC unless specific consent was received. They will therefore likely only be subject to IHL if there is a separate NIAC between the US and al-Shabaab. In three other cases, airstrikes involving drones targeted large groups of fighters, in one instance reportedly killing up to 200. 324 Though foreign fighters feature in al-Shabaab, they in no way comprise a majority of the group, 325 therefore, in cases of drone strikes against large groups, it would have been very difficult for the US to limit their targeting to non-Somali fighters. Thus, it is probable, though not certain, that these strikes also fall outside the scope of this NIAC, though again, specific consent may have been given. Finally, there have been 15 strikes since 2013 in which it has been impossible to confirm the nationality of those targeted, 326 and as such no determination can be made either way as to the inclusion or exclusion of the strikes within the NIAC. Therefore, overall at least three, and possibly as many as 21, drone strikes cannot be included within the NIAC between the Somali government and al-Shabaab.

Al-Shabaab and the US
The final aspect of this analysis will consider whether a NIAC exists (or existed) between the US and al-Shabaab, thereby bringing a greater number of drone strikes into the remit of IHL. It has already been shown that al-Shabaab easily satisfies the organization threshold, so it is necessary only to consider the intensity of the violence.
Since 2011, the US has carried out 27 drone strikes against al-Shabaab 327 but, as stated previously, 328 the actions of a single party cannot, in and of themselves, 321 Byrne (2016), p. 117. satisfy the intensity threshold of a NIAC. Al-Shabaab has engaged in international acts of violence, but these have primarily been in Kenya, and none have been against the US. The only instances of actions taken by al-Shabaab against the US have been its defence against an aborted Navy SEAL attack upon one of the group's commanders, 329 and an al-Shabaab video that called for 'lone wolf' attacks to be carried out in shopping centres in the US, Canada and the UK, though the threat was described as not credible. 330 Therefore it is impossible to conclude that there is a NIAC between al-Shabaab and the US.
A lack of information as to the identities and nationalities of those killed makes it difficult at this stage to confirm with certainty which drone strikes occurred within a NIAC, though some conclusions can be made. It is likely that all six strikes prior to 2013 were within a NIAC. After 2013, three were not within the NIAC absent additional consent, as they have been confirmed to have targeted Somalis. A further three are probably outside of a NIAC, again absent additional consent, as they targeted so many individuals in one instance as to have made it highly likely that Somalis would be killed. Finally, the remaining 15 may have been carried out within the NIAC, though if it transpires that Somali nationals were targeted then these too will fall outside the NIAC. Therefore, once again it is clear that the presumption that IHL applies to drone strikes in Yemen, manifest within US documents classifying these regions as 'outside areas of active hostilities', is mistaken.

Conclusion
This in-depth application of the law governing the classification of armed conflict to the facts of covert US drone strikes has demonstrated that these operations have been carried out both during periods of NIAC and periods which cannot be classified as such. Adopting an objective view of the facts surrounding particular drone campaigns, it has been possible to determine when situations became NIAC and therefore periods during which the lawfulness of drone strikes should be analyzed using IHRL interpreted with reference to IHL and when, conversely, they should be analyzed according to the ordinary meaning of IHRL obligations. Therefore, it is incorrect to presumptively identify IHL as being the relevant paradigm for the analysis of these strikes, as has been done by the US through its designation of their having occurred 'outside areas of active hostilities'. This phrase has no impact upon the character of the fighting to which it applies, and therefore has no bearing on the legal paradigm applicable.
From a purely technical legal perspective, viewing the drone programmes in Pakistan, Yemen and Somalia as a whole, it has been possible to produce an overall picture of the extent to which drone strikes have been undertaken within armed conflict. It is submitted that, by July 2016, 515 strikes appear likely to have been undertaken as part of a NIAC and so may be judged according to IHRL interpreted in light of IHL. Conversely, 46 drone strikes are very likely to have been undertaken outside of NIACs and so their lawfulness must be assessed through the lens of IHRL alone and its attendant restrictions on the use of lethal force. There are 88 other strikes for which the answer is unclear, though a conclusion may be found if the nationality of those targeted can be confirmed (Somalia), if their militant affiliation can be determined (Pakistan) and if the law on the end of NIAC can be clarified (Yemen).
The fact that determinations can be made as to the existence of NIAC in regions classified as 'outside areas of active hostilities' demonstrates the continued relevance of the international law governing the application of IHL. The framework is certainly not straightforward-particularly in terms of the end of NIAC-but it is submitted that it can nonetheless be successfully applied to the facts of modern and ultra-asymmetrical confrontations between NSAs and technologically advanced States. Thus there is no practical need to move away from this method of classifying conflicts. The system of identifying armed conflicts should be continued to be used and should certainly not be replaced with the use of obfuscatory phrases and the presumptive application of IHL in order to expedite the permissible use of military force and undermine relevant IHRL protections.
More generally, it has been shown that the concept of operations 'outside areas of active hostilities' is devoid of normative content in law. Its use by the US centres on the presumptive application of IHL to situations in which drones have been used, in a manner irreconcilable with the objective identification of relevant legal paradigms that is clearly required by international law. Furthermore, the designation of a region as being 'outside areas of active hostilities' also provides no basis for IHL to be applied outside of armed conflict, for instance, in the form of 'self-defence targeting', and it should not be regarded as having such a capacity. It is submitted that these approaches to the application of IHL must be resisted as they greatly undermine the protections of those affected by uses of force. Armed conflict, and the higher levels of force that are permitted as a consequence of the operationalization of IHL, must remain exceptional, and should not be allowed to become normalized by being presumed to exist.
The presumptive application of IHL in Pakistan, Yemen and Somalia is not only contrary to the law governing the identification of armed conflict, it also goes counter to the global trend surrounding the regulation of hostilities towards the increased expansion of IHRL obligations into armed conflict. It has been convincingly articulated by Daragh Murray and others that situations of armed conflict that are not 'active hostilities' will be governed by IHRL as the primary framework of international law. 331 Under this approach, far from opening up the scope for permissive uses of force, classifying a situation as not 'active hostilities' would prohibit certain targeting practices that characterize US drone programmes, such as status-based targeting. 332 It is therefore possible that the methods of combat permitted by US policy on operations 'outside areas of active hostilities', which are more restricted than those available under IHL normally, 333 may ultimately be unlawful in some cases, by virtue of the very fact that they are conducted 'outside areas of active hostilities', therefore rendering IHRL the primary paradigm. This is an enticing line of argument and is one that must be pursued in the future.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.