In recent years there have been calls for the establishment of a variety of arrangements to provide civilians with a modicum of safety in a number of countries experiencing armed conflicts. These have included “safe zones” in LibyaFootnote 1 and “safe havens” or “buffer zones” in Syria (terms used interchangeably).Footnote 2 While the suffering of civilians in these contexts was severe, the calls were not always without ulterior political motives, whether in terms of stemming refugee flowsFootnote 3 or providing some degree of support and legitimacy to opposition forces operating from the proposed areas. Political motives are starkly evident in the “de-escalation zones” established in Syria pursuant to the Astana Agreements of May 2017. Their purported objectives included creating ceasefires between “moderate” opposition groups and the government of Syria; enhancing humanitarian access; facilitating the rehabilitation of basic infrastructure; the creation of conditions to deliver medical aid and meet the basic needs of civilians; and the safe and voluntary return of refugees and internally displaced persons.Footnote 4 While the precise conditions and local dynamics differ in the various zones, in practice the areas have not led to a reduction in violence or to enhanced humanitarian access. On the contrary, locations that fall within these areas, most notoriously Eastern Ghouta, have witnessed some of fiercest fighting in recent months, lending credence to the arguments that the zones were established as a “war management strategy” aimed at weakening the opposition.Footnote 5
Motivations aside, it is important to appreciate the complexities that establishing and operating such areas entail. Their feasibility and success depend on numerous factors, starting from the political will of belligerents to agree to them or, absent such agreement, that of the Security Council and third States to create them. Moreover, the decision to establish so-called “safe areas” is just the beginning. Their implementation in practice raises numerous legal and practical challenges.
The present article focuses on one of the key framing issues that must be addressed when considering the establishment of “safe zones”: the international legal framework. This dimension, which at times is misrepresented or glossed over, is important per se, and also because it highlights some central operational issues that must be addressed if such zones are to meet the promise of safety to those seeking refuge there.
A number of different areas of international law are of relevance to the establishment and operation of “safe areas” in situations of armed conflict. They include international humanitarian law (IHL), also known as jus in bello, the body of law regulating the conduct of hostilities and protecting those not or no longer taking direct part in hostilities; jus ad bellum, the rules regulating resort to the use of force; and also refugee law and international human rights law. The present article will consider the first two areas of law in some detail, and flag some refugee and human rights law concerns raised by “safe areas”. It will also highlight some additional considerations relating to the broader regulatory framework, including Security Council mandates for the establishment of such zones, if belligerents fail to agree to do so, and some operational challenges raised by their implementation in practice.
International humanitarian law
The expression “safe areas” is not used in any treaty; instead, IHL refers to “protected zones”. This term is preferable as it refers to concepts that are defined in international law, but also because it highlights the reality that while the areas may have been accorded special protection in law, this does not necessarily translate to safety in practice for the people seeking refuge.
Three points must be highlighted at the outset. First, as a matter of law, parties to armed conflicts must respect and protect the civilian population and wounded and sick combatants at all times, regardless of whether protected zones have been established. As a matter of practice, it is precisely because belligerents are not complying with this obligation, but are instead targeting civilians, conducting hostilities in an indiscriminate manner or forcibly displacing civilians, that the creation of such zones is considered.
Second, the general rules of IHL regulating the conduct of hostilities continue to apply even if protected zones are established, and remain of fundamental importance both for those who have sought shelter within the zones and for those who have not. These rules are outlined in more detail below.
Finally, the provisions of IHL on protected zones protect the zones and not the people who seek shelter therein. The people must be respected and protected independently, and the zones are merely a way of implementing such protections as effectively as possible.Footnote 6
The aim of the present article is not to exhaustively analyze the relevant rules of IHL, but rather to draw out the key elements of these rules. IHL envisages various types of protected areas in international armed conflicts. They differ slightly, essentially in terms of the categories of people who may access them, but their objective is the same: to create areas where the wounded and sick and the civilian population can seek shelter from hostilities. The key aspect to bear in mind is that, unusually for IHL treaties, which do not ordinarily include merely exhortatory language, the relevant provisions do not require belligerents to establish protected zones. Instead, they merely note the possibility for them to do so. Zones established by one side will benefit from special protection only if and when they have been recognized by its opponent. During the negotiations of the 1949 Geneva Conventions, despite recognizing the humanitarian value of the proposed protected zones, States were unwilling to require their establishment and recognition.Footnote 7 This approach did not change during the negotiations of the 1977 Additional Protocols.
Geneva Conventions I and IV: Hospital and safety zones and localities
Types of protected zones
In situations of international armed conflict, the First Geneva Convention for the Amelioration of the Condition of the Wounded and Sick in Armed Forces in the Field (GC I) and the Fourth Geneva Convention relative to the Protection of Civilian Persons in Time of War (GC IV) foresee the possibility of establishing hospital and safety zones and localities.Footnote 8 As noted above, their special protected status is dependent on agreement between belligerents.
With regard to wounded and sick members of the armed forces,Footnote 9 Article 23 of GC I provides that
[i]n time of peace, the High Contracting Parties and, after the outbreak of hostilities, the Parties to the conflict, may establish in their own territory and, if the need arises, in occupied areas, hospital zones and localities so organized as to protect the wounded and sick from the effects of war, as well as the personnel entrusted with the organization and administration of these zones and localities and with the care of the persons therein assembled.
Upon the outbreak and during the course of hostilities, the Parties concerned may conclude agreements on mutual recognition of the hospital zones and localities they have created. They may for this purpose implement the provisions of the Draft Agreement annexed to the present Convention, with such amendments as they may consider necessary.
The Protecting Powers and the International Committee of the Red Cross are invited to lend their good offices in order to facilitate the institution and recognition of these hospital zones and localities.
In essentially identical terms, Article 14 of GC IV foresees the possibility of establishing similar zones for wounded and sick civilians. As a matter of law, the stark distinction between wounded and sick members of the armed forces and wounded and sick civilians in the 1949 Geneva Conventions was abandoned in Additional Protocol I (AP I) of 1977, where reference to “the wounded and sick” covers both civilians and members of the armed forces.Footnote 10 As a matter of practice, nothing precludes a hospital zone from accommodating wounded and sick combatants and civilians. It is important to note that admission to hospital zones and localities must be granted without adverse distinction and therefore also to enemy wounded and sick, including combatants.
In addition to “hospital zones”, Article 14 of GC IV also refers to the possibility of establishing “safety zones and localities”. These are envisaged as areas where members of the civilian population considered particularly vulnerable and unlikely to pose a threat to the enemy can seek shelter. Article 14 refers to “aged persons, children under fifteen, expectant mothers and mothers of children under seven”, but it seems safe to assume that all civilians may seek shelter in such zones and localities, provided they do not pose such a threat. As is the case for hospital zones, the principle of non-discrimination requires that access to protected zones also be granted to vulnerable civilians of enemy nationality.Footnote 11 In practice, the various types of protected area are likely to be combined, accommodating wounded and sick civilians and members of the armed forces, as well as other vulnerable people.
At the time of the adoption of these provisions in 1949, wounded and sick members of the armed forces, those treating them, and medical establishments had long been entitled to protection.Footnote 12 As far as these persons and facilities were concerned, the creation of hospital areas as envisaged in GC I would simply make it easier to give effect to these protections in practice. The position of civilians was different, however. It was not until the adoption of GC IV that similar protections were extended to wounded and sick members of the civilian population, those treating them, and civilian medical establishments.Footnote 13 Beyond these, the civilian population and civilian objects benefited from few express protections from the effects of hostilities in IHL treaties.Footnote 14 Protections for civilians from the effects of hostilities were not codified until the adoption of the Additional Protocols in 1977. Before then, hospital and safety zones, if established and recognized, were therefore an important way of enhancing the protections to which civilians were entitled as a matter of law, and of reducing their actual exposure to risks.
Article 23 of GC I and Article 14 of GC IV envisage the possibility of creating the hospital and safety zones in times of peace. Although this never appears to have occurred in practice, doing so would enable States to prepare the areas with the necessary equipment and supplies. Even if the zones are only established after the outbreak of hostilities, assembling the wounded and sick in a specifically prepared area can facilitate their treatment and can also help address some of the adverse effects of armed conflict, such as shortages of medical supplies or a breakdown of health services.Footnote 15
As highlighted above, while a belligerent may establish hospital or safety zones and localities, doing so will have no effect unless and until they are recognized by its opponent. Importantly, however, and as also already noted, the absence of agreement does not deprive the people seeking shelter in such zones of the protection to which they are entitled under the general rules of IHL on the conduct of hostilities.
Draft Agreements on hospital and safety zones and localities
The likelihood of recognition of protected zones will depend on a number of factors. Key among them are the steps taken to ensure that the zones are in fact exclusively humanitarian and are not likely to be abused. Annexed to the GC I and GC IV are virtually identical Draft Agreements that provide guidance to parties setting up hospital and safety zones and localities. It is regrettable that these Draft Agreements have received little attention in the literature and have not been considered in the recent discussions on the possible establishment of protected areas, because they set out key conditions that such zones should meet in order to serve their intended purpose of shielding certain categories of people who are hors de combat from the effects of hostilities.
The Draft Agreements contain three key sets of provisions. These warrant highlighting, as the underlying considerations are relevant to all protected zones: those established by agreement as foreseen by IHL, and those established by other modalities, discussed below.
The first set of provisions aim to ensure that those accommodated in protected zones are in no way involved in hostilities, thus preserving the actual and perceived exclusively humanitarian purpose of the zones and not jeopardizing their protected status. They include:
• the requirement that the zones be reserved exclusively for the wounded and sick, civilians, people administering the zones and providing medical care, and people whose permanent residence is within such zones;
• the corollary obligation for those establishing and operating the zones to take all necessary measures to prohibit access to people who are not entitled to reside there; and
• the requirement that no one residing in the zones undertake, within or outside the zones, any activity directly connected with military operations.
Although neither the Draft Agreements nor the Geneva Conventions elaborate this point, the responsibility for ensuring that a zone complies with these conditions lies with the party establishing it. Doing so may be onerous and is likely to require considerable capacities in terms of types of personnel and also the provisions of goods and services. Security personnel may be required to ensure that armed elements do not enter the zones, and to maintain law and order within them.Footnote 16 It will also be necessary to provide food, health care and, if the zones are used for prolonged periods, education. The greater the number of residents, and the longer the zones are employed, the more onerous their operation will be. Although they are not the type of protected zone envisaged by IHL, the experience of the United Nations Mission in South Sudan (UNMISS) with the “protection of civilians sites” discussed below highlights some of the operational challenges involved.
The second set of provisions in the Draft Agreements are measures that aim to enhance the security of the zones themselves. They include requirements that:
• the zones only constitute a small part of the State's territory. Larger zones are likely to impair the opponent's capacity to conduct hostilities and thus may undermine the safety of the zones themselves;
• they be distant from and free of all military objectives, not situated in areas likely to become significant for the conduct of the war, and not defended by military means;
• access roads to the zones not be used for military purposes;
• hospital zones be marked by red crosses/crescents, and other protected zones and localities by other agreed-upon signs.
The final set of provisions in the Draft Agreements are supervisory measures to ensure that the zones operate as envisaged. The party recognizing a zone is entitled to demand that an independent body have access to examine the zone and confirm that it complies with the basic requirements set out above. If shortcomings are found, the party that established the zone must rectify them within a prescribed period of time. If it fails to do so, its opponent may declare that it no longer recognizes the area as protected. This would not, however, affect the protections to which people accommodated in the areas are entitled under the general rules of IHL on the conduct of hostilities.
Geneva Convention IV: Neutralized zones
GC IV foresees the establishment of a further type of protected zone: neutralized zones. Article 15 of GC IV provides that
[a]ny Party to the conflict may, either directly or through a neutral State or some humanitarian organization, propose to the adverse Party to establish, in the regions where fighting is taking place, neutralized zones intended to shelter from the effects of war the following persons, without distinction:
(a) wounded and sick combatants or non-combatants;
(b) civilian persons who take no part in hostilities, and who, while they reside in the zones, perform no work of a military character.
When the Parties concerned have agreed upon the geographical position, administration, food supply and supervision of the proposed neutralized zone, a written agreement shall be concluded and signed by the representatives of the Parties to the conflict. The agreement shall fix the beginning and the duration of the neutralization of the zone.
The drafters of GC IV envisaged three principal differences between hospital/safety zones and neutralized zones. The former were intended to be located at a distance from the combat zone and to constitute a longer-term arrangement for certain categories of particularly vulnerable civilians. Neutralized zones, instead, would be in areas of active fighting to provide temporary shelter to wounded and sick combatants and civilians and the entire civilian population. In practice, on the occasions when protected zones have been established, they have taken the form of neutralized zones: located in areas of combat and accommodating both civilians and the wounded and sick.Footnote 17
While, strictly speaking, the Draft Agreements annexed to GC I and GC IV relate to hospital and safety zones and locations, the measures they set out are equally relevant to neutralized zones – if not more so, considering the latter are located in areas of active combat.
Additional Protocol I: Demilitarized zones
The range of protected zones contemplated by IHL is further expanded by AP I, which foresees the possibility of establishing demilitarized zones.Footnote 18 The aim of these zones is similar to that of the neutralized zones of Article 15 of GC IV: to place localities or zones and their non-combatant population outside the theatre of war.Footnote 19 While the arrangements in the Geneva Conventions do this by foreseeing the establishment of zones to which civilians and the wounded and sick can relocate, demilitarized zones operate by “fencing off” areas from military operations for the protection of all civilians. The party establishing a demilitarized zone must ensure that it is not used for hostile activities – in a broad sense of the term, as will be seen below – and, it if recognizes the zone, its opponent must refrain from extending “military operations” – also a broad notion – to the zone.Footnote 20
Demilitarized zones are established by agreement between belligerents in relation to areas that fulfil a number of conditions:
(a) all combatants, as well as mobile weapons and mobile military equipment, must have been evacuated;
(b) no hostile use may be made of fixed military installations or establishments in the zones;
(c) no acts of hostility may be committed by the authorities or by the population in the zones; and
(d) any activity linked to the military effort in the zones must have ceased.Footnote 21
The agreement to establish demilitarized zones can be oral or in writing. It must specify as precisely as possible the location and limits of the zones, and, if necessary, prescribe the methods for supervising compliance with the conditions set out above.Footnote 22
Belligerents are precluded from extending their military operations to areas to which they have conferred the status of demilitarized zones.Footnote 23 The party recognizing a demilitarized zone is released from its obligations under the agreement in case of a material breach of one of the conditions set out above, or if the zone is used for purposes related to the conduct of military operations.Footnote 24 Should that occur, protections for civilians and the wounded and sick under the general rules on the conduct of hostilities will nevertheless continue to apply.
Non-international armed conflicts
The rules on the various protected zones outlined above are only found in treaties that regulate international armed conflicts. The absence of similar provisions in relation to non-international armed conflicts is not significant, as nothing precludes parties to such conflicts from entering into agreements with a similar effect. In fact, Article 3 common to the Geneva Conventions specifically refers to the possibility of parties concluding special agreements to bring into effect other provisions of the Conventions.
Determining whether a conflict is international or non-international in character can be extremely difficult legally and factually, and is frequently a politically charged question. The conflict in the former Yugoslavia in the 1990s was a case in point. In the interest of reaching agreement on the establishment of protected zones, belligerents agreed, through the good offices of the International Committee of the Red Cross (ICRC), to conclude ad hoc agreements to that effect, without entering into what would have been an inconclusive debate on the nature of the conflict that would have stymied the establishment of the zones.Footnote 25
Protected zones: Their establishment and operation in practice
The defining element of the protected zones foreseen by IHL is the need for agreement between belligerents. This need for agreement is also a key reason why so few have been established. The Geneva Conventions suggest creating protected zones in peacetime, but as noted, this does not appear to have ever occurred. Understandably, reaching agreement after the outbreak of hostilities is extremely difficult. While belligerents could engage in direct negotiations, trusted, neutral intermediaries are likely to play an important role in helping them to reach agreement. The provisions of GC I and GC IV on hospital zones specifically mention the potential role of the protecting powers and of the ICRC in this regard.Footnote 26 The ICRC has played a central role in initiating negotiations and facilitating the conclusion of agreements for the establishment of the majority of the few protected zones that have been created by agreement.Footnote 27
IHL does not specify the format that an agreement should take,Footnote 28 but written agreements have the obvious advantage of clarity. More important than its format are the key issues that an agreement should address, including measures to ensure the exclusively humanitarian nature of protected zones and to enable them to be accurately identified, as outlined in the Draft Agreements annexed to GC I and GC IV.
None of the treaty provisions address the central question of who is responsible for operating the protected zones, including ensuring that they meet the conditions stipulated in the agreement, providing basic services, and maintaining law and order. Although it makes sense for this responsibility to lie with the party to the conflict that establishes the zones, it is interesting to see that in a significant number of the few protected zones that have been established by agreement, it was the ICRC that assumed all these roles.Footnote 29 This was probably feasible in the circumstances for a number of reasons, starting from the limited scope of the areas in terms of size and duration.
While humanitarian actors are likely to make important contributions to the operation of protected zones, they do not have the mandate to carry out all necessary activities, including, most notably, the screening and disarming of people entering the zones, or the maintenance of law and order within them. Moreover, they may have reservations about being involved in zones that have not been set up with the agreement of all belligerents, as doing so could give rise to a misleading impression of the safety of the zones, which would not be warranted if the opponent had not recognized them. It could also undermine perceptions of the organization's neutrality, and put its staff at risk.Footnote 30 Significant reservations are also likely to arise when the zones have been established to prevent people from crossing borders to seek asylum.Footnote 31 The supervision of protected zones to ensure that they meet the key conditions agreed to is a different matter and a role that should be carried out by a mutually trusted intermediary, which can and, as noted above, frequently has been a humanitarian organization.
There have been few instances of protected zones of any kind being established as foreseen by IHL since the adoption of the 1949 Geneva Conventions.Footnote 32 The most frequently cited examples include:
• a number of temporary arrangements established by the ICRC, including in Dhaka in 1971; in Nicosia in 1974; in Saigon and Phnom-Penh in 1975; and in Nicaragua in 1979;Footnote 33
• some areas in Port Stanley, as well as the “Red Cross box” during the Falklands/Malvinas conflict in 1982;Footnote 34 and
• Osijek and other hospitals in the Dubrovnik area in 1991.Footnote 35
Factors that are likely to have contributed to belligerents reaching agreement in these cases appear to include, first, the existence of a trusted and credible intermediary promoting the establishment of the zones, most frequently but not exclusively the ICRC; and second, the limited scope of the zones in terms of size, in terms of categories of people accommodated (most frequently it was just the wounded and sick), in terms of numbers of people and in terms of duration. This increased the likelihood of the zones being considered acceptable, and of their being respected. It made it simpler for those operating the zones to comply with the requirement that they not pose a military threat, and, in view of their limited size, they were unlikely to impede the conduct of military operations.
A more recent example that highlights the risks of establishing so-called protected zones without the agreement of the opponent are the three successive “no-fire zones” established unilaterally by the government of Sri Lanka in early 2009, with the claimed intent of providing safety to civilians who remained in the ever-diminishing areas under the control of the Liberation Tigers of Tamil Eelam (LTTE). The LTTE did not recognize these zones and took no steps to prevent armed elements from entering them. In the final weeks of fighting, the zones were under constant attack by government forces, leading to massive civilian casualties and destruction of the hospitals located in the zones.Footnote 36
Attention has focused on the creation of zones to provide protection to vulnerable people. However, damage or destruction of civilian objects during the conduct of hostilities can also have a severe impact on civilians’ well-being. This is particularly the case for infrastructure providing essential services, such as health-care facilities, electricity generation and distribution networks, and water treatment and distribution facilities. With these considerations in mind, since 2017 the ICRC has been facilitating negotiations between the Ukrainian authorities, the Organization for Security and Co-operation in Europe, and representatives of non-government-controlled areas in Donetsk and Lugansk. The objective is to create an agreement to establish safety zones along the contact line in eastern Ukraine around two water installations: a pumping station and a filtration station. If concluded and respected, the agreement would ensure that infrastructure providing clean water to more than 1.8 million people on both sides of the contact line will continue to operate.Footnote 37
This example highlights the value of initiatives to spare critical infrastructure from the effects of hostilities. This is a type of protected zone that has been largely overlooked but warrants closer attention. The conditions for the conclusion and operation of such humanitarian arrangements are similar to those for protected zones for vulnerable persons: agreement between belligerents, clear identification and demarcation of the areas in question, and supervision of compliance with the terms of the agreement.
The general rules of IHL regulating conduct of hostilities
As repeatedly stated, the protection afforded to protected zones is additional to that to which civilians, persons hors de combat (such as wounded combatants) and civilian objects are entitled under the general rules of IHL regulating the conduct of hostilities. These protections are essentially the same in international and non-international armed conflicts, and include:
• the obligation to respect and protect civilians and the wounded and sick;Footnote 38
• the prohibition of direct attacks against civilians and civilian objects;Footnote 39
• the prohibition of direct attacks against medical facilities;Footnote 40
• the prohibition against conducting indiscriminate attacks, including “disproportionate attacks” – that is, attacks expected to cause incidental loss of civilian life, injury to civilians or damage to civilian objects that would be excessive in relation to the concrete and direct military advantage anticipated;Footnote 41
• the prohibition against using civilians or other protected persons to render areas immune from attack or to shield military operations;Footnote 42
• the obligation, to the extent feasible, to avoid locating military objectives within or near densely populated areas;Footnote 43 and
• the obligation to take constant care in the conduct of military operations to spare the civilian population and civilian objects and to take precautions in attack and defence.Footnote 44
It is important to recall these general protections for a number of reasons: first, even if protected zones are established, and even if they operate as envisaged and provide protection, people who do not seek refuge there remain protected in accordance with these rules. The establishment of the protected zones in no way reduces their protections or belligerents’ obligations towards them.
Second, these general rules continue to apply during the operation of the protected zones – for all parties. For example, should any individuals in the protected zones engage in hostilities, they must comply with the rules on precautions in defence, including the prohibitions against resorting to human shields and locating military objectives within or near densely populated areas, such as the zones themselves. Parties responding to such attacks are similarly bound, including, most notably, by the prohibition against indiscriminate attacks.
Finally, and as already noted, if, for whatever reason, protected zones lose their protected character, the people sheltering there nonetheless remain protected by these general rules.
“Safe areas” established by means other than agreement between belligerents
The assumption underlying the various protected zones foreseen by IHL is that civilians and the wounded and sick are caught up in hostilities and that, provided the necessary measures are taken to ensure the zones are in fact exclusively humanitarian, belligerents will be willing to agree to their establishment. However, this assumption does not hold in situations where the civilian population is being deliberately targeted. The objective of the party attacking civilians is to harm or forcibly displace civilians; that party is therefore unlikely to be willing to set up areas to protect them. This is the most extreme situation, but not the only one in which it may be impossible for belligerents to agree to the establishment of protected zones. In other situations, doing so might simply not have been considered, or insufficient efforts may have been made to facilitate agreement to the zones by belligerents.
In the past twenty-five years, in response to a number of conflicts in which civilians were systematically targeted, protected zones have been established without the agreement of belligerents, including, in some cases, without the consent of the State party to the conflict in whose territory the areas were set up. These include the “safe havens” in Northern Iraq in 1991, the “safe areas” in Bosnia in 1992, and the “safe humanitarian zone” in southwestern Rwanda in 1994.Footnote 45
When so-called “safe areas” are established in such circumstances, in addition to the rules of IHL regulating the conduct of hostilities outlined above – which remain pertinent – another body of law comes into play: jus ad bellum, the rules regulating resort to armed force. Moreover, if the safe areas are established and administered by a multinational force, in addition to requiring the necessary authorization from the United Nations (UN) Security Council to resort to armed force in the first place, the mandate granted to such forces must also be considered, to determine whether it authorizes them to establish the areas and also, crucially, to use force to defend them.
The next section briefly presents the cases of northern Iraq, Rwanda, and Bosnia and Herzegovina. The objective is not to analyze why some areas were more successful than others in providing protection – something that has already been the object of extensive research.Footnote 46 Instead, the article outlines key elements of the regulatory framework in the three cases: the legal basis for the presence of the foreign forces implementing the safe areas, and their mandate.
Jus ad bellum considerations
The safe areas in northern Iraq, Bosnia and Herzegovina, and Rwanda were established in the territory of a State party to the conflict, by the armed forces of third States. This raises questions of jus ad bellum, the rules of international law regulating resort to the use of force.
Article 2(4) of the UN Charter prohibits “the threat or use of force against the territorial integrity or political independence of any state”. There are only two possible exceptions to this prohibition: individual or collective self-defence,Footnote 47 and collective action authorized by the Security Council acting under Chapter VII of the Charter.Footnote 48 The presence of the armed forces of a State on the territory of another State – for whatever reason – amounts to a violation of the prohibition on the use of force, unless it falls within these exceptions, or the territorial State has consented to such presence. This includes armed forces establishing and/or operating a safe area. The three above-mentioned instances in which such areas were established must be analyzed against this framework. States have never justified such areas as a form of individual or collective self-defence; instead, for the most part, they have been cases of the use of force authorized by the Security Council.Footnote 49
In the first case, Operation Provide Comfort in northern Iraq in 1991, Iraq had not consented to the creation of the “safe havens” (at least initially), but Security Council authorization was not apparent. In response to Iraq's repression of the civilian population in the Kurdish-populated areas of the country, in April 1991 the Security Council adopted Resolution 688, in which it insisted “that Iraq allow immediate access by international humanitarian organisations to all those in need of assistance in all parts of Iraq” and appealed “to all Member States and to all humanitarian organizations to contribute to these humanitarian relief efforts”.Footnote 50 Although the Council determined that the repression of the civilian population that led to massive population flows across international borders and to cross-border incursions threatened international peace and security, it did not expressly refer to Chapter VII.
Nonetheless, the resolution was invoked as the basis for a US-led multinational operation. Starting with airdrops, the coalition put ground forces in Iraqi territory to protect displaced persons and build camps. Using ground and air forces, it also established a “safe zone” in northern Iraq to allow civilians to return to their homes.Footnote 51 The government of Iraq and the UN eventually signed a Memorandum of Understanding on the UN's humanitarian activities in Iraq to replace the coalition forces, but the presence of coalition forces and their operations pursuant to Resolution 688 were without Iraq's consent.Footnote 52 This, coupled, with the Security Council's ambiguous language purportedly authorizing the use of force, led the government of Iraq to complain of a violation of its sovereignty and territorial integrity,Footnote 53 a view supported by some commentators.Footnote 54
Neither Resolution 688 nor later ones addressed the details of the “safe havens” – in terms of mandates to establish them or of authorization to use force to defend those seeking shelter there.Footnote 55 While the lawfulness of the presence of the coalition forces is questionable, as is the actual mandate to establish the zones, the coalition forces adopted a robust approach to protecting the zones. In the immediate aftermath of the expulsion of Iraq from Kuwait, the threat of military action in defence of the zones had the requisite deterrent effect.Footnote 56
In contrast to northern Iraq, the lawfulness under jus ad bellum of the presence of the UN Protection Force (UNPROFOR), the multinational force eventually tasked with protecting the “safe areas” in Bosnia, was unquestionable. The Security Council established UNPROFOR through Resolution 743, with the consent of the government of Yugoslavia.Footnote 57 Resolution 758 expanded its mandate to Bosnia and Herzegovina.Footnote 58
The basis for the establishment of the safe areas in Bosnia was equally clear: in Resolution 819, acting under Chapter VII, the Security Council demanded “that all parties and others concerned treat Srebrenica and its surroundings as a safe area which should be free from any armed attack or any other hostile act”.Footnote 59 Resolution 824, also adopted under Chapter VII, declared that “Sarajevo, and other such threatened areas, in particular the towns of Tuzla, Zepa, Gorazde, [and] Bihac”, also constituted “safe areas” and should be free from armed attacks.Footnote 60
The Security Council's approach to safe areas in Bosnia suffered from a different shortcoming. Resolution 819 requested the Secretary-General, “with a view to monitoring the humanitarian situation in the area”, to increase UNPROFOR's presence in Srebrenica and its surrounding areas, and demanded that all concerned parties cooperate with UNPROFOR for this purpose.Footnote 61 Resolution 824 contained a similar request in relation to the other locations designated as safe areas.Footnote 62 The Council did not, however, grant UNPROFOR the mandate to use force to defend the safe areas until June 1993, and even then did not allocate additional troops for this task.Footnote 63 This, coupled with the fact that the areas were not demilitarizedFootnote 64 (which meant that military operations were carried out from the areas, and led to responses from opposing forces), made it impossible for UNPROFOR to protect the areas from Bosnian Serb attacks when the hostilities intensified, leading to the massacres of the Bosnian men and boys who had been seeking refuge in the areas.
In this instance, the presence of the multinational forces was unquestionably lawful from a jus ad bellum point of view, and they had a clear mandate to establish and administer the safe zones. However, the combination of the failure to demilitarize these zones and the lack of a mandate to use force to defend them led to the devastating outcome.Footnote 65
In Rwanda, the Security Council first established a UN force, the UN Assistance Mission for Rwanda (UNAMIR), with the consent of the government of Rwanda and of the Rwandan Patriotic Front in Resolution 872.Footnote 66 Although its presence was lawful, UNAMIR did not have a mandate to establish safe areas. Instead, the so-called “protected sites” emerged spontaneously, either as people fled to areas where UNAMIR personnel were known to be stationed, or as UNAMIR troops were dispatched to sites where civilians had congregated. Although UNAMIR contributed to saving lives, it simply never had the capacity to respond to the scale of the crisis.Footnote 67 In view of the magnitude of the crisis and the delays in bringing UNAMIR up to strength, nine months later, acting under Chapter VII, the Security Council adopted Resolution 929. This authorized the deployment of a “temporary force under national command and control aimed at contributing, in an impartial way, to the security and protection of displaced persons, refugees and civilians at risk in Rwanda”.Footnote 68 From the outset, this force was authorized to use all necessary means – including the use of force – to contribute to the security and protection of displaced persons, refugees and civilians at risk in Rwanda, including through the establishment and maintenance, where feasible, of secure “humanitarian areas”.Footnote 69 Led by France, and provided with significant troop numbers and equipment (including helicopters and fighter aircraft), Operation Turquiose established a “safe humanitarian zone” covering one fifth of Rwanda's territory.Footnote 70 Regrettably, this only occurred once the genocidal violence was subsiding, and the force was deployed for less than two months.
The presence of UNAMIR was lawful from a jus ad bellum point of view. It assumed the mandate to provide protection but, in view of its size, was overwhelmed by the extent of the violence. As far as Operation Turquoise was concerned, its presence also did not raise jus ad bellum concerns, and from the outset the force had the requisite mandates: to establish safe areas – which ended up being far larger than those envisaged by IHL – and to use force to ensure their protection. In view of the size of the “safe humanitarian zones”, one can but wonder whether the force would have actually had the ability to ensure protection to those within the zones had they been established at an earlier stage in the violence.
Mandates, “concepts of operations” and directives on the use of force
While, absent the consent of the territorial State, Security Council authorization is necessary for the presence of a multinational force not to violate that State's sovereignty and territorial integrity, as well as the prohibition on the use of force, authorization is only the first step. As highlighted by the UNPROFOR experience, it is not sufficient per se to ensure the safety of any safe area that is established. The force will also require a mandate to establish and administer the areaFootnote 71 and, importantly, a sufficiently robust mandate to deter and respond to attacks against the population there.
In addition, the various internal documents elaborated by a multinational force to implement the mandate, including the “concepts of operations” and directives on the use of force, must address the range of issues likely to be raised by the existence of a safe area in a context of ongoing hostilities. These should include, for example, measures to deter and put an end to violence by hostile forces against those seeking shelter in the areas and to ensure that relief consignments can reach the areas. They will also need to include measures to ensure safety and security within the safe areas, starting with screening and disarming those entering the areas, ensuring no military activities take place within the areas, and maintaining law and order within the areas. Some of these activities require military personnel and others police staff; as highlighted in the next section, they are likely to involve close liaison with humanitarian actors, so will also need staff with experience in civil–military coordination. Missions must be appropriately staffed.
A new model of safe areas: “Protection of civilians sites” in South Sudan
The safe areas discussed thus far, whether established by agreement between belligerents or by the Security Council, were all planned in advance. A different form of safe area came into being in South Sudan following the outbreak of fighting in late December 2013: spontaneous “protection of civilians sites” (PoC sites) formed when civilians fleeing violence sought refuge within and in close proximity to UNMISS bases.
Until now, attention has focused principally on the modalities for the establishment of safe areas, but the PoC sites in South Sudan raise a number of challenges relating to their operation in practice. The PoC sites are frequently described as “unprecedented” or as presenting “unique challenges”.Footnote 72 While it may be correct that there have never been so many people seeking refuge within peacekeeping bases for such prolonged periods of time,Footnote 73 this is by no means the first occasion on which this has occurred,Footnote 74 nor is it likely to be the last.Footnote 75 The sites raise innumerable operational challenges, and it is well beyond the scope of the present article to attempt to present and analyze them all;Footnote 76 nor is this article intended to be a criticism of the sites, which have played an important role in providing protection in South Sudan. Instead, it will highlight a small number of issues that have arisen in the operation of the PoC sites, as they are likely to occur when safe areas are established – by whatever means.
In terms of regulatory framework, the UN Security Council, acting under Chapter VII of the UN Charter, established UNMISS in July 2011. From the outset, its mandate included a protection of civilians dimension, and the authorization to use force to implement it. Originally, this focused on providing support to the government of Sudan to develop its capacity in this area.Footnote 77 Violence broke out at the end of 2013, but it was only in May 2014 that the mandate was changed from one of support to the government to one that required UNMISS to take measures to respond to a number of threats to civilians, including deterring violence against civilians within and outside of PoC sites and maintaining public safety and security within PoC sites.Footnote 78 Thus, while the Security Council never mandated UNMISS to administer the PoC sites, it eventually expressly tasked it with protecting the people seeking shelter there and with maintaining public safety and security within the sites.
The mere presence of thousands of people for prolonged periods of time on or in close proximity to UNMISS premises meant that the people seeking shelter there looked to the Mission not only for protection, but also for assistance and other basic services. UNMISS was reluctant to conduct activities that fell beyond its mandate and for which it lacked capacity, and this required it to cooperate with a range of humanitarian actors operating in South Sudan.Footnote 79 This cooperation was frequently fraught, starting with the reluctance of some humanitarians to be associated with armed actors by providing assistance on a military compound.Footnote 80 Difficulties also arose in terms of allocation of responsibilities for particular tasks; where ultimate decision-making authority lay; minimum standards that the sites should meet in terms of adequate food, water, sanitation and medical care; who should meet the costs of improvements to the sites, such as the construction of perimeter fences; and decisions on whether to close the sites.Footnote 81 Eventually, guidelines were adopted laying down respective roles and responsibilities for operations in the PoC sites.Footnote 82
Even implementation of the activities with which UNMISS had been expressly tasked by the Security Council, most notably ensuring the safety of the PoC sites from external threats and maintaining safety and security within the sites, gave rise to operational challenges. The presence in the sites of former combatants and people who had not fully disarmed, the availability of weapons, and inter-communal violence related to the conflict gave rise to significant security concerns.Footnote 83 In addition, as is often the case when large numbers of people are accommodated in close quarters, criminality within the sites was a problem, with frequent incidents of violence – including sexual or gender-based violence, and violence related to gang, community or family disputes – as well as theft and drug smuggling.Footnote 84
Preserving the civilian character of the PoC sites is key to preventing attacks from hostile forces. One of the principal challenges to this was the presence of former combatants in the sites. UNMISS eventually adopted guidelines stipulating how armed combatants seeking access to the sites should be treated.Footnote 85 However, it is extremely difficult to distinguish a combatant from an armed civilian in South Sudan, and the risk remains that combatants could take advantage of the PoC sites to seek temporary safety.Footnote 86 This undermines the safety of the sites and, as the sites now predominantly host people from the ethnic group that opposes the government, may give the impression that those administering the sites are not neutral but are indirectly providing support to that party to the conflict.Footnote 87
The PoC sites are frequently compared to sites for internally displaced persons (IDPs), but they actually differ in important ways.Footnote 88 From a regulatory point of view, the fact that the sites are located within UNMISS bases means that, pursuant to the Status of Forces Agreement concluded between the UN and South Sudan, they are on territory that is “inviolable” and “under the exclusive control and authority of the UN”.Footnote 89 While ordinarily in IDP camps it is the host State that carries out numerous administrative functions (including, notably, preventing and responding to criminality within the camps), this is UNMISS's responsibility with regard to the PoC sites, as a result of the status of the bases.Footnote 90 However, UNMISS's capacity to maintain security within the sites has been hampered by the absence of a law enforcement or judicial authority dimension to its mandate (a so-called “executive mandate”), allowing it to investigate crimes, conduct pre-trial detention, and prosecute and detain people for criminal activity.Footnote 91 This, coupled with the weakness of South Sudan's criminal justice institutions and the frequent impossibility of transferring suspects to the local authorities as doing so might expose them to the risk of human rights violations, has required the Mission to develop alternative approaches to dealing with criminality.Footnote 92 These have included community watch groups, intended to monitor the situation within the sites and alert UNMISS police of disturbances; and an informal mediation and dispute resolution mechanism, which deals with breaches of security that do not pose a substantial risk to public order or safety within the sites.Footnote 93 Traditional justice mechanisms have also continued to operate within the sites.Footnote 94 When breaches of security are more severe and involve persons who pose significant threats to public security, these are transferred to UNMISS, which has held such people, sometimes for prolonged periods, and expelled some of them from the PoC sites.Footnote 95 As discussed below, in the absence of an executive mandate and a legal framework regulating deprivation of liberty by UNMISS, this raises questions of compliance with international human rights law.
The PoC sites have saved tens of thousands of lives, but they have also highlighted numerous operational challenges in terms of allocation of responsibilities and coordination between missions, humanitarian actors and host States. These challenges are relevant to the operation of most safe areas.
Protection in safe areas: Refugee law and human rights law considerations
Safe areas also raise a number of refugee and human rights law-related questions. Most starkly, there are serious concerns that in some contexts the establishment of the areas was not driven by a desire to create zones of shelter for vulnerable people; rather, the primary motive was to prevent or end refugee flows across borders or to promote or effect returns of refugees at a time when the conditions on the ground did not warrant this.Footnote 96
In law, the position is simple: the existence of safe areas must not be used to limit people's entitlement under refugee law to seek asylum,Footnote 97 nor to promote returns of refugees before they are safe. In practice, safe areas have been established or suggested for this very purpose on a number of occasions. This has raised complex questions for humanitarian actors, most notably the Office of the UN High Commissioner for Refugees (UNHCR). Should UNHCR push for respect for the principles of refugee law, including access to asylum and non-refoulement, and refuse to carry out activities for people in safe areas, so as not to be seen as supporting arrangements that undermine the essence of refugee protection – even if this means depriving people in need of essential services?Footnote 98
Safe areas, however they are established, also raise questions of human rights law. The parties that operate such areas have assumed a degree of control over their residents and, with that, human rights obligations towards them.Footnote 99 These situations raise questions about the scope of extraterritorial application of human rights obligations and, if the areas are under the control of multinational forces, the manner in which human rights law applies to such forces. Both topics have been the subject of legal proceedings and considerable academic debate in recent years. A detailed discussion is beyond the scope of this article, but for present purposes it suffices to note, first, that recent human rights jurisprudence has imputed extraterritorial obligations on States when they have assumed “effective control” over areas of foreign territory,Footnote 100 or when their agents exercise physical control and authority over individuals.Footnote 101 This is understood in a broad sense, and can include overseas detention operations, but also situations in which State agents exercise control over people passing through checkpoints, and even through targeting or the use of force.Footnote 102 Second, multinational forces established by the UN Security Council must comply with international human rights law. This is because the UN must itself respect human rights, and because troops participating in multinational forces remain bound by the sending States’ human rights obligations.Footnote 103
Two related cases of 2013 before the Supreme Court of the Netherlands confirmed both the extraterritorial application of the Netherlands’ obligations under the European Convention on Human Rights (ECHR)Footnote 104 and the International Covenant on Civil and Political Rights (ICCPR),Footnote 105 and the responsibilities of States that contribute troops to multinational forces. The cases related to the involvement of the Dutch battalion in UNPROFOR in the “safe area” of Srebrenica. Inter alia, the Court found that the battalion had exercised effective control over the people in the enclave and that this had brought into play its human rights obligations. By not allowing the plaintiffs’ relatives to remain in the compound where it was based, which led to their murder by Bosnian-Serb forces, the battalion and therefore the Netherlands had violated the right to life and the prohibition on torture and cruel, inhuman or degrading treatment or punishment in the ECHR and the ICCPR.Footnote 106
According to current jurisprudence, extraterritorial human rights obligations arise only in relation to those rights that are actually subject to a State's control in a particular situation.Footnote 107 In situations of occupation, States must ensure the full spectrum of human rights. In other situations, particularly when the responsibility arises as a result of the exercise of effective control over a person, it will be more limited in scope, and may include the right to life, the prohibition on arbitrary deprivation or life or liberty, and the prohibition on torture or cruel, inhuman or degrading treatment or punishment.Footnote 108 Accordingly, the precise nature and extent of extraterritorial human rights obligations in relation to safe areas will depend on a number of factors, including the degree of control exercised over the residents of the safe areas by the multinational forces; the extent to which the territorial State has been excluded and thus prevented from discharging its human rights obligations towards the residents; and the specific human rights in question.Footnote 109 By way of extreme example, the human rights obligations of a State operating a no-fly zone to protect civilians will be significantly less onerous than those of a State that is operating a confined safe area.
Where multinational forces have been mandated with particular tasks in the safe areas (as is the case, for example, for UNMISS in the PoC sites), in discharging them they must comply with their human rights obligations that are relevant to the tasks in question. For example, in maintaining public safety and security within the PoC sites, UNMISS must comply with human rights standards relating to the use of force in law enforcement. If it deprives people of their liberty, it must ensure they are treated in accordance with human rights standards and are afforded due process.Footnote 110
While UNMISS may not have publicly stated that it has human rights obligations towards the residents of the PoC sites, its approach to particular issues indicates that it considers this to be the case. This includes its reticence to hold people suspected of serious breaches of security, because, as discussed above, in the absence of an “executive mandate”, holding people in such circumstances could amount to an arbitrary deprivation of liberty.Footnote 111 Similar concerns also underlie the draft memorandum of understanding submitted by UNMISS to the Ministry of Justice of South Sudan in relation to transfers of suspects to national authorities. This was an effort to give effect to the Mission's obligation under human rights law not to transfer people if a real risk exists that they may be subjected to torture or ill-treatment, a trial that does not meet minimum standards, or the death penalty.Footnote 112
Conclusion
The track record of the “safe areas” that have been established since the Second World War has been mixed at best. Their success depends on belligerents’ willingness to respect them, something that can be achieved either by establishing the areas by agreement or, absent such agreement, by the taking of robust measures to defend the areas.Footnote 113
Not surprisingly, the most effective safe areas have been those established by agreement between belligerents, as envisaged by IHL. In particular, smaller, demilitarized zones accommodating limited numbers of particularly vulnerable people for short periods of time appear to be the most likely to succeed.
In view of this, should IHL be revised so as to require belligerents to agree to establish and respect protected zones? This is improbable. First, and more generally, States are extremely unlikely to open the Geneva Conventions and the Additional Protocols for revision just to amend the provisions on protected zones. Any change to these rules would have to be part of a broader process of revision, for which there is no appetite at the moment, not least because of well-founded concerns that doing so in the current political climate would reduce existing protections rather than enhance them. Second, States had the opportunity to adopt provisions requiring belligerents to establish protected zones twice: during the negotiations of the 1949 Geneva Conventions and of the 1977 Additional Protocols. On both occasions they chose not to do so. It is unrealistic to think that they would take a different approach now.
That said, experience has shown that if safe areas are to function effectively, it is essential for belligerents to agree to the details of their establishment and operation. Not only does this set clear limits for the areas, but it also addresses what are likely to be the key concerns: demilitarization and measures of supervision. Moreover, the process of negotiating an agreement may help to build a climate of trust and cooperation between belligerents – something that will be essential to the areas’ effective functioning in practice, and which may also have a beneficial effect on compliance with IHL more generally.
The establishment of safe areas by agreement between belligerents has a further advantage: it might alleviate some of humanitarian actors’ reservations about operating in the areas, at least in terms of involvement with arrangements imposed by one side to a conflict. Reservations would – rightly – remain if the objective of the zones was stemming refugee flows, as this would undermine the residents’ right to seek asylum.
While frequently the requirement of consent by belligerents is perceived as impeding protective measures for civilians, in the case of safe areas it may actually have the opposite effect, by responsibilizing parties, building trust and ensuring that there is clarity about arrangements to enhance their effectiveness.
In view of all this, rather than calling for a reform in the law, far greater investment should be made in encouraging belligerents to reach agreement on the establishment and recognition of protected areas. Quiet diplomacy is more likely to succeed than calls in the political limelight of the Security Council, and as always, humanitarian negotiations should be kept totally separate from discussions of a political nature. Calls for the establishment of safe areas without belligerents’ agreement should be made judiciously, so as not to undermine negotiations that may be ongoing. Those suggesting such arrangements should also bear in mind the experiences of the past, in terms of mandates, will and capacity to enforce safe areas, and also the range of operational challenges that will need to be addressed when operating the areas, as highlighted by the experience of the UNMISS PoC sites.
