In 1991, Croatia and the Socialist Federal Republic of Yugoslavia concluded an agreement during their armed conflict to declare a protected zone around Osijek hospital under ICRC supervision. Access to the zone was restricted to sick and wounded civilians and military personnel; family members visiting patients recovering in the hospital; people over 65 years of age, children under 15, expectant mothers and mothers of children under seven; and the hospital’s medical and administrative personnel.
6.1 What are they?
Another arrangement whose establishment is frequently called for to provide some security to civilians caught up in active hostilities is that of ‘safe zones’, also referred to in various other ways, including ‘safe havens’, ‘safe areas’, ‘buffer zones’, ‘protected areas’ and ‘safe humanitarian zones’. These are areas that host people who are not or are no longer taking a direct part in hostilities: civilians, wounded and sick fighters, and those taking care of them.
These zones should be ‘demilitarized’: they should only accommodate people who are not taking a direct part in hostilities and hostile acts should not be conducted from them. Consequently, there should be no reason for attacks to be directed against the zones, thus promoting the security of people sheltering in them.
This paper uses the term ‘protected zones’. This is the generic expression used in IHL; it reflects the reality that even if established and accorded special protection in law the areas might not in fact provide security to those sheltering there.
Protected zones are a formal arrangement, ideally established by agreement between belligerents, where a party assumes control for the people sheltering in the zones and certain responsibilities towards them. At times civilians move to the proximity of objects afforded special protection under IHL, such as hospitals, or that they perceive as safe, such as UN compounds. This does not render the locations protected zones. However, if IHL is complied with, the protection afforded to hospitals and UN facilities would mean that areas in their proximity are less affected by military operations and thus safer for civilians.
Protected zones are similar to other humanitarian arrangements considered in this paper. First, they are a practical way of giving effect to the protections that IHL already accords to certain categories of people. Second, IHL envisages the possibility for belligerents to establish a variety of protected zones but does not require them to do so. Third, although it is possible for a belligerent to set up a protected zone unilaterally, there needs to be agreement between the parties if the zone is to be truly protective.
6.2 What does the law say?
6.2.1 The protected zones foreseen by IHL
IHL envisages the possibility for belligerents to establish a number of different types of protected zones. The principal difference between them is the category of persons they accommodate. This is just a reflection of the specific focus of the Geneva Conventions, rather than representing significant differences in the nature of the zones or the modalities for their establishment.
In practice, the various types of protected area could be combined, accommodating wounded and sick civilians and members of the armed forces, as well as other vulnerable people. Rather than the specific categories of people accommodated, it is the size of the zones and duration of the arrangements that pose the greatest challenges to the establishment and safe operations of the zones.
Hospital zones
The First Geneva Convention foresees the possibility of establishing hospital zones for wounded and sick members of states’ armed forces; and the Fourth Geneva Convention does the same for wounded and sick civilians. In practice, nothing precludes an arrangement whereby the same hospital zone accommodates wounded and sick members of the armed forces as well as civilians.
Admission to hospital zones must be granted without adverse distinction, and thus wounded and sick members of enemy forces must also be admitted.
Safety zones and localities
The Fourth Geneva Convention also foresees the possibility of establishing ‘safe zones and localities’ to protect particularly vulnerable civilians from the effects of hostilities. Although the convention specifically refers to ‘wounded, sick and aged persons, children under fifteen, expectant mothers and mothers of children under seven’, there is no reason why all civilians should not be able to seek shelter in such zones and localities.
Neutralized and demilitarized zones
The Fourth Geneva Convention refers to the possibility of establishing ‘neutralized zones’. The difference between hospital/safety zones and neutralized zones is that the former are intended to be located at a distance from combat zones, and to constitute a longer-term arrangement for certain categories of particularly vulnerable civilians. Neutralized zones, instead, are in areas of active fighting, and aim to provide temporary shelter to wounded and sick combatants and civilians and the entire civilian population.
The range of possible protected zones is further expanded by Additional Protocol I, which foresees the possibility of establishing ‘demilitarized zones’. Their objective is similar to that of the neutralized zones: placing localities and their non-combatant population outside the theatre of war. The difference is in the approach to doing this.
While the Fourth Geneva Convention does this by the establishment of zones to which civilians and the wounded and sick can relocate, the demilitarized zones envisaged by Additional Protocol I ‘fence off’ certain 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 – and, if it recognizes the zone, its opponent must refrain from extending ‘military operations’ – also a broad notion – to the zone. To date, no such zones have been established.
Non-international armed conflicts
In relation to non-international armed conflicts, neither Common Article 3 to the Geneva Conventions nor Additional Protocol II refers to any type of protected zone. This is not significant, as nothing precludes parties from entering into agreements with a similar effect. In fact, these are among the types of special agreements that Common Article 3(3) suggests that parties could conclude to bring into effect other provisions of the Conventions.
The points made below about modalities for reaching agreement and the practical operation of protected zones are equally relevant whether they are established in international or non-international armed conflict.
A flexible arrangement
As the law does not require particular zones to be established, the zones that have been set up were based on what was needed and achievable in the circumstances, rather than on the models foreseen in the Geneva Conventions and Additional Protocol I. The zones accommodated the most vulnerable categories of people – usually the wounded and sick, as in the cases of the zones established in Dhaka in 1975, Nicosia in 1974, Saigon in 1975, and Osijek in 1991. They were located in the areas of active combat or violence, and hosted relatively small numbers of people.
Similarly, although this type of arrangement was not expressly referred to in IHL treaties during the Falklands/Malvinas war in 1982, the UK and Argentina agreed to establish a neutral zone on the high seas where hospital ships from both sides could operate in safety. Referred to as a ‘Red Cross Box’, this was a zone, some 20 nautical miles in diameter, where four British and three Argentinian hospital ships were stationed. These vessels could receive the wounded and provide treatment without hampering military operations, and the proximity between the two sets of ships allowed for the easy exchange of patients. The arrangement allowed hundreds of British and Argentine casualties to be treated.
6.2.2 Establishment of the protected zones
Belligerents can establish protected zones but until and unless they are recognized by their opponent, such zones do not benefit from special protection. However, the absence of such recognition does not deprive those people sheltering in the zones of the protection they are entitled to under IHL as civilians or persons hors de combat.
Elements of agreement
The likelihood of recognition of a protected zone depends on a number of factors. Key among them are the steps taken to ensure that the zones are exclusively used to shield civilians and other protected people from the effects of hostilities and to reduce any risk that they will be abused.
Annexed to the First and Fourth Geneva Conventions are virtually identical model agreements, referred to as ‘Draft Agreements’, that provide an indication of the kinds of issues that should be addressed by parties setting up hospital and safety zones and localities. Key elements of these agreements are set out in section 6.3.2.
The elements of the Draft Agreements are only indicative, and parties can agree to other measures. However, they provide a valuable indication of the types of measures that could provide reassurance and lead to recognition of protected zones.
Intermediary to reach agreement and supervisory body
As for the other humanitarian arrangements addressed in this paper, reaching agreement is likely to require an intermediary to facilitate the negotiations. This intermediary might also play the supervisory role envisaged in the Draft Agreement.
The First and Fourth Geneva Conventions expressly invite the ICRC and any protecting power to ‘lend their good offices’ to facilitate the institution and recognition of hospital and safety zones. The ICRC assumed this role, facilitating the conclusion of many of the agreements to establish protected zones since the Second World War, and it also administered many of the zones.
As far as supervision of protected zones is concerned, although it was not involved in the negotiation of this arrangement, the ICRC acted as a supervisory body for the ‘Red Cross Box’ established by agreement between Argentina and the UK in the South Atlantic in 1982.
6.2.3 Interplay with other rules of IHL
As a matter of law, it is not the establishment and recognition of the protected zones that provides protection to the civilians and the wounded and sick sheltering there. IHL already requires those people to be respected and protected. Recognition of the zones, however, gives an additional layer of protection in practice: the belligerent that recognizes the zones undertakes not to direct hostilities into or in the proximity of the zones, provided they are used in accordance with the agreed terms. This is sometimes referred to as ‘special protection’ because, for the most part, IHL protects objects not areas.
Like other humanitarian arrangements discussed in this paper, the zones are a practical way of enhancing the security of protected people. Hostilities and other military operations should not be conducted into or in the proximity of the areas, thus reducing civilians’ exposure to risks. In gathering the wounded and sick in one location, zones can facilitate their treatment by addressing adverse effects of armed conflict such as shortages of supplies and the breakdown in health services.
The rules of IHL regulating the conduct of hostilities and humanitarian relief operations continue to apply. They apply unmodified to people who have not sought refuge in the protected zones, or who choose to leave them. The establishment of the zones in no way affects belligerents’ obligations towards those people.
The general rules of IHL also continue to apply to those within the protected zones. If, for example, in violation of the agreement establishing the zones, people who are sheltering there engage in hostilities, they must comply with the rules regulating the conduct of hostilities. In addition to the rules regulating attacks, other rules are also relevant: the people conducting attacks must not use human shields; and must, to the maximum extent feasible, take precautions against the effects of attacks, including avoiding locating military objectives within or near densely populated areas.
It should be recalled that conducting hostile acts from the protected zones would undermine the special protection afforded to them. The agreement establishing the zones should regulate how to address such situations. For example, it might require the party targeted by the acts to issue warnings, and the party administering the protected zones to take measures to put an end to such activities.
If, nevertheless, hostile activities from the protected zone continue, leading to attacks directed against the zone, these must comply with the general rules of IHL. Attacks must be directed exclusively against military objectives; they must not be indiscriminate; and they must comply with the rule of proportionality; and feasible precautions must be taken to spare civilians and civilian objects. The context in which such operations would be conducted – likely to be a small and restricted area hosting high numbers of protected people – will significantly limit the permissible use of force.
6.2.4 Interplay with refugee law
Protected zones raise questions of refugee law. In some situations, calls for their establishment were driven by a desire to prevent or put an end to population flows across borders, or to promote the return of refugees at a time when this was not warranted by conditions the ground.
Notable examples include Turkey’s closure of its borders to Iraqi Kurds in 1991, which led to the adoption of Security Council resolution 688 (1991) and the establishment of Operation Provide Comfort in northern Iraq. More recently, calls were made at various junctures during the conflict in Syria to establish safe zones within Syria. None was ultimately established.
These are not new problems, but they continue to present difficult policy choices for humanitarian actors, and for UNHCR in particular, when deciding whether to join calls for the establishment of protected zones, and on the nature of their involvement in any that might have been set up. The establishment of safe zones must not be used to limit the entitlement to seek asylum, or to promote returns of refugees before it is safe for them to return.
6.2.5 Human rights considerations
In addition to rules relevant to the establishment of the protected zones, consideration must also be given to the rules relevant to their functioning. Depending on the nature of the zones, a state or coalition that is administering the zones may be considered as having effective control of the people who have sought shelter in the zones. With such effective control come certain human rights obligations.
In addition to meeting basic needs in terms of food, water, shelter and medical care, it is also necessary to maintain safety and security within the zones.
6.2.6 Protected zones established by the Security Council
IHL envisages that protected zones are established by agreement between belligerents. In recent decades, there have been a small number of instances when protected zones were established by or on the basis of UN Security Council resolutions. They include the ‘safe havens’ established in northern Iraq in 1991; the ‘safe areas’ in Bosnia and Herzegovina in 1993; and the ‘safe humanitarian zones’ in Rwanda in 1994.
In addition to IHL considerations, the establishment of protected zones in this manner raises questions regarding the rules of international law on resort to force. The presence of the armed forces of a state on the territory of another state without the latter’s consent is a violation of the prohibition on the use of force, unless such a presence is authorized by the Security Council, or amounts to individual or collective self-defence (not relevant in the instances when safe zones have been established to date).
Safe havens were established in northern Iraq in 1991 when hundreds of thousands of Iraqi Kurds fled to the mountains in the face of repression from the regime of Saddam Hussein. Security Council resolution 688 (1991) condemned the repression and required Iraq to allow immediate access by humanitarian organizations to all those in need of assistance. The resolution did not establish the zones; consent by the government was not entirely clear. The resolution was nonetheless the basis for a US-led multinational operation, ‘Provide Comfort’, that carried out airdrops and put forces on the ground to protect displaced people and build camps. These safe havens were also protected by means of the no-fly zone referred to in Chapter 7.
The Security Council established ‘safe areas’ in Bosnia in 1993, but there were problems of enforcement. The Security Council demanded that parties to the conflict treat Srebrenica and the other locations referred to as safe areas, but it was only several months later that it granted the UN Protection Force (UNPROFOR) the mandate to use force to defend these areas. Even then, it did not allocate additional troops for this task. This, coupled with the fact that the areas were not demilitarized (which meant that military operations were carried out from the areas, in turn provoking responses from opposing forces), made it impossible for UNPROFOR to protect the areas from Bosnian Serb attacks when the hostilities intensified. This led to massacres of the Bosnian men and boys who sought refuge in the areas.
Rwanda had agreed to the presence of the UN Assistance Mission for Rwanda (UNAMIR), established in 1993, so Security Council authorization was not necessary for its presence. The force did not have a mandate to establish safe areas, but so-called ‘protected sites’ emerged spontaneously in April 1994 as civilians sought security in the areas where UNAMIR forces were stationed. However, UNAMIR never had the capacity in terms of troops and equipment to respond to the scale of the crisis.
Nine weeks after the start of the genocide in Rwanda, in view of the magnitude of the crisis, the Security Council adopted a resolution under Chapter VII of the UN Charter, authorizing the deployment of a force under national command to contribute to the protection of civilians. This was authorized to use force including to establish and maintain ‘secure humanitarian areas’. Led by France, Operation Turquoise established a ‘safe humanitarian zone’ that covered a fifth of Rwanda’s territory.
Tragically, this only happened as the genocidal violence was subsiding, and the force was deployed for less than two months. Had the violence continued, it is questionable whether – given the size of the ‘safe humanitarian zone’ – it would have been feasible to provide protection to the people located there.
Rwanda was not the only context where civilians sought security in the proximity of UN forces. The same occurred following the outbreak of fighting in South Sudan in 2013. ‘Protection of Civilians’ sites (POC sites) came into being as civilians spontaneously sought security near and within UN Mission in South Sudan (UNMISS) bases. Over 35,000 civilians sought refuge within three weeks of the start of the violence, and more than 200,000 people were still sheltering in the sites three years later. The POC sites raised a multitude of operational challenges that are relevant to all zones that accommodate large numbers of people for prolonged periods of time. A small number of these challenges are flagged below.
As regards its Security Council mandate, UNMISS had a mandate for the protection of civilians, and authorization to use force to implement it from when it was established in 2011, but initially this was only to develop the government’s capacity in this area. Violence broke out in December 2013, but it was only in May 2014 that the mandate was amended to allow responding to certain threats against civilians, including deterring violence against civilians within and outside POC sites, and maintaining safety and security within the sites. UNMISS was thus never tasked with establishing the POC sites, and had only a very limited mandate in terms of operating the sites. This contributed to the challenges they raised.
6.3 Operation in practice – challenges, good practice and recommendations
Although their establishment is frequently called for, protected zones have been established only on a very small number of occasions – whether by agreement or further to Security Council authorization. This is a reflection of the challenges in reaching agreement to establish the zones, as well as in administering them.
6.3.1 Unilateral zones are unlikely to afford protection
There have been instances when one side to a conflict established purportedly safe zones unilaterally. These have not proved protective.
For example, in early 2009 the government of Sri Lanka unilaterally established three successive ‘no-fire zones’, with the claimed intention of providing security 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 mass civilian casualties and destruction of the hospitals located in the zones.
6.3.2 Reaching agreement
As for other humanitarian arrangements, reaching agreement on the establishment of the zones is challenging. Many factors affect the likelihood of whether agreement can be reached. Two important elements are the existence of an intermediary to facilitate discussions, and the terms of the agreement. The agreement must provide assurance that the zones will be used for exclusively humanitarian purposes and that measures will be taken to prevent the zones from being abused.
The intermediary
As far as the intermediary is concerned, as already noted, the ICRC facilitated the conclusion of many of the agreements to establish protected zones since the adoption of the Geneva Conventions in 1949. The ICRC was also responsible for administering the majority of these zones.
At times, the ICRC took a remarkably flexible approach to securing agreement. In Nicosia in 1974, for example, civilians spontaneously sought refuge in various hotels. Of its own initiative, the ICRC marked these with a red cross flag, and only after this negotiated a written agreement with local authorities. The arrangements operated for a very brief period of time. Similarly, ahead of the expected fall of Phnom Pehn in 1975, the ICRC itself created a ‘hospital and safety zone’ in a hotel and merely informed the parties to the conflict of this, urging them to respect it. The arrangements, which operated for two days, provided temporary shelter to 2,000 people.
Experience since the 1970s has highlighted the very significant security risks that zones established and administered without the express agreement of all belligerents can pose to the people seeking shelter and to those administering them. It is perhaps doubtful that the ICRC would take a similar approach today.
Issues to be addressed in an agreement
A second aspect that will influence the likelihood of reaching agreement to recognize any zones and the possibility that, if established, they will actually provide security are the measures that the party establishing the zones undertakes to ensure their exclusively humanitarian character.
An agreement that was successful in providing protection to the people sheltering in the zone was that concluded in 1991 by Croatia and the Socialist Federal Republic of Yugoslavia (SFRY), under ICRC auspices, to create a protected zone around Osijek hospital. The agreement followed the approach suggested in IHL relatively closely. The zone was placed under ICRC supervision. Access to the zone was restricted to sick and wounded civilians and military personnel; family members visiting patients recovering in the hospital; people over 65 years of age, children under 15, expectant mothers and mothers of children under seven; and the hospital’s medical and administrative personnel.
As so few agreements have actually been concluded, it is difficult to identify good past practice. However, as mentioned above, annexed to the First and Fourth Geneva Conventions is a virtually identical Draft Agreement that highlights the types of issues that should be addressed in an agreement to establish hospital and safety zones and locations; it can provide helpful guidance to all protected zones.
The Draft Agreements contain three key sets of provisions. The first set aim to ensure that the people 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.
The second set of provisions are measures to enhance the security of the zones themselves. They include requirements that:
- The zones constitute only a small part of a state’s territory. Larger zones are likely to impair the opponent’s capacity to conduct hostilities and thus decrease the likelihood that it will recognize them;
- The zones 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; and
- Hospital zones be marked by red crosses/crescents, and other protected zones and localities by other agreed-upon signs.
The final set of provisions are supervisory measures to ensure that the zones operate as envisaged. The Draft Agreements suggest that the party recognizing a zone is entitled to demand that an independent body has access to examine the zone and confirm that it complies with the requirements agreed to. 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. As already noted, this would not affect the protections to which people accommodated in the areas are entitled under the general rules of IHL on the conduct of hostilities.
The elements of the Draft Agreement are only indicative, and parties can agree to other measures. However, they provide a valuable indication of the types of measures that could provide reassurance and lead to recognition of protected zones.
In those instances when agreement to recognize a zone has been reached in practice, and where the zones were actually respected, they were limited in scope in terms of size, categories of people accommodated (most frequently just the wounded and sick), numbers of people accommodated, and duration. This made it simpler for the party administering the zones to ensure that they did not pose a military threat or impede the conduct of military operations. It also made meeting the needs of those accommodated in the zones and addressing any security concerns within the zones less onerous.
6.3.3 Administering the protected zones – selected challenges
Responsibility for ensuring that a zone complies with the agreed terms of the agreement lies with the party administering it. Doing so is onerous, and requires considerable capacities in terms of personnel and resources. Identifying the range of operational issues that must be addressed is beyond the scope of this paper, which highlights only two overarching issues: security; and meeting the needs of the people in the zones.
Security
Security has two principal dimensions: ensuring the ‘demilitarization’ of the zones; and maintaining security within the zones. If the zones are not established by agreement, as in the case of those established by the UN Security Council, military force may also be necessary to create the spaces and to protect them
from attacks.
The ‘demilitarization’ of protected zones is the basis for their acceptance. Preserving their civilian character is key to preventing attacks from hostile forces and thus the safety of the zones. Doing so can be extremely challenging, especially in zones that accommodate large numbers of people such as the POC sites in South Sudan, which more than 200,000 people. It is less of an issue in the smaller zones, such as the hospital zones administered by the ICRC in the former Yugoslavia.
Maintaining the civilian character of zones requires the adoption of clear guidance on who is allowed into the zones, criteria for screening, security personnel to conduct such screening, and also the capacity to constantly monitor the activities conducted in the zones.
The party administering the zones must also ensure safety and security within the zones. Again, due to the large number of people accommodated and the duration of the arrangements, the POC sites in South Sudan faced a number of practical challenges. The presence in the sites of former combatants and people who had not fully disarmed, the availability of weapons, and intercommunal violence related to the conflict all posed significant security concerns. In addition, 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.
UNMISS’s capacity to maintain security within the sites was hampered by the absence of a law enforcement or judicial authority dimension to its mandate that would have allowed it to investigate crimes, conduct pre-trial detention, and prosecute and detain people for criminal activity.
Meeting basic needs
The party that is administering a protected zone is also responsible for providing essential goods and services for the people who have sought shelter. This includes food, water, shelter, medical care. If the zones are established for longer periods of time, education can also become a concern.
Again, the POC sites in South Sudan highlight some of the issues that need to be addressed. These issues were exacerbated in this instance because of the spontaneous nature of the zones and their size and duration. Most significantly, UNMISS was never mandated to administer the sites, so it lacked the capacities and funds to adequately respond to the tens of thousands of people who looked to it for assistance and other basic services for several years.
This meant UNMISS had to reach out to a range of humanitarian actors operating in South Sudan. However, cooperation was often difficult, not least because some humanitarians did not want to be associated with armed actors by providing assistance on a military compound. There was lack of clarity in terms of allocation of responsibilities for particular tasks; the minimum standards that the sites should meet in terms of adequate food, water, sanitation and medical care; and who was responsible for providing the necessary goods and services. Eventually, guidelines were adopted that set out respective roles and responsibilities for operations in the POC sites.
6.3.4 Establishment by the Security Council
Although a Security Council resolution adopted under Chapter VII of the UN Charter is binding on all states, the experience in Bosnia shows that, without the consent of all relevant belligerents, the mere establishment of protected zones by the Security Council risks giving rise to false expectations of security. It also emphasizes the undesirability of promising security when there is insufficient enforcement capacity.
The instances when the Security Council established protected zones highlight the importance of granting clear and express mandates to use force both to establish the zones, and to protect them if this is necessary. Mandates given to any multilateral forces administering the areas must address the range of responsibilities that will be entailed, such as maintaining security and order within the zones.
All aspects of the mandate must be reflected in the various internal documents elaborated by a multinational force, including the ‘concepts of operations’ and directives on the use of force. They 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 safe areas, and to ensure that relief consignments can reach those 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.
Missions must be appropriately staffed. Some of the tasks require military personnel and other police staff; they are likely to involve close liaison with humanitarian actors, so will also need staff with experience in civil–military coordination.
6.4 In conclusion
The recommendations above help to identify the range of issues to be considered in response to calls to establish protected zones. In particular, while it is possible for a belligerent to set up a protected zone unilaterally, there needs to be an effective agreement between the parties if the zone is to be truly protective.
Although there are frequent calls for protected zones, they have been established only on a very small number of occasions. Even then, they have not always provided the hoped-for protection. As for most of the humanitarian arrangements discussed in this paper, it is important to make clear – to those who are calling for the zones and to civilians who wish to benefit from them – what limitations of such zones are.
