In recent weeks, Israel has reportedly “scaled back” its operations in some areas in Gaza, while it prepares to ramp up in others. At present, the humanitarian situation among displaced Palestinians in South Gaza remains extremely dire, with the UN warning of impending famine. The humanitarian situation in the area was central to the International Court of Justice’s recent ruling on provisional measures in South Africa’s case against Israel.

The majority of those sheltering in the Rafah area in South Gaza have fled from North Gaza, when in the weeks following the horrendous October 7 attacks by Hamas, Israel called on all residents of the area to evacuate south of Wadi Gaza. With the worsening humanitarian situation, the United States, as well as other actors, stressed that residents should be allowed to return to North Gaza as soon as possible. While the area is itself devastated after months of war, this might, at the very least, alleviate the overcrowding in the camps in the South. Should Israel expand its operation into Rafah, the need for people to move back north, it seems, would be even more accentuated.

In its initial calls to leave North Gaza, the Israeli military declared that the evacuation was temporary, and that its sole purpose was to protect civilians from the intensive bombardments in the area. Nonetheless, the policy was criticized by some international organizations as potential acts of forcible displacement, and the United Nations General Assembly has called upon Israel to rescind the policy (operative paragraph 5). Furthermore, some worry that the Netanyahu government’s “true” intention behind the evacuation– following statements by far right ministers that Palestinians should be “encouraged” to leave Gaza, and that Israel should rebuild it settlements there – is to permanently ban evacuees from returning, or at least to elongate the evacuation for political reasons, such as to pressure Hamas to release Israeli hostages. For his part, the IDF Chief of Staff reemphasized, on January 13, that on his view the evacuation is temporary, and that “[w]hen we know there is no danger to the population [in North Gaza], we will be able to consider bringing them back.”

To the extent that the initial evacuation from North Gaza was unlawful, on whatever grounds, then it obviously cannot serve as a justification for prevention of return to the area. Furthermore, any suggestion that people should be “encouraged” to leave Gaza is so blatantly unlawful, that it does not require serious legal engagement at all. However, this essay is not about the consequences of unlawful evacuation. Rather, it discusses the rules that should determine return assuming that an initial evacuation could be grounded in law as a temporary measure, as claimed by the IDF. Furthermore, this essay does not deal with obligations to ensure humanitarian access to civilians in Gaza, which remain in place in any case.

Assessing Two Possible Legal Grounds

Before proceeding, it should be emphasized that there are extremely narrow grounds that could allow, under international humanitarian law (IHL), for the temporary evacuation of civilians during armed conflict. Evacuation that runs counter to these – as well as prevention of return when the relevant grounds have expired – may amount to an international crime of forcible transfer (when the evacuation is within the territory) or deportation (when the evacuation is to an area outside the territory).

The first potential legal ground for temporary evacuation is when a warning of impending attacks is given as a precautionary measure. Sometimes, such an evacuation can be a consequence of the duty to give an advance warning before an attack that may endanger the civilian population. Usually, such a warning is given before a specific attack. However, in certain circumstances, an advance warning could relate to a wider area, where attacks on a large scale are expected. If not an advance warning in the strict sense, such a call could also follow the more general duty to take constant care to spare the civilian population. To emphasize, as with all precautions, under no circumstances does a warning alone transform any protected object or person into a lawful target. Furthermore, issuing such a warning does not release the party to the conflict from its obligations to ensure humanitarian access to those remaining in the area.

More crucial for our purposes is that such a measure – if viewed as a type of advance warning or precaution – is extremely limited in its ability, if at all, to justify prolonged evacuation. Crucially, an advance warning does not create a legal obligation for civilians to evacuate – as the attacker has no legal authority over them – nor does it in itself justify preventing persons from returning. Indeed, if the purpose of the call to leave an area is to advise civilians to evacuate for their own safety, and does not create an obligation to do so, it obviously cannot provide a legal basis for preventing return. This reveals a broader conundrum: if the earlier evacuation call is phrased as a warning (which people are free to accept or reject), there needs to be another source to actively bar their return. Even setting this conceptual problem aside, a threshold issue concerns the scale of hostilities. When the intensity of the fighting in the relevant area decreases below the level that putatively justified the initial broad warning, it is no longer possible to rely on the previous level of danger to civilians to prevent them from returning to the area now.

Therefore, the IDF Chief of Staff’s statement that once the danger subsides people would be permitted to return to North Gaza is in the right direction; however, it is wrong on the law by requiring that this will be the case only when there is “no danger” in the area. Unfortunately, there is always danger to civilians during hostilities, particularly in urban areas. The question is not whether there is any danger, but rather whether the danger remains on a level that justified the initial evacuation, or perhaps whether the situation is imminently expected to return to such a level. It would be hard to argue this is the case any longer in North Gaza, especially as IDF operations have been reportedly “scaled back” in the area and considering the extent of the aerial bombing that already took place.

The second ground for evacuation stems from the law of occupation, and as opposed to the previous ground, may also provide legal authority for evacuations. Under Article 49(2) of the Fourth Geneva Convention, the occupying power may evacuate a population from a certain area if the safety of the population or imperative military reasons require it. However, it is impossible to invoke powers under the law of occupation without the corresponding duties. Crucially, this power is preconditioned on an obligation to ensure as much as possible that proper living conditions exist in the area to which civilians are evacuated, and furthermore, the evacuees must be allowed to return to their homes as soon as hostilities in the area in question cease. In any case, since there is a significant risk of abuse in the powers recognized in Article 49(2) – and in particular concerning military reasons for evacuation – it should be interpreted narrowly.

Now, many of those that criticize Israel’s initial evacuation, proceed from the assumption that Israel was an occupying power in Gaza at the time the order was issued, and has failed to ensure proper living conditions in South Gaza. The question whether Gaza was fully, “functionally,” or not occupied at the time is beyond the scope of this essay. What is important is that in recent weeks the Israeli military has claimed to possess ”operational control” in North Gaza, which supports the view that at least now it is an occupant in North Gaza (although, this control might be in flux in some areas). This remains the case even if there are some clashes in the area, as the mere existence of such clashes does not necessarily negate the existence of occupation. Article 49(2) itself concedes this possibility by recognizing that some hostilities can take place within occupied territory, as grounds for evacuation. Arguably, then, even if it is accepted that the initial evacuation was predicated on an advance warning during active hostilities, after which Israel proceeded to occupy the area, any prevention of return post-occupation becomes a de facto evacuation order also under the law of occupation and subject to its regulation.

This de facto transformation of the warning into an order under the law of occupation has two key implications. First, the occupant is under a strong affirmative obligation to restore public order and to act for the benefit of the local population. This obligation requires that the military commander act positively to facilitate the residents’ return. Second, the occupying forces must work to ensure that proper living conditions exist in the area to which people were evacuated. That, indeed, goes beyond the occupant’s general obligation to ensure the welfare of civilians in the area it controls.

In terms of the safety of the population as grounds for evacuation from occupied territory, here the analysis seems to merge with that above concerning advance warning. The occupant cannot rely on reasons for evacuation that were predicated on intense and sustained aerial bombardment of the area, to justify prevention of return after the area is occupied. Namely, it is clear that evacuation under Article 49(2) cannot stand until the end of the armed conflict as a whole, and it is likewise clear that this provision cannot allow prevention of return as long as there is fighting of any intensity in the area.

Notably, Article 49(2) does not refer only to the security of civilians as grounds for evacuation, but also to imperative military considerations. Furthermore, some claim – as the US Department of Defense Manual states in Section 5.19 – that in sieges, belligerents may prevent civilian access to certain areas. Israel might argue, in this context, that if return to North Gaza is permitted, there is fear that Hamas fighters will impersonate civilians and return to the area. However, this risk cannot justify a blanket denial of the return of all civilians. Just as the presence of some enemy fighters in a civilian area cannot alone justify the complete removal of civilians from the area to begin with, so it is that the fear that some enemy fighters may return cannot justify total prevention of return of civilians. Even if one accepts that the law of siege might recognize such restrictions – an issue not dealt with here – it is difficult to view the situation in the whole of North Gaza as one of siege, considering that Israel claims to exercise operational control within the area. A siege is about exercising external control. One cannot have it both ways.

Furthermore, should an argument be made that imperative military considerations can include the creation of a security buffer zone by evacuating civilians from an area, this would not hold on any reasonable reading of Article 49(2). Not only is the article to be read narrowly so as not to weaken the prohibition on forcible transfer or deportation, but it is also explicit that evacuees should be allowed to return as soon hostilities in the evacuated area cease. This quite clearly excludes any preventive rationale as an imperative military consideration.


Whatever rationale a state invokes for evacuations, the principle of proportionality has special significance. Although this principle usually applies to “attacks,” Israel’s own Supreme Court has ruled that proportionality also applies to other measures undertaken in armed conflict and occupation, such as the route of the West Bank Wall. Other sources also apply proportionality beyond attacks. The San Remo Manual, as part of a broader trend, applies proportionality to blockades, and the updated version of the DoD Manual also adopts proportionality in the context of measures intended to starve enemy forces (Section 5.20.2). It is also possible to argue that residual “laws of humanity and dictates of public conscience,” which are part of IHL, require that military considerations cannot justify unlimited harm to civilians, even when the law does not explicitly speak the language of proportionality. This also makes sense in terms of law’s coherence. It would be unsustainable to argue that a single attack that might risk, say, a few people, would be subject to a proportionality assessment while military measures that might affect millions would not.

Thus, it seems that even in cases where civilian safety or imperative military considerations could justify evacuation, if a humanitarian crisis occurs in an area to which civilians have been evacuated, proportionality may under certain conditions create an obligation to allow the civilians to return. Concerning North Gaza, any security advantage potentially sought by preventing disguised Hamas fighters from returning to North Gaza, must be balanced against the acute humanitarian crisis that exists in Southern Gaza due, among other things, to the overcrowding there. Likewise, any risk for civilians in North Gaza today, does not seem to outweigh the risk in south Gaza, due to the humanitarian conditions there – not to mention the risk to evacuees if significant military operations will take place in Rafah itself.

The war in Gaza raises complex dilemmas, owing to the extent of embeddedness of armed groups in the urban infrastructure both above and underground, the offensive capabilities and motivation they displayed on October 7 and since, as well as the presence of hostages. These factors implicate both the safety of the local population and give rise to operational challenges across the board. When military considerations arise, however, they must remain military and not political; and even when they remain as such, some rational balance must exist between benefit and harm. Proportionality as a general principle under IHL best captures this balance and closes the normative void that otherwise exists. The conditions in South Gaza are a quintessential example of harm that one would be incredibly hard-pressed to justify, even in relation to civilian safety or military considerations in North Gaza.

In sum, there are two potentially lawful grounds for evacuations: safety of civilians and, in cases of occupation, imperative military considerations. Any evacuation of civilians on these grounds must be temporary and closely related to the underlying lawful cause. The initial grounds for the evacuation of North Gaza in October –which were officially predicated on the safety of the civilian population – cannot be said to persist at the level of intensity that putatively justified the evacuation almost four months later, and accordingly cannot alone justify maintaining the situation. If the cause of evacuation is related to imperative military considerations, these should be concrete, defined, and limited. Perhaps most importantly, proportionality requires Israel to take into account the humanitarian calamity in South Gaza. Owing to the scale of this crisis, that humanitarian cost almost certainly outweighs any putative safety or military consideration in letting civilians return to relative safety in North Gaza; indeed, especially when those competing safety and military interests could be afforded through other measures.

Photo credit: Displaced Palestinians are pictured along with makeshift tents on December 21, 2023 in Al-Mawasi, Rafah, Gaza (Ahmad Hasaballah/Getty Images)