In Honor of Yoram Dinstein – Civilian Evacuations and the Law of Armed Conflict

Editors’ note: This post is part of a series to honor Professor Yoram Dinstein, who passed away on Saturday February 10, 2024. These posts recognize Professor Dinstein’s work and the significant contribution his scholarship has made to our understanding of international law.

Operational environments and their attendant challenges vary widely in armed conflict. Few would doubt, however, that conducting military operations in urban terrain presents extraordinary risks to its civilian inhabitants. Indeed, within highly populated areas collateral damage is too often exceedingly tragic. This is especially true considering that, in some conflicts, parties may intentionally endeavor to exploit local civilian populations (e.g., human shielding). But examples are few in which the course of lawful military operations does not result in incidental civilian harm.

It is against this grave backdrop that evacuations have long served as an important tool with which parties to an armed conflict may (and in some cases, must) protect affected civilian populations. Following the outbreak of the Second World War, for example, the British famously evacuated children and others away from London and other cities to live with family or strangers in rural areas in Operation Pied Piper. More recently, Ukrainian authorities have evacuated civilians near the Russian border due to the dangers of artillery bombardments. Similar operations have characterized the Israel-Hamas conflict in Gaza (see, e.g., here and here), which this post assumes is non-international in character. Evacuations are, therefore, not new to warfare.

Far from a panacea, evacuations may themselves lead to humanitarian crises. Their lawful execution is therefore of critical importance. Accordingly, this post provides a brief, general overview of the law of armed conflict rules pertaining to evacuations of civilians in international and non-international armed conflicts. Given its limited focus, it does not discuss obligations that may apply to other categories of persons, such as detainees, the wounded and sick, and the besieged, or under other bodies of law, including international human rights law. Nor does it address so-called noncombatant evacuation operations (NEOs), such as the one conducted by the United States during its withdrawal from Afghanistan. It begins from the perspective of those conducting attacks before turning to those with territorial control over the civilian population.

The Duty to Warn

A long-observed rule of customary international law is that parties must exercise feasible precautions to minimize the risk of any incidental civilian harm when planning and conducting attacks on otherwise valid military objectives. A corollary to these so-called precautions in the attack (or active precautions) is the more specific obligation to provide effective advance warning of attacks that may cause death or other physical harm to the civilian population, unless circumstances do not permit (on the requirement of physical harm, see Tallinn Manual 2.0, commentary to rule 120, para. 4; Baruch & Neuman, p. 374-75). As another Articles of War post clarifies, there is general consensus that the customary duty to warn applies in both international and non-international armed conflicts (but see Dinstein, para. 813).

A firmly established caveat to the rule, however, is that only those measures that an attacker determines in good faith and based on available information to be feasible are required. Such determinations are, of course,highly contextual. As the U.S. Department of Defense (DoD) Law of War Manual helpfully explains, “[f]easible precautions are those that are practicable or practically possible, taking into account all circumstances ruling at the time, including humanitarian and military considerations” (§ 5.2.3.2). With respect to feasibility in the context of evacuations, warnings are not required if they would compromise an attack’s operational effectiveness by, for example, sacrificing the elements of surprise, speed of maneuver, or security of the attacking or friendly forces, all critical components of battle-proven principles of war (§ 5.11.5.4; International Committee of the Red Cross (ICRC), Customary International Humanitarian Law (CIHL) Study, rule 20; Dinstein, para. 580).

Contrast, for instance, attacks on military objectives that vary based on their nature, location, purpose, or use. Whereas warnings to evacuate one or more residential buildings used to store small arms or ammunition would likely allow enemy forces to remove those materials before an attack, it is more feasible, relatively speaking, to evacuate military objectives that would retain their character and all or some of the targeting value thereafter (e.g., a bridge, munitions factory, or military airfield). Still, there are myriad sound operational reasons why it may ultimately be infeasible to evacuate those latter objectives beforehand. Some common means of conveying evacuation warnings presume a degree of air or electromagnetic control (e.g., radio broadcasts, leaflet drops, etc.). As these examples illustrate, considering feasibility’s inherently contextual nature, “it would be inaccurate to conclude that specific precautions are required as a general rule” (DoD, Law of War Manual, § 5.2.3.2).

If feasible, warnings must be effective. In this context, it is helpful to recall that

[t]he purpose of a warning is to facilitate the protection of the civilian population so that civilians and the authorities in control of the civilian population can take measures to reduce the risk that civilians will be harmed by military operations. Although there is no set form for warnings, a warning should be designed to accomplish this purpose to the extent feasible” (DoD Law of War Manual, § 5.11.5.2; ICRC, Commentary to Additional Protocol I (AP I), art. 57, para. 2225; Dinstein, para. 579).

To be effective, therefore, an evacuation warning must be communicated at such a time and in such a manner as to provide the affected civilian population with a reasonable opportunity to meaningfully protect itself.

As this implies, the duty to warn is prospective in nature. Effectiveness thus does not turn on whether, or if so how, the civilians in question heed a warning. Their decision to evacuate pursuant to an attacker’s duty to warn is inherently voluntary. Although an effective warning must allow civilians a reasonable amount of time to evacuate a military objective if feasible, they are not obliged to do so. Accordingly, if the civilians refuse to evacuate, they retain their immunity from being made the object of attack and should still be accounted for in determining the proportionality of any collateral damage.

Nor is there a firm requirement that evacuation warnings provide details of an attack, such as when, where, or how it will occur, or otherwise specify how civilians should evacuate or otherwise avoid harm. Commensurate with the feasibility of precautions generally, as Colonels Baruch and Neuman prudently explain, “the degree to which a warning must be specific and detailed is dependent on the context and circumstances of the situation” (p. 383; Air and Missile Warfare (AMW) Manual, commentary to rule 37, para. 10). This no doubt applies to calls to evacuate as well. After all, considering the uncertainty with which military operations historically unfold, “it is not always clear where the fighting will take place, what targets will be attacked and which areas will be safer than others. Much depends, of course, on the actions of the enemy forces” (p. 377).

As Professor Dinstein succinctly observed, “[i]t is not easy to determine what kind of advance notice would constitute an effective warning, nor is it clear how specific and direct the warning has to be” (para. 580; Baruch & Neuman, p. 377-78). In light of such difficulties, and although parties have a duty to exercise precautions before each discrete attack, it is important to note that evacuation warnings may also be general in character. That is, whereas for the reasons previously mentioned attackers may find it infeasible to provide a specific warning before a particular attack, they may likely (and often do) reach different conclusions with respect to providing, for instance, a generalized description at the outset of an operation of the types or locations of military objectives that may be attacked throughout its duration (DoD, Law of War Manual, § 5.11.5.2; ICRC’s 1987 Commentary to AP I, art. 58, para. 2225; ICRC, CIHL Study, commentary to rule 20). At the outset of the invasion of Gaza, for example, Israel announced a call “for the evacuation of all civilians of Gaza City . . . an area where military operations take place.” Indeed, as the commentary to the AMW Manual notes with respect to timing, even “an imprecise warning issued well in advance of the attack may be more effective than a precise warning immediately preceding it” (commentary to rule 37, para. 9).

In addition to active precautions, the law of armed conflict also imposes special warning requirements that implicate civilian evacuations (DoD, Law of War Manual, § 5.11.5.1). In the event that medical facilities, for instance, are used to commit “acts harmful to the enemy,” in which case their special protections from attack are forfeited, they may be attacked “only after due warning has been given, naming, in all appropriate cases, a reasonable time limit, and after such warning has remained unheeded” (Geneva Convention IV (GC IV), art. 19). The obligation applies equally to non-international armed conflicts (ICRC, CIHL Study, rule 28). Prior to raids on the al-Shifa hospital, for example, Israeli officials complained for “weeks” that the hospital was the “‘beating heart’ of [Hamas’s] command infrastructure in northern Gaza,” which if true made it a valid military objective.

In case of either duty, genuine calls to evacuate must be distinguished from unlawful threats. Regardless of a conflict’s character, customary international law prohibits “acts or threats of violence the primary purpose of which is to spread terror among the civilian population” (ICRC, CIHL Study, rule 2; AP I, art. 51(2); Additional Protocol II (AP II), art. 13(2)). Their critical distinction is the subjective intention of the warning party. If the aim of a call to evacuate is to minimize or avoid incidental harm to the civilian population, it would not constitute a terroristic threat. On the other hand, a knowingly false warning that civilians may be made the object of attack in the event they do not evacuate, likely would.

The Duty to Remove

The Prohibition Against Forced Movement

Before turning to the duty to remove, it must be emphasized that in many cases involuntary displacement of the civilian population is unlawful. With respect to international armed conflicts, Article 49 of Geneva Convention IV provides that “[i]ndividual or mass forcible transfers, as well as deportations of protected persons from occupied territory to the territory of the Occupying Power or to that of any other country, occupied or not, are prohibited, regardless of their motive.” Violation of the prohibition, which is also customary in nature (ICRC, CIHL Study, rule 129; 1863 Lieber Code, art. 23), is a grave breach of the Convention, a war crime, and depending on its scale, a crime against humanity (GC IV, art. 147; Rome Statute, arts. 7(1)(d), 8(2)(b)(viii)). Recall that only those qualifying as “protected persons” within occupied territory fall within its protections (GC IV, art. 4).

With respect to non-international armed conflicts, parties similarly “may not order the displacement of the civilian population, in whole or in part, for reasons related to the conflict . . .”(ICRC, CIHL Study, rule 129; AP II, art. 17). Doing so is also a war crime (e.g., Rome Statute, art. 8(2)(e)(viii)). As is apparent, an element of force or other coercion is required to violate either rule; voluntary movement does not suffice (Blagojević and Jokić, para. 596; ICRC, Commentary to AP II, art. 17, para. 4851; Bothe et al., p. 795).

Critically, however, these otherwise broad prohibitions carve out an exception for evacuations. After setting forth the prohibition, for instance, Article 49 further observes,

Nevertheless, the Occupying Power may undertake total or partial evacuation of a given area if the security of the population or imperative military reasons so demand. Such evacuations may not involve the displacement of protected persons outside the bounds of the occupied territory except when for material reasons it is impossible to avoid such displacement. Persons thus evacuated shall be transferred back to their homes as soon as hostilities in the area in question have ceased. 

The limits and conditions of such evacuations are self-apparent. In non-international armed conflicts, it is similarly permissible to wholly or partially evacuate the civilian population if “the security of the civilians involved or imperative military reasons so demand” (ICRC, CIHL Study, rule 129; AP II, art. 17). That these exceptions allow for the permissive, albeit conditional, evacuation of civilians is manifest.

Precautions From the Effects of Attack

Under certain circumstances, evacuation is not merely permitted but is in fact required. The duty to exercise precautions by those subject to attacks (or passive precautions) is set forth in treaty form in the 1977 Additional Protocol I to the 1949 Geneva Conventions. In relevant part, Article 58 provides that “[p]arties to [a] conflict shall, to the maximum extent feasible . . . endeavour to remove the civilian population, individual civilians and civilian objects under their control from the vicinity of military objectives.” Although AP I is not universally ratified, this provision is widely recognized as reflecting customary international law applicable to both international and non-international armed conflicts (ICRC, CIHL Study, rules 22 and 24; Kupreškić, para. 524). Note, in contrast to active precautions, which apply to all attacks that may affect the civilian population, the duty to exercise passive precautions prescribes actions that must be taken only in territory under a defender’s control (ICRC, 1987 Commentary to AP I, art. 58, para. 2239).

As with active precautions, only feasible measures are required. Relevant considerations in the evacuation context include, for instance, the risk(s) that removal operations would pose to the party’s overall mission or its forces, the availability of necessary resources (e.g., troops, time, funding, etc.) compared to operational requirements, and the extent to which the civilian population, either in whole or in part, would likely benefit from evacuation. As the latter suggests, it may only be feasible to evacuate the more vulnerable portions of the civilian population, such as children, the elderly, or the infirm (ICRC, 1987 Commentary to AP I, art. 58, para. 2247). Given the uncertainty of military operations, evacuations may also do more harm than good. But assuming evacuation is feasible, failure to do so raises the specter of human shielding, a topic beyond the scope of this post.

Due to the overlap between an attacker’s duty to warn, a defender’s duty to take passive precautions, and the prohibitions against displacement, their practical application can present challenges (see, e.g., here). Whether Israel’s call to evacuate northern Gaza was a warning or an order to involuntarily displace, for example, has been the subject of much debate (see, e.g., here, here, here, and here). The distinction between the two is, in my view, highly fact-dependent and contextual. Concerning the latter, I believe an essential factor is the degree to which a warning party exercises control over the civilians in question. While effective control over the territory may not be a necessary condition to all manners of effecting displacement (Ntaganda, paras. 550, 559; Dörmann, p. 473), I struggle to see how the call to evacuate could be objectively interpreted as a compulsory order under the circumstances.

In this regard, it is helpful to consider the practical relationship between active and passive precautions; in many cases, an attacker’s call to evacuate enables the public authorities with control over the civilian population to carry out such a warning. Nor am I persuaded that the warning should be construed as such an order in concert with the overall conduct of Israeli military operations against Hamas. However one characterizes Israeli’s actions in Gaza, and even if parties have a duty to avoid displacement indirectly caused by their unlawful acts (ICRC, CIHL Study, commentary to rule 129; Willms), I find a contrary conclusion difficult to reconcile with (and likely to have detrimental implications for) the affirmative duty to warn.

Duty of Care

Such interpretive challenges are not merely academic in nature, for evacuations may trigger humanitarian obligations. With involuntary evacuations of protected persons within occupied territory, for example, “[t]he Occupying Power undertaking such transfers or evacuations shall ensure, to the greatest practicable extent, that proper accommodation is provided to receive the protected persons, that the removals are effected in satisfactory conditions of hygiene, health, safety and nutrition, and that members of the same family are not separated” (GC IV, art. 49(3)); ICRC, CIHL Study, rule 131). Occupying powers are also obliged to consider the destination of those evacuated, which should not include “area[s] particularly exposed to the dangers of war unless the security of the population or imperative military reasons so demand” (GC IV, art. 49(5)).

Similar obligations apply in the context of non-international armed conflicts. In the event of an order to displace, for instance, “all possible measures shall be taken in order that the civilian population may be received under satisfactory conditions of shelter, hygiene, health, safety and nutrition” (AP II, art. 17). State practice confirms this duty of care—debate over which has permeated the Israel-Hamas conflict—is customary in character (see, e.g., ICRC, CIHL Study, rule 131; DoD, Law of War Manual, § 17.9.1). In anticipation of a ground incursion into Rafah, for example, Israel has reportedly incorporated the concept of evacuating civilians to so-called “humanitarian islands” into its operational planning. While details are few, an Israeli spokesman indicated that such enclaves would provide civilians with “temporary housing, food, water, and other necessities” in cooperation with international actors. It is difficult to determine at this stage whether Israel’s call to evacuate Rafah would constitute an order to displace, and if so, whether the islands in question would comply with the customary obligations described above. Regardless, should civilians relocate to an area within Israel’s effective control, its forces would, in my view, owe them similar obligations akin to those observed by occupying powers in international armed conflicts in addition to any human rights obligations that may apply.

Conclusion

Although the character of armed conflict changes frequently, its nature seldom does. Few understand this better than the civilian victims of war. One estimate approximates that civilians account for up to 90 percent of wartime casualties, in large part due to the fact that modern warfare in urban terrain is increasingly prevalent and devastating. Accordingly, though subject to the law of armed conflict, evacuations remain a critical tool with which to protect civilians from the brutal effects of hostilities.

***

Lieutenant Colonel William C. Biggerstaff is a military professor at the Stockton Center for International Law at the U.S. Naval War College, where he co-teaches a course on the Law of Armed Conflict.

Photo credit: Yan Boechat/VOA