Ukraine Symposium – The Escalating Military Use of the Zaporizhzhia Nuclear Plant

By Tom Dannenbaum, Assistant Professor of International Law at The Fletcher School of Law and Diplomacy at Tufts University

In early March, I argued on these pages that Russian forces’ reported attack on the Zaporizhzhia Nuclear Power Plant likely violated the Additional Protocol I provisions protecting works containing dangerous forces, but probably fell short of the associated war crime. Following the attack, Russia quickly gained control of the facility, subsequently drawing scrutiny for its potentially criminal mistreatment of plant officials. The Zaporizhzhia plant has now returned to headlines amid reports that Russia has deployed weapons, troops, and other military assets to the site and is using the facility as a base for strikes against nearby Ukrainian towns. News of further attacks on the plant (attributed by each side to the other, with Ukraine alleging Russian false flag operations) have also emerged. In response, UN Secretary-General António Guterres has called for the parties to agree to a demilitarized zone around the plant. Meanwhile, G7 foreign ministers have demanded immediate Russian withdrawal from the site. Describing Russia’s tactics as “unconcealed nuclear blackmail,” Ukrainian President Volodymyr Zelensky has warned: “Every Russian soldier who either shoots at the plant, or shoots using the plant as cover, must understand that he becomes a special target for our intelligence agents, for our special services, for our army.”

Supplementing my prior analysis of the law governing attacks on nuclear power plants and military objectives in their vicinity, this post draws on Additional Protocol I and the International Criminal Court (ICC) Statute to analyze the legal issues arising from Russia’s alleged military use of the plant, the call for a demilitarized zone, and the threatened Ukrainian response. Both Russia and Ukraine are States Parties to Protocol I, rendering the treaty straightforwardly applicable to the international armed conflict between them (see arts. 1 and 96). ICC jurisdiction arises from Ukraine’s 2015 declaration under Article 12(3) of the Court’s Statute.

Protocol I and the Co-Location of Military Objectives and Dangerous Forces

Article 58(c) of Additional Protocol I requires that parties to the conflict take “necessary precautions to protect the civilian population, individual civilians and civilian objects under their control against the dangers resulting from military operations.” Clarifying the application of this principle to the context at hand (even in contexts in which a nuclear power plant were to qualify as a military objective under the ordinary rules specified in Article 52), Article 56(5) specifies that belligerents must “endeavour to avoid locating any military objectives in the vicinity” of nuclear power plants. Within that provision, an exception is made for

installations erected for the sole purpose of defending the protected works or installations from attack … provided that they are not used in hostilities except for defensive actions necessary to respond to attacks against the protected works or installations and that their armament is limited to weapons capable only of repelling hostile action against the protected works or installations (emphasis added).

The first question this raises is whether Russia’s deployment of military assets to the plant qualifies for the exceptional permission to the prohibition articulated in Article 56(5). If so, this would leave only the general passive precautions rule in Article 58. The exception is a narrow one, including a cumulative requirement that the use, purpose, and capability of the military installation be exclusively defensive.

Given their apparent offensive use against targets in nearby Ukrainian towns, Russia’s deployments at the Zaporizhzhia plant would appear not to fall within that exception. It might also be inferred from the same reports that their deployment is at least partially offensive in purpose. As the requirement for an exclusively defensive use, purpose, and capability is cumulative, this renders the more complicated question of their capability moot. That complexity arises from the plant’s proximity to areas of active hostilities. As emphasized in the International Committee of the Red Cross Commentary to Additional Protocol I, the requirement that the weapons deployed be capable only of repelling hostile action is most plausibly dispositive when the protected work is far behind the front lines, in which case defending the installation entails protecting it from “attacks mounted by combatants who have been infiltrated or parachuted, or against attacks by guided missiles or projectiles dropped from aircraft. Thus there will be needed, on the one hand, a military guard equipped with light individual weapons, and on the other hand, anti-aircraft artillery” (para. 2175). Equipment other than that would be difficult to justify as solely defensive. It is harder to define the scope of armaments capable only of repelling hostile action against the plant when the facility is proximate to the area of combat, given the broader array of threats that could arise in such a scenario (e.g., para. 2176).

With the exception inapplicable on the basis of use and probably purpose, the question is whether Russia can be said to have violated the primary prohibition. Here, it is important to emphasize that parties to the conflict are not precluded from locating non-exempt military objectives in the vicinity of nuclear plants in all circumstances. Rather, they are required by the terms of Article 56(5) to “endeavour” to avoid such co-location. As suggested above, this is best understood as a specification of the general passive precautionary requirement that belligerents act to the “maximum extent feasible” to take “necessary precautions to protect the civilian population, individual civilians and civilian objects under their control against the dangers resulting from military operations” (art. 58(3)).

Assessing feasibility (and thus whether a party can be said to have endeavored to avoid co-location) is inevitably context-specific, requiring a careful analysis of alternative locations and the implications of placing military objectives in those positions. In tight conditions of urban warfare, for example, it can be impossible to avoid some proximity between military objectives and protected persons and objects (e.g., para. 64). Precisely this complexity has been central to the controversy surrounding Amnesty International’s assessment that Ukraine may be falling short of its passive precautionary obligations in the current conflict. Such determinations can involve multiple points of uncertainty and disagreement.

However, when troops, military equipment, or other military objectives are deliberately located in or around protected persons or objects with a view to deterring attacks, as has been alleged at the Zaporizhzhia plant, the case is clear. The deliberate exploitation of co-location entails an endeavor to co-locate, rather than an endeavor to avoid it, clearly violating the requirements of article 56(5). As elaborated below it may also implicate the ban on using human shields.

The War Crime of Using Human Shields

Article 51(7) of Protocol I provides, “The presence or movements of the civilian population or individual civilians shall not be used to render certain points or areas immune from military operations, in particular in attempts to shield military objectives from attacks or to shield, favour or impede military operations.” The associated ICC war crime prohibits “[u]tilizing the presence of a civilian or other protected person to render certain points, areas or military forces immune from military operations.”

In evaluating whether Russia’s alleged deployments at the Zaporizhzhia plant violate the human shields ban, three legal technicalities warrant consideration. The first relates to whether that Protocol I prohibition and particularly the associated ICC war crime attach to exploiting co-location with protected objects, such as nuclear power plants, as distinct from co-location with protected persons or populations. The second is whether it is plausible that those engaged in the alleged deployment of non-defensive military objectives to the facility might satisfy the criminal intent standard at the ICC. The third goes to the meaning of rendering the shielded military objective “immune” from attack.

On the first point, one of the elements of the ICC war crime specifies that the crime attaches only when “[t]he perpetrator moved or otherwise took advantage of the location of one or more civilians or other persons protected under the international law of armed conflict” (p. 30) (emphasis added). Notably absent from this provision are scenarios in which the perpetrator moved or otherwise took advantage of the location of an object protected under the international law of armed conflict, although the latter might be thought to be included in the associated customary prohibition (see, for example, the United States Department of Defense Law of War Manual, § 5.16.2). At the ICC, the exploitation of certain objects may qualify for the separate war crime of “[m]aking improper use of a flag of truce, of the flag or of the military insignia and uniform of the enemy or of the United Nations, as well as of the distinctive emblems of the Geneva Conventions” where such improper use results in death or serious personal injury (art. 8(2)(b)(vii)). However, although Article 56(7) of Protocol I recognizes an emblem specific to works or installations containing dangerous forces, a leading interpretation would exclude it from the improper use war crime on the grounds that the dangerous forces emblem, while recognized by Protocol I, is not a distinctive emblem of that treaty or the Geneva Conventions—a status reserved for the red cross, red crescent, and red crystal (p. 454). In any event, it is not clear that such an emblem is currently displayed at the Zaporizhzhia plant.

Focusing instead on Article 8(2)(b)(xxiii), the difficulty that might be raised is that Russian forces appear to be exploiting their military assets’ proximity to the plant (and not their proximity to civilians) as a way of gaining special legal, moral, and political protection. The crime, on the other hand, attaches only when the alleged perpetrator takes advantage of the location of civilians or other protected persons to confer immunity on its military objectives.

However, to read the reference to the “presence” or “location” of civilians in the ICC war crime as excluding co-location with a nuclear power plant would be contrary to the object and purpose of the criminal ban, which is to prohibit the deliberate endangerment of civilians for the purpose of deterring enemy military operations (p. 345; see also § 5.16.2). The essence of the ban on attacking nuclear power plants and military objectives in their vicinity inheres also in the imperative to protect civilians from the harms associated with such an attack—it obtains only when the operation “may cause the release of dangerous forces from the works or installations and consequent severe losses among the civilian population” (art. 56(1)). To shelter military objectives in order to exploit that protection is to deliberately endanger protected persons for the purpose of deterring enemy military operations.

The reference to the “presence” and “location” of civilians in the text of Article 8(2)(b)(xxiii) and the associated element must be read in this light. The nature of nuclear fallout is such that attacks on nuclear power plants can affect civilian populations distant in space and time from the targeted objective. This enables a party to the conflict to take advantage of a particularly expansive location or presence of at-risk civilians. Although civilians may be located at a greater remove from the relevant military objective than is ordinarily the case in a situation of human shielding, the deterrent arising from the co-location of military objectives with a nuclear plant arises from the risk to those populations due to their current or prospective presence or location in a contaminated zone. The status of the plant as an object should not preclude the applicability of this element of the crime.

The second issue warranting attention here is that the war crime of human shielding requires the specific intent to deter attacks through exposing civilians to risk. This is arguably what is meant by “utilising the presence of a civilian … to render certain points, areas or military forces immune from military operations” (art. 8(2)(b)(xxiii)). It is also the implication of the requirement that the perpetrator “moved or otherwise took advantage of” civilian persons’ presence, while “intend[ing] to shield a military objective from attack or shield, favour or impede military operations” (p.30).

This kind of exploitative intent is the central allegation of the Ukrainian authorities in describing Russia’s tactics as “unconcealed nuclear blackmail.” It is also a plausible inference from the reporting indicating that Russia is using the plant to launch regular attacks on the surrounding area. The prosecutor would, of course, bear the burden of proving mens rea of any particular individual in a criminal case, but there is reason to believe that there could be criminal intent in this situation.

The third issue relates to the notion that both the crime and the underlying prohibition attach only when the actor in question seeks to render the relevant points, areas, or military forces “immune from military operations.” In the ICC elements, this is framed as the intent to “shield” the relevant military objective or “favour” the relevant military operation.

The ban on attacking military objectives collocated with nuclear power plants is unusually robust. It attaches whenever such an attack “may cause the release of dangerous forces … and consequent severe losses among the civilian population” (art. 56(1)) (emphasis added). If an attack would risk such losses, it can be undertaken “only if [the military assets co-located with the plant] are used in regular, significant and direct support of military operations and if such attack is the only feasible way to terminate such support” (art. 56(2)(c)) (emphasis added). Even in that scenario, the ordinary precautions and proportionality requirements would very likely preclude most attacks. For these reasons, deploying military objectives to the plant almost certainly renders them legally immune from most conceivable Ukrainian operations.

However, even if it were thought that the exceptional conditions necessary to provide a legal basis for a Ukrainian strike on military objectives at the plant might obtain, co-location could still implicate the human shields bans. The concept of rendering the objective immune from military operations should be interpreted broadly to include taking advantage of co-location to deter attacks due to their human (or political costs) even if the strike to which the deterrent attaches would not qualify straightforwardly as unlawful (p. 345). Notably, the elements of the ICC crime do not require the exploitation of formal immunity. Instead, the key element is that the action was taken “to shield a military objective from attack or shield, favour or impede military operations.” (p. 30).

In the current case, Russian forces may be focused primarily on exploiting Ukraine’s unwillingness to expose its own population to catastrophic fallout from an attack on the plant, rather than on exploiting the legal constraints that would prohibit Ukraine from engaging in such an operation. However, on the question of whether the crime attaches, the distinction is immaterial. What matters is whether Russia is deliberately using the risk to civilians or civilian populations to deter attacks on its military objectives. Doing so would amount to shielding the relevant objective and favoring the Russian operations in which it is involved.

Proposals for a Demilitarized Zone

Given the potential for nuclear catastrophe, Secretary-General Gutteres’s motivation in urging the establishment of a demilitarized zone (DMZ) in the plant’s vicinity is obvious. Vasyl Nebenzia, Russia’s permanent representative to the UN, has rejected this call. Two points are of note with respect to this aspect of the situation.

The first is that Russia’s dismissal of the proposal implicates two provisions of Protocol I. Article 56(6) provides that the parties “are urged to conclude further agreements among themselves to provide additional protection for objects containing dangerous forces.” Article 58(3) requires that belligerents “take … necessary precautions to protect the civilian population, individual civilians and civilian objects under their control against the dangers resulting from military operations.” The use of the term “urged” in Article 56(6) indicates that it is hortatory, rather than strictly binding. However, the general precautionary requirement in Article 58(3) might be understood to demand a good faith consideration of the viability of such an agreement, particularly given the human stakes. It would, of course, be difficult to establish a clear violation of such a requirement in the current case.

The second point of note is that to qualify for demilitarized status, such a zone would need to be agreed by the parties to the conflict. Ordinarily, per the terms of Article 60 of Protocol I, the agreement would (a) require the evacuation of mobile weapons and military equipment, (b) preclude the use of fixed military installations in the zone, (c) prohibit acts of hostility by the authorities or population within the zone, and (d) require the cessation of any activity in the zone linked to the military effort (a condition the meaning of which is to be specified by the parties). An agreement to this effect could heighten the protections applicable in the environs of the Zaporizhzhia plant (and could apply to a broader area than the prohibitions in article 56). Additionally, per the terms of article 85(3)(d) of the Protocol, attacking a DMZ can constitute a grave breach of Protocol I.

However, given the apparent violations of existing requirements at the plant, there is reason to be cautious in assuming that the obligations that would follow from a DMZ agreement would improve the situation on the ground. Of course, a specific bilateral commitment may prove to be action-guiding in a way that the general commitments associated with ratification of Protocol I appear not to have been, particularly if the DMZ agreement were combined with a robust independent monitoring mechanism, but even this is far from guaranteed. It is also worth noting that a material breach of the DMZ agreement by one party would release the other from its obligations under the agreement (art. 60(7)). In that scenario, the zone would lose its demilitarized status, but other protections, including those in Article 56, would continue to apply (art. 60(7)).

Ukraine’s Threat to Russian Forces Involved in the Militarization of the Zaporizhzhia Plant

A final question arises as to how Ukraine might respond to Russia’s apparent military use of the nuclear plant. In contrast to the rules applicable to demilitarized zones, a violation by one party of its obligations under Article 56 of Protocol I does not release the adversary from its obligations under the article. Additionally, Article 56(4) precludes reprisals against the works, installations, or military objectives protected by Article 56. As noted above, anything but a narrowly targeted and tightly limited Ukrainian strike on a military objective in the plant’s vicinity would almost certainly be prohibited, despite Russia’s apparent abuse of the site. Quite apart from the legal reasons not to do so, it seems highly unlikely that Ukraine would consider a prohibited attack on the plant or a military objective in its vicinity, given the profound risk to its own population.

However, as noted above, Ukrainian President Volodymyr Zelensky has promised that “[e]very Russian soldier who either shoots at the plant, or shoots using the plant as cover” will become “a special target for our intelligence agents, for our special services, for our army.” This appears to be a commitment to use targeted lethal force to punish and deter those involved or potentially involved in the military use of the nuclear facility. There is no basis in the law of armed conflict for using lethal force as an extrajudicial method of punishing the adversary. Moreover, the practice of targeted killing has generated a range of legal concerns (e.g., here and here), particularly insofar as such strikes are used against persons other than traditional combatants (implicating legal debates as to the contexts in which such persons can be targeted in war and the certainty with which target status must be determined) and in contexts that might be thought to extend beyond the geographic or temporal scope of the armed conflict(s) upon which they are purportedly predicated. Such objections also often implicate debates as to the proper interaction between human rights law and the law of armed conflict.

As things stand, however, these concerns are unlikely to be determinative in the current case. An international armed conflict is clearly ongoing in Ukraine and Zelensky’s statement is specific to “soldiers”—i.e., formal combatants in the clearest sense (art. 43(2)). In such circumstances, Russian forces are lawful targets by their status as combatants until such time as they are discharged from the military or rendered hors de combat by wounds, sickness, capture, or surrender. Thus, although punishment is not a legal basis for targeting them, their status as combatants is and the (retributive or deterrent) motive for such targeting is legally irrelevant. Additionally, although it has generated critical reflection in other contexts, the fact that specific individuals would be identified for targeted killing does not itself implicate any prohibition under the current law of armed conflict.

To be clear, the other rules governing the conduct of hostilities would still apply in constraining operations against such individuals. Moreover, if such individuals were in a situation such that they could be straightforwardly captured rather than killed, it has been argued that the law of armed conflict would require the use of that less harmful means. Certainly, human rights law would demand that level of care, if applied cumulatively with the law of armed conflict or using a process of coordinated interpretation (rather than through the law of armed conflict as lex specialis). However, as articulated, Zelensky’s statement does not itself indicate that these rules would be violated in operations against soldiers involved in the militarization of the plant.


Ultimately, based on current reporting, the Russian Federation appears to be violating its passive precautionary obligations (both under Article 56 and under Article 58 of Additional Protocol I) by locating non-defensive military objectives in the vicinity of the Zaporizhzhia nuclear power plant. Indeed, there is good reason to believe that Russian forces involved in that co-location could be criminally liable for using human shields under the terms of Article 8(2)(b)(xxiii) of the ICC Statute. A demilitarized zone around the plant would specify a particular and direct commitment and could broaden the protected area and bolster monitoring. However, the necessary agreement for such a zone appears unlikely at this stage and, given the existing flouting of clear rules, it would be no panacea. Even as Russia continues to use the plant to launch attacks, it is likely that anything other than a very tightly targeted Ukrainian response would violate Article 56 of Protocol I. However, Ukraine’s threat to target those involved in the military use of the plant is at least potentially compatible with the law of armed conflict, depending on the details of how it is pursued.

This piece is republished from Lieber Institute.

Leave a Reply