The Attack at the Zaporizhzhia Nuclear Plant and Additional Protocol I

On March 4th, the International Atomic Energy Agency (IAEA) announced a report by Ukrainian authorities that Russian forces had attacked the Zaporizhzhia Nuclear Power Plant in southeast Ukraine. Russia denied this characterization. The IAEA reported that a building in the vicinity caught fire, but that essential equipment was not damaged and that there were no reported changes in radiation levels at the plant. Later that day, United Nations (UN) Under-Secretary-General for Political and Peacebuilding Affairs Rosemary DiCarlo briefed the Security Council, describing Russia’s activities as “highly irresponsible” (at 1:18) and in violation of Article 56 of Additional Protocol I (at 2:04). U.S. Ambassador to the UN Linda Thomas-Greenfield characterized the attack as “incredibly reckless and dangerous” due to the threat posed to “the safety of civilians across Russia, Ukraine, and Europe.” In the same meeting, French Ambassador to the UN Nicolas de Rivière described the action as “dangerous and illegal.” The U.S. Embassy in Kyiv tweeted, “It is a war crime to attack a nuclear power plant.”

In this post, I analyze both the legal framework provided in Article 56 of Additional Protocol I and the associated war crime codified in Article 85(3)(c) of the Protocol.

Additional Protocol I

As both the Russian Federation and Ukraine are States Parties, Protocol I is applicable to the international armed conflict between them, pursuant to the terms of Articles 1 and 96. The rule codified in Article 56 of Protocol I is more demanding than is the analogous customary rule, at least as that rule is understood by the International Committee of the Red Cross (ICRC) (see also p. 228), so my focus here is on the treaty rule. To inform what follows, it may be of some use to replicate the first two paragraphs of Article 56 in their entirety:

Article 56 – Protection of works and installations containing dangerous forces

1. Works or installations containing dangerous forces, namely dams, dykes and nuclear electrical generating stations, shall not be made the object of attack, even where these objects are military objectives, if such attack may cause the release of dangerous forces and consequent severe losses among the civilian population. Other military objectives located at or in the vicinity of these works or installations shall not be made the object of attack if such attack may cause the release of dangerous forces from the works or installations and consequent severe losses among the civilian population.

2. The special protection against attack provided by paragraph 1 shall cease:

(a) for a dam or a dyke only if it is used for other than its normal function and in regular, significant and direct support of military operations and if such attack is the only feasible way to terminate such support;

(b) for a nuclear electrical generating station only if it provides electric power in regular, significant and direct support of military operations and if such attack is the only feasible way to terminate such support;

(c) for other military objectives located at or in the vicinity of these works or installations only if they are used in regular, significant and direct support of military operations and if such attack is the only feasible way to terminate such support.

These first two paragraphs include a core prohibition and a narrow exception to that prohibition. They are worth evaluating in turn.

The Core Prohibition

The core prohibition is shaped by two extraordinary features. First, attacks are proscribed even if the nuclear power plant “glaringly” qualifies as a military objective under the standard articulated in article 52 (p. 227). In other words, even if the plant is such that by its “nature, location, purpose or use” it unambiguously makes “an effective contribution to military action,” such that its “total or partial destruction, capture or neutralization, in the circumstances ruling at the time, [would] offer a definite military advantage,” it is still protected from attack (compare art. 52(2)). Second, attacks on military objectives located in the vicinity of the station are also prohibited. In both respects the rule deviates from the ordinary law of armed conflict framework, pursuant to which military objectives may be targeted, with civilians and civilian objects protected from the effects of such targeting by the cumulative requirements of discrimination, proportionality, and precautions in attack.

By prohibiting the targeting of what would otherwise qualify as clear military objectives, Article 56 shifts the rule of distinction. This heightened protective framework is triggered by the combination of two elements. First, the work or installation in question must be a dam, dyke, or nuclear electrical generating station containing dangerous forces. Second, it must be the case that the attack (whether on the installation or on military objectives in its vicinity) “may cause the release of dangerous forces … and consequent severe losses among the civilian population.”

The first element is relatively straightforward in the case of the Zaporizhzhia Plant. It is the largest nuclear power plant in Europe and the ninth largest in the world. The second element is more complicated. It requires an assessment of the relationship between the attack and the possible release of dangerous forces and any consequences arising from it. In the case of the Zaporizhzhia Nuclear Power Plant, the IAEA reported “no release of radioactive material.” However, for reasons elaborated below, this does not dispose of the legal question, as the prohibition attaches to the risk of release ex ante.

At first glance, the prohibition in paragraph 1 of Article 56 might look like a specific operationalization of the proportionality rule (p. 228). As articulated in Article 51(5)(b) of Protocol I, the latter supplements the rules on distinction, discrimination, and precautions by prohibiting “an attack which may be expected to cause incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof, which would be excessive in relation to the concrete and direct military advantage anticipated.”

However, the core prohibition in Article 56 is different. It is internal to distinction. If the threshold is met, the object is presumptively not a legitimate target and attacking it is prohibited, regardless of the countervailing military advantage and regardless of whether every feasible precaution was taken to minimize civilian loss. Additionally, and of critical importance here, the epistemic threshold is not the same as that applicable to proportionality. Article 56(1) focuses on whether the attack “may” cause the release of those forces and consequent severe losses among the civilian population. That stands in contrast to the use of “may be expected” in Article 51(5)(b) (emphasis added).

As such, the Article 56(1) prohibition on targeting nuclear power plants should not be understood to be triggered only when those engaged in the attack expect or ought to expect that their operation will release radioactive material and cause severe civilian loss. It is enough that the attack entails that risk. Put another way, for the prohibition not to attach, it must be the case that the attacking force is certain that radioactive material will not be released or that if such a release were to occur, the civilian population would not suffer severe losses. As the ICRC Commentary puts it, an attack would avoid the Article 56(1) prohibition if it “cannot cause severe losses” (para. 2153) (emphasis added).

Certainty that there will be no radioactive leak is conceivable in the context of a precisely targeted and limited operation that avoids any danger to stocks of radioactive material, to elements essential to the cooling system, or to the core of the installation. Asserting its compliance with this standard in the context of dams and dykes, the Farabundo Martí National Liberation Front (FMLN) claimed in 1988 that its forces focused exclusively on “annihilating the troops stationed” at hydro-electric plants in El Salvador and “damaging the machinery,” but specifically not destroying the dams or dykes precisely so as to comply with the requirements of Article 56 (p. 382). As noted below, the Russian Federation appears to have adopted a similar posture in the Zaporizhzhia case.

Alternatively, certainty that a release of the dangerous forces would not cause severe civilian losses could arise if the object were in a remote area without significant civilian habitation (para. 2154; p. 228). Given the enduring impact of a radioactive leak, such an analysis is more straightforwardly conceivable in the case of dams and dykes. Any such analysis would also need to consider Articles 35(3) and 55 of Protocol I, which prohibit attacks or methods of warfare that “may be expected” to cause “widespread, long-term and severe damage” to the natural environment, regardless of the military advantage such attacks may return.

Early analysis after the reported attack on the Zaporizhzhia plant suggested that the risk of a radioactive release was “low” during the fighting. However, it is less clear that those engaged in the operation could have been certain of that at the time. Certainly, shelling the vicinity does not offer the tight control that might ordinarily underpin certainty that there would be no radioactive release. The Russian Federation has denied engaging in such activity, claiming instead:

Russian military were heavily fired at from windows of several levels of a training facility building that was located outside the territory of the ZNPP. This was done to provoke retaliation. Russian patrol started retaliatory fire and eliminated the gun posts of Ukrainian sabotage group in the training center. When leaving the building, Ukrainian saboteurs set it on fire. Let me emphasize that the facility was located outside the ZNPP territory.

Notably, IAEA Director General Rafael Mariano Grossi “remained gravely concerned about the situation” at the power plant, even once it was established that the initial fighting had not caused a leak. Reporting since this blog post was written has indicated that the attack entailed significant risk and has made the official Russian position difficult to sustain.

Crucially, Article 56(1) imposes an ex ante prohibition. The absence of a radioactive leak or severe civilian losses does not preclude the unlawfulness of the attack, if perpetrated despite the risk of such an outcome. The virtue of codifying a rule of this kind is that it preempts the ambiguities inherent in assessments of feasible loss mitigation and proportionality, replacing those ambiguous principles with a simple presumption: do not target nuclear electrical generating stations or military objectives in their vicinity.

Exceptions

Deviation from that presumption is permissible only in two scenarios. The first, as already noted, is when the commander can be certain that the dangerous forces will not be released, or that their release will not cause severe civilian losses. In that case, the heightened protection is not triggered, so the station or objects in its vicinity could qualify as legitimate targets, although any attack would still need to comply with the applicable rules on doubt, discrimination, precautions, and proportionality.

The second basis for deviating from the presumptive ban on attacking nuclear power plants or military objectives located in their vicinity applies even if the plant meets the criteria for heightened protection noted above. This exception is framed slightly differently in its application to attacks on the power plant and to attacks on military objectives in its vicinity.

Paragraph 2(b) of Article 56 specifies that the special protection provided for a station meeting the criteria discussed above shall cease “only if it provides electric power in regular, significant and direct support of military operations and if such attack is the only feasible way to terminate such support” (emphasis added). By the terms of this provision, there can be no other exception to the heightened protections provided in paragraph 1. Additionally, the exception is a demanding one, including three substantive criteria and one procedural criterion.

Substantively, the station must provide electric power in “direct support of military operations.” This requires a significantly tighter connection than is specified in the general definition of military objectives, which refer only to those that “make an effective contribution to military action,” whether or not that contribution is direct (Art. 52(2)). It is analogous instead to the permissibility of targeting civilians, which applies only for such time as they take a “direct” part in hostilities. (Art. 51(3)). An influential (albeit disputed) analysis understands the latter to attach only when the activity is a single causal step from the infliction of harm on the adversary (pp. 52-54). Interpreted in that light, the threshold would be very difficult to satisfy. In particular, it is disputed whether “merely supplying electricity” to an integrated grid with “pool[ed] generating capacity,” from which the military then draws, would be sufficient to constitute “direct support of military operations” (para. 2165). Certainly, the electrical supply must directly “benefit military operations themselves and not merely intermediary objectives which themselves would be related to such operations” (para. 2162). Even a nuclear power plant delivering power to “industrial works of the defense industry” might be deemed “debatable” on the grounds that “[w]hat directly supports warfare in such a case is the industry supplied with the power, not the power station in itself” (p. 231).

Additionally, the support to military operations must be “regular.” This criterion would not be satisfied by the military occasionally drawing on this power source or having it as a backup power source. The ICRC Commentary suggests that “there must be some continuity in the use, or at least some rhythm.” (para. 2162). Finally, on the substantive criteria, the support must be “significant.” Here, it must be established that the contribution to military operations is “sizeable” and therefore has “a real and effective impact.” (para. 2162).

Even when each of these substantive thresholds is satisfied, an attack would still be precluded if there is any alternative feasible way to terminate that support. In addition to warnings and requests for cessation, one obvious alternative way to terminate the electrical support of the power station would be to target the main circuit lines or transformer stations at a remove from the station sufficient to eliminate the risk of releasing the dangerous forces contained within (para. 2166; p. 230).

Paragraph 2(c) of Article 56 applies a similar, but not identical, framework to military objectives at or in the vicinity of the station. Specifically, such objectives may be attacked “only if they are used in regular, significant and direct support of military operations and if such attack is the only feasible way to terminate such support” (emphasis added). The only significant difference here is that the exception covers support other than the provision of electric power. Each of the substantive and procedural thresholds remains the same. Even when an attack would be permitted under the exclusive permission provided in paragraph 2(c), paragraph 5 of the article clarifies that “installations erected for the sole purpose of defending the protected works or installations from attack” are not to be made the object of 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.” In short, purely defensive installations attached to the nuclear power plant and used in purely defensive ways cannot be understood to qualify as military objectives providing regular, significant, and direct support of military operations.

It has been suggested that military objectives regulated in paragraph 5 are exempt from the standard articulated in paragraph 2(c) (e.g., para. 2155). However, this is contrary to the text. Paragraph 2(c) is clear in specifying the exclusive context in which military objectives “located at or in the vicinity of” a nuclear power plant may be targeted. Paragraph 5 instead regulates passive precautions and the exceptions thereto. It supplements the prohibition in paragraph 2(c); nothing in the provision suggests that it is meant to create an exception to that prohibition.

Before turning to passive precautions, it is important to emphasize, per the terms of paragraph 3 of Article 56, that the application of this framework at the level of distinction does not replace the requirements of discrimination, precautions, and proportionality. Those are cumulative requirements that would apply to constrain an attack even if it were not precluded by the terms of Article 56.

Notably, the Russian account of the situation makes little attempt to frame the operation in the specific terms of the exceptions in Article 56. Instead, as noted above, the Russian Federation has framed its operations on March 4th as not endangering the plant in the first place.

Passive Precautions

In addition to the rules on targeting military objectives, paragraph 5 of Article 56 also imposes specific passive precautions associated with the location of military objectives in the vicinity of such stations. Specifically, parties to the conflict are to “endeavour to avoid” locating military objectives in that vicinity, but for defensive installations of the kind mentioned above. If Russia’s claim to have been “heavily fired at” from the vicinity of the power plant were to be borne out, this could indicate a violation of Article 56(5) on the part of those who initiated that firefight.

Significantly, however, a party’s failure to comply with that requirement would not itself eliminate the protection accorded to any military objectives in the vicinity of the power plant. As noted above, paragraph 2(c) lays out the exclusive conditions under which such objectives may be targeted. Defensive installations used defensively cannot be targeted at all, but even military objectives that are not defensive and that ought not to have been located in the vicinity of the power plant cannot be targeted, except as permitted by the terms of Article 2(c).

The explicit specification in paragraph 4 of Article 56 further affirms that it is “prohibited to make any of the works, installations or military objectives mentioned in paragraph 1 the object of reprisals.” The fact that one side has breached the requirements of paragraph 5 does not allow breach of paragraph 2(c) as a mechanism of law enforcement. This is consistent with the general posture of Additional Protocol I on the protection of the civilian population as articulated in Article 51(8).

A War Crime?

The threshold for the war crime, as specified in the grave breaches regime of Additional Protocol I, is notably higher than is the threshold for the underlying provision. Article 85(3)(c) identifies that breach as “launching an attack against works or installations containing dangerous forces in the knowledge that such attack will cause excessive loss of life, injury to civilians or damage to civilian objects, as defined in Article 57, paragraph 2 (a) (iii).” Additionally, per the terms of the chapeau, this rises to the level of a grave breach only when the attack is “committed wilfully, in violation of the relevant provisions of this Protocol, and causing death or serious injury to body or health” (emphasis added).

Three things are worth noting about this articulation of the war crime, as compared to the underlying rule. First, the epistemic threshold is far higher. For an individual to be criminally liable for such an attack, that individual must have acted in the knowledge that the attack will cause excessive loss, injury, or damage to civilians or civilian objects. This, of course, requires a very different analysis from whether the context was such that dangerous forces may be released with severe losses for the civilian population.

Second, the reference to Article 57(2)(a)(iii) indicates that the war crime attaches only to attacks that would cause civilian loss, injury, or damage that would be “excessive in relation to the concrete and direct military advantage anticipated.” In other words, it is derivative of a proportionality assessment. The underlying rule, on the other hand, precludes attacks that risk severe losses, regardless of military advantage.

Third, the grave breach has a consequence element—the illegal attack must cause death or serious injury to body or health. To be clear, it need not be established that severe civilian losses occurred. It is relevant in that respect that two individuals were reported injured in the Zaporizhzhia operation. However, the underlying rule is violated when an attack is launched despite the risk of releasing dangerous forces with consequent severe civilian losses, whether or not any death or serious injury in fact occurs.

Conclusion

Ultimately, if it is established that Russian forces engaged in the shelling of the Zaporizhzhia plant or objectives in its vicinity in a way that risked a radioactive leak, it is almost certain that this operation violated Article 56. The Russian Federation, of course, contests the claim that its operation entailed any such risk. It has not, however, provided information indicating that if such a risk were taken, it would have been permitted pursuant to the exception provided in Article 56(2)(c).

It is less likely that the operation satisfied the threshold for the associated war crime, as articulated in article 85(3)(c). Given that there was in fact no radioactive leak and that there seems to have been relatively little collateral damage, it does not appear that those who engaged in the attack would have known at the time that excessive civilian loss would arise from it.

The virtue of the underlying rule is that it provides a bright line to commanders: do not target nuclear power stations or military objectives in their vicinity, except in very rare and specifically identified circumstances. The clarity of the presumption is such that any decision that one of the exceptions applies ought to be made at a high level of command (para. 2159) and the justification ought to be clear cut. In cases such as this, it is crucial not to allow the difficulty of establishing the war crime to muddy the clarity of that underlying prohibition.

***

Tom Dannenbaum is Assistant Professor of International Law at the Fletcher School of Law and Diplomacy.

Photo credit: Ralph1969 via Wikimedia Commons