Indirect or Reverberating Excessive Collateral Damage in Modern IHL
Article 51(5)(b), of the 1977 Additional Protocol I to the Geneva Conventions of 1949 codifies the rule or principle of proportionality in offensive or defensive attacks in the following terms,
Among others, the following types of attacks are to be considered as indiscriminate: … b) 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.
For many years, the incidental damages considered under this provision were held to be those which occur at some range or radius around the place of impact of a military ordnance of whichever type. In other words, the provision was considered to apply only to direct collateral (or in legal jargon “incidental”) damage (i.e. to those which are felt by persons or which affect objects around the place of impact just mentioned).
In more recent years, there has been a move away from this quite narrow reading. Legal writings and practice progressively require the commander to take account also of indirect or reverberating collateral damage to places and persons not directly surrounding the place of impact (see International Committee of the Red Cross (ICRC) & Université de Laval Study, p. 43; see also Robinson & Nohle, p. 107-45).
A breach of the rule requires that the damage be “excessive” in the sense of the provision, and that it is also was reasonably foreseeable, a thing which is less obvious in cases of knock-on effects than in cases of direct effects. Indirect collateral damage occurs on a causal chain containing more than two terms. Thus, direct damage occurs when there is just one causal step (e.g., a missile A causes destruction to buildings B around the place of impact). Meanwhile, indirect damage occurs where more than one causal step links the initial impact to the effect of the damage (e.g., missile A causes effect B in the surroundings which in turn causes effect C elsewhere or in the future).
The more causal steps that exist between the original impact and the final damage to be considered, the less it is probable that one can speak of foreseeable damage. Only foreseeable damage can be excessive under the provision we are considering. Finally, while not being the focus of this post, it may be worth mentioning that only death or injury to humans or destruction of objects is sufficient to implicate the proportionality rule for attack.
It is difficult to say much more about the subject matter on an abstract analytical level. To better grasp the implications of such prohibited or non-prohibited indirect causations of damage, the most promising path is to consider typologies of concrete situations and attempt to order them according to their relevant legal reasoning. The easiest way to do so quickly and simply is to approach the cases under the lens of easy cases (clear indirect excessive collateral damage, or clear indirect non-excessive collateral damage) and hard cases (doubtful cases).
This post will briefly analyze some cases which were in part made available to me through the IHL section of the Swiss Military High Command of which I am a member. To keep symmetry, I will discuss three cases under each category, which shall leave us with nine cases overall.
Easy Cases of Excessive, Indirect Collateral Damage
Let us start with cases in which there is clearly excessive collateral damage, albeit only of an indirect nature.
Case 1
A member of the technical staff of the ICRC reported to the Swiss High Command section some years ago the following situation occurring in Syria. A mission in the town of Aleppo reported that this town was set in a desert area where the water supply was limited. A supply station situated outside the town provided drinkable water to the town and was linked to it through aqueducts.
Some military objectives, near that station or its aqueducts would be attacked in the wake of the armed conflict taking place there. It stands to reason that the situation required the commander to take account not only of the direct damage to the station and its canalizations, but also the impact of their impairment on the water supply of the town.
The incidental damage in this case would be indirect. There would be three steps including: effects on the place of impact; effects on water installations around the place of impact; and effects in the town downstream. Moreover, the situation would be highly foreseeable. Even an untalented commander and his legal staff would easily be aware of the importance of these water installations for the supply of the town or at least would be able to obtain the relevant information.
Therefore, an attack without the necessary precautions causing a fully expected disruption of the water supply in the town would be characterized as excessive collateral damage under Article 51 AP I and related customary international law (also for NIACs). An attack whose effects can be expected to cause such effects would be unlawful under IHL.
Case 2
On 6 June 2023, presumably Russian forces attacked the Kakhovka dam in eastern Ukraine. A breach of the Nova Kakhovka dam caused flooding, displacing tens of thousands of civilians, with multiple villages and thousands of hectares of agricultural land submerged or flooded. The long-term impact was severe. The reservoir was a critical drinking and irrigation water source for the region (see here, here, and here).
There is no need to canvass at length this situation. The effects of the flooding are excessive direct collateral damage. The long-term effects are indirect collateral damage. This is the case in particular regarding the function of the dam as a water source (similarly to the installations outside Aleppo). This effect of the attack was fully foreseeable; the role of the dam as a water reservoir was information that could not be ignored by a reasonable attacking commander. And the effects in the longer run of its destruction were also plain. That such an attack yielded excessive direct and indirect collateral damage seems thus manifest. Indeed, it seems hardly possible that any tangible military advantage (stopping the advance of adverse forces?) could justify collateral damage of such an amplitude.
Case 3
In 2014, reports indicated that Israeli forces shelled the only local electricity plant in Gaza (see here). According to reports, the shutdown of the Gaza power plant had an important impact on the civilian population beyond the curbing of energy production. It drastically curtailed the pumping of water to households and the treatment of sewage, both of which require electric power. It also caused hospitals, already straining to handle the surge of war casualties, to increase their reliance on precarious and apparently insufficient generators. And it affected the food supply because the lack of power shut off refrigerators and forced bakeries to reduce their bread production.
These knock-on effects may seem disproportionate regarding the military advantage searched for. The assessment would probably be different if alternate power stations had existed and the attack would therefore not affect all the essential civilian functions. But in such a case, the military advantage pursued by the attack might also have become evanescent in the first place, electricity supply remaining fully available also to the military forces.
Similarly, in 1991, it was estimated that coalition attacks on Iraq’s electrical power system resulted in approximately 70,000 civilian deaths. The attacks reduced Iraq’s power capacity to 15% of its pre-conflict levels. The attack had a significant impact on health services (reduced hospital capacity, inability to refrigerate adequate quantities of vaccines) and sanitation (inability to treat and dispose of raw sewerage) (see Robinson & Nohle, p. 123).
However, by contrast with the preceding cases, at the time there was insufficient awareness of such long-term effects. They may not have been foreseeable. The lessons learned from that armed conflict have now brought into common military knowledge such reverberating effects.
Easy Cases Lacking Excessive, Indirect Damage or Foreseeability
We can now move to cases where there is, for some reason, clearly no excessive indirect collateral damage, or insufficient foreseeability.
Case 1
The case of the bombing of the radio television station (RTS) in Belgrade by NATO forces provides an example. It gave rise to the famous Banković case at the European Court of Human Rights. On 23 April 1999, just after 2:00 a.m., one of the RTS buildings at Takovska Street was hit by a missile launched from a NATO aircraft. Two of the four floors of the building collapsed, and the master control room was destroyed. Sixteen persons were killed and another sixteen were seriously injured in the bombing of the RTS. The station was repaired in a matter of some hours and the broadcasting could thus be resumed.
The foreseeability of that quick repair and the direct proportionality of causing an expected number of civilian casualties can be discussed (with regard to a military advantage to, apparently, curb some military messaging from this station, which, however, could be performed by some other broadcasting station). But this is not the issue here. The point is merely that of indirect collateral damage.
The main such indirect damage was the deprivation of civilian broadcasting during a certain number of hours, and thus also of the right to obtain information held by the populations. This damage is in all probability too limited to constitute “excessive” damage in the sense of Article 51(5)(b) of AP I of 1977. The lack of “excess” disposes of the matter in this context. Moreover, it would be correct to say that already in the first place this damage was below the article’s ratione materiae threshold: neither deprivation of information nor of broadcasts amounts to death or injury to persons or damage or destruction to an object. Under this lens, there is no relevant collateral damage at all.
Case 2
The following case is taken from an exercise. A commander purports to attack a target placed near an electricity generating station, which is crucial for essential services rendered to the civilian population. He takes the measures of precaution provided in Article 57 of 1977 AP I to minimize the expected collateral damage.
There remains, however, a foreseeable but remote probability that the electricity generating station is impaired, albeit in a quite limited way. The commander may then reasonably think that if this occurs, the services of the attacked State will be able to repair the light damage caused to that plant within a short time. What is not foreseeable to him is that due to exceptional shortage of spare parts, this repair cannot be undertaken within a reasonable timeframe and the collateral damage to civilians is therefore significantly greater than expected. The main point is here the lack of reasonable foreseeability. Thus, the damage is not “expected” in the sense of Article 51(5)(b) of AP I.
Case 3
The following case is again taken from an exercise. The bombing in a densely built area could damage a building in which a laboratory holds dangerous substances, such as viruses. If the blast of a bomb expanding horizontally damages that building, these substances may be released uncontrollably and provoke a chain of consequences, notably epidemic diseases.
The commander chooses the explosive ordnance in a way that the blast is limited; he also orders that the ordnance explode not on the ground, but in the air at a certain distance from the targeted building. This course is ordinarily chosen by military commanders to concentrate the main flow of the blast vertically from the top to the bottom, rather than horizontally from the place of impact to the surroundings. The risk of damaging bystanding buildings and facilities is thereby greatly reduced.
Having taken these precautions in line with the requirements of Article 57 of AP I of 1977, the probability of producing the knock-on collateral damage of releasing dangerous substances, which then in turn will cause diseases, seems too remote to substantiate prohibited incidental damage in the sense of Article 51(5)(b), of quoted AP I. The issue is here one of expected occurrence and thus of probability. Again, the damage is not “expected” in the sense of Article 51(5)(b) AP I. However, even a very remote risk of damage, because of its considerable gravity, could still lead the commander to renounce the attack for sound policy reasons. But this is a question which is spared here.
Hard(er) Cases
We now move to the hard cases (i.e., to the ones where there is reasonable doubt about the outcomes to be expected). Under IHL, if there is doubt, a commander is not obliged to adopt the worst-case scenario. He may accept those uncertainties and proceed to order the attack even if it is unclear whether excessive collateral damage can be expected. Only when excessive harm is expected can there be a breach of Article 51(5)(b) of AP I. Absent this condition, the proportionality rule does not apply.
But that is only half of the story. In Western armies for some years at least, commanders have considered collateral damage beyond the mere aspect of compatibility with IHL. The law may be quite permissive in that context, but morality and public expectations, as well as national rules of engagement, have figured equally if not more so. No commander of a Western State can afford to risk an attack where massive effects on civilians could even somewhat remotely be possible (in contrast to those which are unforeseeable and are flowing from bad luck).
Thus, in many such cases, Western commanders will be reluctant to give the order to attack, perhaps unless the military advantage expected to be reaped is really of paramount importance. There is thus a double layer here: what the law permits; and what the commander can politically afford. The same is not necessarily true for commanders of armies with less respect for IHL, that lack the check of a free press and for armies of autocratic régimes, such a Russia and perhaps for commanders of a State using more relaxed standards for targeting in this regard, such as Israel.
Case 1
On 16 March 2022, Russian military forces targeted the Mariupol theatre in Ukraine. The theatre, in the city’s Tsentralnyi district, was a significant hub for the distribution of medicine, food, and water. It also served as a designated gathering point for people hoping to be evacuated via humanitarian corridors. The building was made clearly recognizable as a civilian object (see here, here, and here).
The strike caused extensive collateral damage on the spot of the attack (direct collateral damage). But it also affected the role of the theatre as a hub for humanitarian aid (indirect collateral damage). The foreseeability criterion was likely fulfilled. If the attacking commanders correctly executed their duties of precautions in attack under Article 57 of AP I and related customary international law, it is likely they could not have ignored this function of the building. The storing and distribution of humanitarian aid supposes significant movements of humanitarian commodities in and out the building, a fact that cannot go unnoticed.
However, doubt arises in regard of an additional criterion. It is unclear to what extent this building was the only one in the town able to perform such functions. If it was possible to move the humanitarian hub function to another building, there was then no noticeable impact on the humanitarian aid function (apart from the fact that the relief materials present in the building at the time of bombardment were destroyed, which has some indirect collateral effects).
It is difficult to presume, a priori, that the hub function could not be moved to another building in the whole town. If that was true, the disruptive effect on future relief, with expected deaths or physical impairments (the core of the indirect collateral damage argument) is considerably limited. The assessment must here be made considering a series of facts which cannot be easily evaluated from the outside including: the size of the building; the facilities of storage within it; the connections with the main road axes; etc.
One can contrast this case with a hypothetical one where the move to other buildings is impossible. Suppose belligerent A attacks electricity generating station which is a military objective because it produces electric energy delivered among others to the army. But its electricity is also a vital asset for three huge hospitals (suppose their own batteries can overcome a blackout only for some hours).
Assume belligerent A destroys the electric station. It is common knowledge that a repair of the station that permits resumed production of electric energy will take weeks, if not more. It also stands to reason that all the persons cared for in the hospitals cannot be evacuated to other hospitals, either on the gravity of their medical condition, or because there are no places in other hospitals, or for other reasons. In such a case, the situation of the hospitals requires to be considered under the lens of relevant indirect collateral damage. There is no “replacement” possibility which could be envisaged as in the Mariupol case.
Case 2
Another example comes from a Russian attack on the Ukrainian satellite ViaSat in 2022, with the aim to cut military communications within the Ukrainian army (see Kazi, Kazi, & Bhosale, p. 59 et seq.). The attack destroyed satellite modems used to deliver Internet service to distributors in several European States, which in turn allowed civilians to connect their devices.
The attack therefore had indirect effects. For example, 9,000 French subscribers were cut off from the Internet network delivered by this satellite for two weeks. Quaere whether this indirect collateral damage is excessive. A specific element is that the effect is felt in third, neutral or non-belligerent, States. The doubt which arises here has to do with what is foreseeable or excessive. Considering the ratione materiae threshold of Article 51(5)(b), such damage not leading to death or physical injury, or the impairment of objects, could at first sight fall outside the domain of applicability of that provision. If, however, the attack had some impact on the Internet flow by impairing its hardware or its material functioning, there could be, arguably, a “damage to civilian objects”, as required by the provision. If that were the case, we would remain within the domain of Article 51 of AP I.
To what extent must a commander consider the possible number of cut-offs from a satellite system? Clearly, that depends on the importance of the satellite, however, normally individual satellites are among many others potentially that may take over functioning. Moreover, guessing the precise impact of a blackout of one system on the ubiquitous Internet systems is not a task which a commander is normally fully equipped to undertake. Thus, the situation remains doubtful.
Case 3
The following example is taken from an exercise. In an armed conflict, a belligerent attacks and destroys a military satellite (alternatively: a series of military satellites) of the adverse belligerent. There is no direct collateral damage; these satellites are not located near civilian objects or persons in outer space. Nor are the satellites performing dual-use civil functions. However, much spatial debris results. The debris remains in orbit and can in the future damage other civilian space objects of the adverse belligerent or of third States (indirect collateral damage).
This damage is foreseeable; it is known that free-floating debris can cause significant impairment of objects with which they impact. What remains unclear is the probability that such damage arises. If merely one military satellite is attacked, the result will not be to add a sufficiently significant amount of debris to the already existing ones, so the possible future collateral damage will hardly be excessive.
But the situation might look different if a greater number of such satellites are destroyed. There will be a threshold after which the excessive, indirect collateral damage will be at once foreseeable and sufficiently probable. When this exactly is the case will most likely be a matter of doubt. Contrary to case number 1, the issue is here not the possibility to replace, but rather the probability of causing damage and its gravity.
Conclusion
This short case analysis has shown that a series of contextual criteria must be mustered to reach a legally sound result when assessing indirect collateral damage for purposes of proportionality determinations. The commander must check various points, among which the four following stand out as being particularly important: (1) excess, as with direct collateral damages; (2) foreseeability, to a reasonable degree (past experiences and lessons learned also substantiate what is foreseeable); (3) possibility to replace an item in its functions to the benefit of the civilians, or possibility to repair it (which then interacts with the criterion of excess: the possible damage would then not be excessive); (4) the probability of the occurrence of indirect collateral damage (again, this interacts with the criterion of excess: the occurrence of the damage is not sufficiently certain, or its amplitude remains all too doubtful, there is then no “expected excessive damage”).
As usual, the law on targeting is based on a cluster of factual-legal considerations to be applied by the reasonable commander. For example, commanders must take account of the type of attack (defensive responses often allow for less time of preparation), the capacity of getting relevant information, the precise contextual-factual setting, the concrete possibility to take precautions or to substitute a target for another to obtain a similar military advantage, etc. Any exercise of this type must be performed in the full context of each case and a certain point of time.
Finally, access to relevant information by the commander and his legal advisors is a highly crucial element in the assessment of all collateral damage and often even more of indirect incidental damage (see here).
***
Robert Kolb is a Professor of Public International Law at the University of Geneva and former legal staff member of the ICRC. Prof Kolb is also a member of the legal section of the Swiss military high command (IHL section).
The views expressed are those of the author, and do not necessarily reflect the official position of the United States Military Academy, Department of the Army, or Department of Defense.
Articles of War is a forum for professionals to share opinions and cultivate ideas. Articles of War does not screen articles to fit a particular editorial agenda, nor endorse or advocate material that is published. Authorship does not indicate affiliation with Articles of War, the Lieber Institute, or the United States Military Academy West Point.
Photo credit: Teteria Sonna
