Use of Grad Rockets in Populated Areas: What Lessons from Gotovina?

Written by

Maya Brehm PhotoMaya Brehm is a researcher in weapons law at the Geneva Academy of International Humanitarian Law and Human Rights (ADH) and a consultant with Article 36 and PAX. Her recent work focuses on the humanitarian impact of explosive weapon use in populated areas and on framing the policy debate on autonomous weapons systems.

In a recently published report, Human Rights Watch (HRW) documents harm to civilians from the use of 122mm Grad rockets apparently fired by Ukrainian government forces and pro-government militias into Donetsk and its suburbs. In four attacks investigated by HRW at least 16 civilians were killed and many more wounded. According to HRW insurgent forces also recently used Grad rockets. The image below from HRW shows attacks in and around Donetsk (click to enlarge). The organization has also posted a video online presenting its findings.

 The problem with Grad rocketsGrad rockets

Grad rockets are unguided rockets fired from a multiple-barrel rocket launcher (MBRL) that can deliver up to 40 rockets within a very short time to a range of 20 kilometers. Like other unguided, indirect fire  weapons, Grad rockets are considered ‘area weapons’, suited for attacks against targets of significant dimensions, because due to ballistic and other factors, the area over which the rockets can spread out is relatively wide.

The dimension of the area affected by a rocket attack (the area of potential impact of the rockets combined with the blast/fragmentation zones of the individual rockets) is a function of many variables, including fuzing, ballistic and firing technique-related factors. As that area can be very wide, the use of Grad rockets in populated areas carries a high risk of harm to civilians.

An indiscriminate weapon?

The remarkable aspect of this HRW report is that it goes beyond the usual exhortations on parties to a conflict to fully comply with the rules of international humanitarian law (IHL), and calls on ‘all parties to the conflict in eastern Ukraine, particularly Ukrainian government forces, to stop using Grad rockets in or near populated areas because of the likelihood of killing and wounding civilians.’ HRW considers that

Grad rockets cannot be targeted with sufficient precision to differentiate military targets … from civilians and civilian structures…, which are immune from attack. As such, their use in populated areas violates the laws-of-war prohibition against indiscriminate attacks.

IHL experts will be quick to point out that Grad rockets are not specifically prohibited under treaty law. What is prohibited under customary IHL is the use of ‘weapons which are by nature indiscriminate’ (Rule 71, ICRC, Customary IHL study, 2005). There is, however, no agreement on what weapons concretely, if any, should be deemed ‘indiscriminate weapons’ whose use would be prohibited on this basis.

According to the ICRC, the two criteria most frequently cited to determine what is an indiscriminate weapon are ‘whether the weapon is capable of being targeted at a military objective and whether the effects of the weapon can be limited as required by international humanitarian law’. These criteria are drawn from the prohibition on indiscriminate attacks, which are described in Article 51(4) of 1977 Additional Protocol I to the Geneva Conventions.

The prohibition on indiscriminate attacks is, however, silent on what means of combat are deemed sufficiently dirigible or limited in their effects. In its customary IHL study, the ICRC notes that weapons cited in practice as being indiscriminate in certain or all contexts include Scud missiles and Katyusha rockets. The Grad, the most widely used multiple barrel rocket launcher in the world, is not mentioned.

Indeterminacy and case-by-case assessments

According to the dominant doctrinal view a weapon that can be used in compliance with IHL in some circumstances cannot be deemed indiscriminate per se (or ‘inherently’ indiscriminate). This view is premised on the proposition that we can distinguish between ‘inherent’ characteristics of a weapon that make it impermissibly difficult to direct, on the one hand, and characteristics pertaining to a weapon’s use, on the other hand.

But because the legal assessment of whether a given weapon can be directed as required by IHL is treated as context-dependent it has proven difficult in practice to agree that specific weapons are illegal per se. Proponents of a weapon will usually be able to come up with a scenario in which a weapon’s use would likely not result in an indiscriminate attack (e.g. an attack on a large military objective in an area devoid of civilians).

In the absence of a weapon-specific treaty that resolves the indeterminacy of the general rules, therefore, we are unable to determine the legality of a weapon. This, of course, ‘is tantamount to saying that States consider the general principle on indiscriminate weapons to have become valueless’. (A. Cassese, ‘The Prohibition of Indiscriminate Means of Warfare’, in A. Cassese, P. Gaeta and S. Zappalà (Eds.), The Human Dimension of International Law: Selected Papers of Antonio Cassese, 2008.)

What we are left with are case-by-case (attack-by-attack) assessments, which are not ideally suited to draw more categorical lines. The Gotovina case (IT-06-90) before the ICTYa ruling that was meant to shape the law pertaining to artillery use against targets in populated areas, grappled with these issues and has left a dangerous legacy.

An attempt at drawing a line

Gotovina, a Croatian commander, was prosecuted before the ICTY, inter alia, for crimes committed in connection with the shelling, among other weapons with 122mm Grad rockets, of several towns, including Knin. (For detailed discussions, see the articles by Clark, Huffmann, and Vallentgoed.)

The Trial Chamber, considering a variety of factors, determined that although ‘MBRLs are generally less accurate than Howitzers or mortars, their use by the HV [Croatian army] in respect of Knin on 4 and 5 August 1995 was not inherently indiscriminate’. (§1897, Gotovina et al., ICTY Trial Chamber Judgment, 15 April 2011.)

As it was contested what had actually been targeted in all cases (there were indications of ‘area bombardment’), the Trial Chamber decided, with respect to a sub-set of all projectile impact sites under consideration, that projectiles that impacted within a distance of 200 meters of an identified artillery target, were presumed to have been fired at that target, whereas impact sites further away were treated as indicators of an unlawful attack on civilians. (§1898)

This approach was rejected on appeal, which caused the rest of the case to collapse and led to the reversal of the convictions.

Gotovina’s legacy: the end of indiscriminate attacks?

The Appeals Chamber dismissed the 200 meter ‘margin of error’ among other reasons on the ground that firing a weapon from a greater distance could well have resulted in shells and rockets impacting more than 200 meters away from a target. The wide spread of impact sites could then be plausibly explained by a higher margin of error, and it could not be excluded that the shells were all aimed at legitimate military targets. (§§60-65, Gotovina et al., ICTY, Appeals Chamber Judgment, 16 November 2012.)

This line of argument is also evidenced in expert reports submitted by the Defence on appeal. Lieutenant General (ret.) Shoffner, for example, rejected the 200 meter standard on the basis that under the given firing conditions, more than 50% of the rockets could be expected to fall more than 300 meters from the aim point. (Appellant Ante Gotovina’s motion to admit new evidence pursuant to Rule 115 (Public redacted version), 4 November 2011, Exhibit 21, 3540.)

That the Majority of the Appeals Chamber subscribed to this logic bodes ill for the protection of civilians in populated areas from the effects of explosive weapons. As Judge Agius pointed out in his dissenting opinion, the approach makes it impossible to classify any attack as indiscriminate in the absence of an established margin of error. (§21, Dissenting Opinion of Judge Agius, Gotovina et al., ICTY, Appeals Chamber Judgment, 16 November 2012.)

The Trial Chamber had clearly realized this. In a footnote, it noted that

had these impacts which were at a distance of up to 700 metres from artillery targets been the result of the inaccuracy of the artillery weapons used, that would require a further consideration of whether such inaccurate weaponry can be used in the context of an artillery attack on specific targets within a town. (Gotovina et al., ICTY Trial Chamber Judgment, 15 April 2011, footnote 932.)

The Appeals Chamber, however, did not give further consideration to this matter, and it has been criticized for its ‘apparent disregard for the consequences’ of its ruling.

Toward a presumption against wide-area effects explosive weapons in populated areas

Can Grad rockets impacting in residential areas of Ukrainian towns now simply be justified by reference to the relative inaccuracy of the rocket system, and the large distance they were fired from? I certainly hope not.

In the Gotovina case, several artillery experts invited by the Prosecution to comment on the Defence reports did not think that the use of Grad rockets was appropriate in populated areas (even though they also rejected the 200 meter standard as too simplistic). Lieutenant General (ret.) Applegate, for example, considered that the use of MRL systems in built-up areas was surprising in view of the wide dispersal of their rockets. He pointed out ‘that the MLRS rocket system – a superior system to that used by the HV … was not deployed to support UN forces in Bosnia because of its inherently larger spread of fire and lack of discrimination when engaging targets in built-up areas occupied by non-combatants.’ (Notice of Filing of Public Redacted Version of Prosecution Response to Gotovina’s Second Rule 115 Motion, and Request for Change in Status of Corrigendum, Public Annex I, 5678.)

The Gotovina case illustrates that artillery experts, even if broadly in agreement on questions of targeting doctrine and weaponeering, can come to opposite conclusions about the appropriateness and acceptability of using Grad rockets and other explosive weapons in a populated area. The answer to what is permissible is not going to be settled on the basis of ‘error margins’ alone.

The law provides the corner posts: to protect civilians from harm, attackers must ensure that their weapons can be and are directed at a specific military objective and that their effects are limited. Rather than deferring to indeterminate legal rules or determinations on an attack-by-attack basis, a more categorical determination of what explosive weapons are acceptable and ultimately permissible for use in a populated area can and must be made.

Grad rockets, typically used for their capacity to affect a wide area at great range, have been shown in a number of conflicts to result in unacceptably high levels of harm to civilians when fired into a populated area. (See, e.g. the Report of the Independent International Fact Finding Mission in Georgia, Vol. II, 2009, 338-343; AOAV, Syria’s Dirty Dozen: The Grad, 2013.) A categorical presumption against the use, in populated areas, of Grad rockets and other explosive weapons with wide area effects (due to low accuracy of delivery, high dispersal and/or large blast/fragmentation radius) can help prevent and reduce harm to civilians.

Print Friendly, PDF & Email


Leave a Comment

Comments for this post are closed


Luka Misetic says

July 30, 2014

Having served as Defence Counsel for Ante Gotovina, I read this blog post with great interest. The author discusses the use of GRAD rockets and appears to call for a general ban on their use in urban areas, but fails to discuss Article 51(4) of the Protocol Additional to the Geneva Conventions. I believe that if the Gotovina Appeals Judgment is analyzed in the context of the text of the law (i.e. Article 51(4)), it becomes clear that the Appeals Chamber got it right.

Moreover, the author uses Gotovina as the case around which to build her case that a "categorical presumption" against the use of GRAD rockets in civilian areas should be implemented in order to "prevent and reduce harm to civilians." However, in the Gotovina case, the Prosecution was not able to identify a single victim of shelling anywhere (no civilian was identified who was killed or injured by a single shell or rocket used during the operation).

In light of this absence of civilian casualties, one wonders whether the Gotovina precedent is really the right case to use in support of the call for a "categorical presumption" against GRAD rockets.

Maya says

August 4, 2014

Dear Luka Misetic

Thank you very much for your comment. I think your characterisation of my argument is accurate: Rather than deferring determination of permissibility to attack-by-attack assessments under art. 51(4) API I would call for a more categorical orientation.

There is nothing revolutionary about this proposition. I see no reason why we should not give effective meaning to the prohibition on weapons that are by nature indiscriminate. The ICTY’s jurisprudence demonstrates that this can be done: it found it surprisingly easy to make such a categorical determination in the Martic case, for example.

As you know, that case dealt with an attack on another populated area, Zagreb, with M-87 Orkan rockets equipped with cluster warheads and launched from another multiple barrel rocket launcher (MBRL). Neither at the time of the attack nor at the time of the Trial judgment were rockets, MBRLs, the Orkan rocket system or cluster munitions prohibited under int’l law. The Trial Chamber considered expert testimony on accuracy/dispersion and blast/fragmentation zones, found that Martic knew what effects such weapons would have in a populated area and, in the absence of an ‘established error margin’, was able to determine that the Orkan was ‘an indiscriminate weapon’, incapable of hitting specific targets.(§463) The Appeals Chamber upheld this finding. (§255)

Today, cluster munitions are banned under the 2008 Convention on Cluster Munitions, among other reasons, due to their ‘indiscriminate area effects’ (Art. 2(2)(c)). A number of states have already pointed out that cluster munitions are not that different in their effects from other explosive ‘area weapons’. I would agree.

The case in favour of a presumption against the use, in populated areas, of explosive weapons wide area effects rests, of course, on far more than the Gotovina case. Data collected by different actors over several years and across different regions shows that such use results in a pattern of civilian harm that is unacceptable from a humanitarian standpoint. (See e.g.

As to your last remark, let me borrow Judge Orie’s words: ‘If I'm shelling…in an area full of civilian population, and if for whatever reason they all had been down in their cellars… and only one stayed upstairs… would that allow for a final judgement on whether it was appropriate or not when we have only one victim? That's not the test, would it be?’ (Transcript of hearing 19 January 2009,

Luka Misetic says

August 7, 2014

Dear Ms. Brehm:

I will disagree with your post on the following basis, at least as it relates to the Gotovina case:

1. Your initial post makes clear that you concede Grad rockets are currently not considered an inherently indiscriminate weapon under international law (“Grad rockets are not specifically prohibited under treaty law.” … “The Gotovina case illustrates that artillery experts, even if broadly in agreement on questions of targeting doctrine and weaponeering, can come to opposite conclusions about the appropriateness and acceptability of using Grad rockets and other explosive weapons in a populated area.”)

Grad rockets clear are “capable of being directed at specific military objectives,” and therefore are not per se indiscriminate weapons for purposes of Article 51(4)(b) of the Additional Protocol. Accordingly, the use of these rockets in Operation Storm was not a per se violation of international law in 1995.

2. I don’t agree with your interpretation of the Martic decision. The Appeals Chamber was very careful to find that the rockets were “incapable of hitting specific targets in the circumstances of the case as presented before the Trial Chamber. “ (Paragraph 252 of the Appeals Chamber Judgement, emphasis added). Accordingly, the Appeals Chamber was careful to emphasize that the holding was limited to the facts of that case and did not have broader implications generally for the use of Grad rockets in urban areas.

Moreover, unlike in Gotovina, the Prosecution introduced evidence that the dispersion error in the facts of the Martic case was 1km in each direction from any potential target, and the Trial Chamber relied on this evidence for its conclusion that the use of such a weapon with such a wide dispersion was inherently indiscriminate (Appeals Chamber Judgement, paragraph 250). I am therefore unsure why you claim that there was an “absence of an ‘established margin of error’” in the Martic case. It seems to me that the opposite is true in Martic: because of an established 1 km dispersion in either direction, the Trial Chamber concluded that the weapon was inherently indiscriminate in the specific context of the Martic case.

The Prosecution in Gotovina submitted no evidence of margins of error.

3. I disagree with your assertion that the Majority's opinion in Gotovina “tmakes it impossible to classify any attack as indiscriminate in the absence of an established margin of error.” The Majority said no such thing. Rather, the Majority was rejecting the Trial Chamber’s methodology of determining the lawfulness of the intent of the attackers by measuring the distance of impact locations from known military objectives by use of an artificial 200m standard that had no basis in evidence (unlike the evidence adduced in Martic, for example). The Appeals Chamber did not state that evidence of unlawful intent could not be adduced through other means, even in the absence of a known margin of error.

Moreover, the Trial and Appeals Chambers were focused on the issue of the intent of Gotovina in launching the artillery attack, rather than on whether the weapons themselves were inherently indiscriminate. The facts are that the Trial Chamber itself found that at least 795 out of 819 known artillery shells fired, were fired at pre-planned legitimate military objectives. (See Annex A here: -06-90-A/BRF286R0000349392.pdf). This amounts to 97% found to have been fired at legitimate military objectives, and not civilians or civilian objects. By necessity, the Trial Chamber implicitly found that these 97% of shells must have landed within 200m of military objectives, because the Trial Chamber’s own 200m standard stands for the proposition that if the Croatian Army fired at a military objective, there could be no explanation that would allow for such a shell to fall more than 200m away from that military objective.

In summary, the entire appeal was centered on the issue of whether Gotovina could be found to have intentionally targeted civilians with indiscriminate attack because 3% of shells fell beyond 200m of the intended targets. That was the Trial Chamber’s finding, and the Appeals Chamber Majority quite correctly found that this 3% could have been the result of a margin of error beyond 200m, and not because the only inference to be drawn from the 3% is that Gotovina was intending to target civilians.

4. Finally, on your citation to Judge Orie’s comment from the trial. Let me state what I have already stated publicly: Judge Alphons Orie has been reversed in every single Judgement he has delivered at the ICTY since 2003, in whole or in part (except for cases where the Accused plead guilty and there was no trial, or in contempt cases). Accordingly, it is safe to say that Mr. Orie has the worst judicial record of any judge in the history of international criminal law. I therefore do not find his comments to be authoritative.

Furthermore, Article 85(3) of the Additional Protocol in fact does make “causing death or serious injury to body or health” an element of the crimes of indiscriminate attack, direct attack against civilians, disproportionate attack, etc. when charged as a Grave Breach. It is true that the ICTY has eliminated the need for proof of death or injury when the crime is charged as a crime against humanity rather than as a Grave Breach. I find this proposition dubious in that the graver crime (crime against humanity) requires less than the “lesser” crime (grave breach), but admittedly this is the ICTY law at present.

Nevertheless, my point was not that in the Gotovina case, the lack of death or injury means that an element of the crime was not proven. Rather, my point was that as a matter of evidence, the lack of death or injury must be a strong consideration when assessing the overall intent of those launching the attacks. For this, I cite you to the Kordic Appeals Chamber Judgement:

“The crime of unlawful attack on civilians does not require that a specific number of civilians be killed or seriously injured. However, in a circumstantial case such as the present one, the scale of civilian casualties may be relevant to determine whether an attack is directed at civilians.” (Kordic Appeals Judgement, paragraph 446).

Judge Orie’s failure to consider the “scale of civilian casualties” – as he should have done in light of the Kordic Appeals Chamber Judgement – is yet another reason why the Trial Chamber’s Judgement in Gotovina was so flawed.

Finally, let me note that I am not expressing an opinion as to whether the international community should or should not institute a ban on the use of GRAD rockets going forward. My only point is that in the circumstances of Operation Storm in 1995, the use of artillery was consistent with the then-applicable laws of war.