31-03-2000 Article, International Review of the Red Cross, No. 837, by A. P. V. Rogers
A. P. V. Rogers is a Fellow of the Lauterpacht Research Centre for International Law, University of Cambridge, and a former Director of Army Legal Services, United Kingdom.
There can be no doubt that the political and military planners of the NATO air campaign in the Kosovo conflict of 1999 were determined that it should be carried out strictly in accordance with the law of armed conflict.
“The targets were exclusively military -- every effort was made to avoid collateral damage -- planes only fire at targets when we are confident that we can strike accurately -- some aircraft in the first operation returned without dropping ordnance. Targets are carefully selected and continuously assessed to avoid collateral damage.” [1 ]
“Others, and I have said continually this week that our aim is to hit the target and not to cause collateral damage to any surrounding areas. You have seen the effects of the bombs that we have dropped and the missiles that we have launched. We need to put those on the target, so if someone is in the cockpit and he has got to see the target all the way down to the ground he has to make sure that he can see that target and if he doesn't he is professional enough to not hit that target and bring his bombs back.” [2 ]
“Rules of engagement and concerns about collateral damage placed on us by our political masters are, rightly, very tautly drawn. The discipline of our people is such that they will not drop their weapons unless they have clearly identified t heir assigned target and know precisely what they're doing.” [3 ]
These are sentiments that experts in the law of armed conflict would heartily endorse.
Despite problems caused by cloud cover, care was taken to ensure that only military objectives were hit. There were, inevitably, some instances of collateral damage [4 ] . But, by and large, things seemed to go well until an attack on a vehicle column by NATO aircraft operating at 15,000 feet raised an important question about the law of armed conflict: what is the precise legal obligation on an attacker to identify the target as a military objective before he launches his attack?
Until the development of air power in this century, wars tended to be fought by opposing armies or navies in a series of short but bloody battles. While civilians suffered from the effects of armies on the march and from evacuation, billeting and requisitioning, they were rarely directly affected by the fighting unless they were inhabitants of a besieged town [5 ] During the static trench warfare of the First World War, they were usually evacuated from the front line area and out of range of artillery fire.
Air power enabled the belligerents to strike well behind enemy lines, at the lines of communication and logistics that kept enemy armies supplied. By cutting them off from reinforcements, fuel, food and ammunition, their ability to fight was severely curtailed. However, this meant striking at areas that were likely to be populated by civilians, thus increasing the risk of civilian casualties. In the First World War military casualties were still extremely heavy, since the tank was in its infancy and only sheer weight of numbers -- allowing for such losses -- could win through against the hail of enemy shrapnel and machine-gun bullets; civilian casualti es were relatively light.6 This was about to change. The senior commanders of the war, of course, deeply regretted the death, wounds and sickness of their men, but the sense of military duty to defeat the enemy at any cost seemed to prevail. The more junior commanders, who would rise to high rank in the next war, were determined to ensure that in future military casualties would be reduced.
Things were to change in other ways too. The heavy demands of the large but mobile conscript armies that were put into the field in the Second World War and their increasing reliance on mechanization meant that civilians had to be employed in factories producing weapons, warships and military aircraft, their armaments and components, and in installations producing the fuel to drive vehicles and the raw materials such as steel needed to build ships. The emphasis of targeting was shifting away from enemy combatants to the equipment and supplies on which they depended, but at great cost to the enemy's civilian population. Though still protected by the law of armed conflict from direct attack, they were not protected from incidental damage caused as a by-product of attacks on war production facilities. Bombing was still far from precise.
Guerrilla warfare, in which fighters merge with the civilian population, tends to increase civilian casualties. It has taken its sad toll in many post-1945 conflicts. Guerrillas prefer to launch attacks out of civilian anonymity at the enemy's weak points, often using tactics such as ambush very successfully. Finding it hard to identify and grapple with the guerrillas, their adversaries, the regular forces, are inclined to overreact with their countermeasures or with heavier weapons, in either case to the detriment of the civilian population. The situation is exacerbated when guerrillas, partisans, freedom fighters or other armed factions are engaged in combat in towns or populated areas. Regular forces are unable to deploy th eir armoured vehicles to best advantage and their opponents, from the cover of civilian buildings, can deploy light weapons and anti-tank grenades very effectively at short range. Such fighting entails severe casualties, not only for the troops involved but also for the civilian inhabitants who get caught up in the hostilities. In circumstances like that, the Russian proposal to set up safe corridors to allow the civilian population to escape the fighting in Grozny is in keeping with the principles of the law of armed conflict. Of course, those corridors must genuinely be safe and, preferably, monitored by international organizations such as the ICRC or OSCE.
The understandable modern attitude of military commanders that they will not expose their subordinates to unnecessary risks tends to increase the danger to civilians. This attitude is not limited to the West. After the casualties suffered by the Russian army in its attempts to invade the Chechen capital in January 1995, it tried new tactics in its campaign of 1999. Marshal Igor Sergeyev, the Russian Defence Minister, was reported as saying that Russia was relying on the massive use of firepower to keep down its casualties and that artillery and air force was being used to “destroy the Chechen gangs and their camps”. It was estimated that Russia had deployed 100,000 troops against President Maskhadov who had 20,000 to 30,000 experienced guerrillas with light weapons. [7 ]
There has also been growing pressure, especially in the West, to reduce military casualties to a minimum, otherwise public support for the campaign is likely to fall away. Indeed, there is almost an expectation now of few, if any, casualties, especially if the country is not directly threatened by military operations. The British task force suffered 256 military battle deaths and 777 military wounded out of a mobilized military strength of 9,500 in its campaign to retake the Falkland Islands in 1982 [8 ] This was tolerated by the British public, perhaps because it was thought that British interests were at stake. But there would have been lower tolerance in the Gulf War of 1991 or in the Kosovo conflict of 1999, since neither directly affected the British people. In the event, the casualties were slight. There were 17 British military battle deaths in the Gulf [9 ] ; there has been none yet in Kosovo. [10 ]
The result of these trends has been a steady increase in the ratio of civilian to military casualties in armed conflicts despite the efforts of diplomats and international lawyers to emphasize the principle of civilian immunity [11 ] . There has been growing public awareness, possibly since the Vietnam War, of these civilian casualties. Media reports, often broadcast at the same time as events unfold in a campaign, have a very immediate impact and influence on the public. This is known as “the CNN factor”. The public are shocked by images of women and children killed or injured in war and are likely to have a low tolerance of such casualties, especially, again, if they perceive that national interests are not at stake. The situation is probably exacerbated by their unduly high expectations of the accuracy of smart weapons. As the NATO air campaign in Kosovo went on and mistakes occurred, there was an almost tangible sense of public discomfort. This could have led to pressure to call off the air strikes.
Requirements concerning methods and means of warfare
So the political and military imperative is for weapons and tactics that, so far as possible, prevent own casualties, hit military
objectives accurately and reduce incidental death, injury and damage to an absolute minimum [12 ] . It has been said that:
“The next key requirement [for air power in a new era ] is to increase stand-off capability for weapons which have pinpoint delivery accuracy to achieve maximum strategic e ffect, with minimum collateral damage and minimum risk to the weapon carrier and launcher.” [13 ]
Smart weapons, a universal panacea?
In the Gulf War of 1991, Royal Air Force (RAF) Tornado aircraft were used initially to attack Iraqi airfields by dropping JP233 bombs from a low altitude above the target. Low-altitude attacks were meant to defeat enemy radar and increase the element of surprise. But the RAF suffered the loss of five aircraft in seven days and so aircraft were switched to dropping 1,000lb iron bombs from 20,000 feet where they were out of range of anti-aircraft guns but from where the bombing was less accurate [14 ] . In response to demands for smart technology, the RAF used Buccaneer aircraft fitted with laser target designators to pinpoint targets, along with Tornados carrying laser-guided weapons in order to minimize collateral damage. The thermal imaging and laser designation pod (TIALD) was rushed into service. These changes dramatically improved targeting accuracy. [15 ]
Some argue that stand-off weapons remove any element of humanity from warfare, which is reduced to a cold and ruthless activity carried out with clinical precision. The further from the target an attack is launched, the less the attacker is aware of the human cost involved because he does not see the effects of the attack. It is sometimes said that this results in a dehumanization of warfare. It may be thought that because personnel launching stand-off weapons are attacking inanimate objects, such as airfields or bridges, but do not know of the destruction and misery that is being caused on the ground, they are not so aware of the problem of collateral damage [16 ] . This is not always true in the case of smart technology. The laser designator system used by the Buccaneers had a built-in video camera that recorded what appeared on the navigator's screen as he guided the laser designator on to the target. As a result, air crews and ground staff could see exactly what was happening. [17 ]
Media reports of allied bombing during the Gulf War and Kosovo air campaigns gave the impression that the weapons deployed were mainly smart weapons and that these weapons were unerringly accurate. Of course, that is far from the truth. Smart weapons are capable of delivery with considerable accuracy, but that capability is only one of many factors that affect accurate targeting. The effects of air defence measures, which can deflect missiles from their course; the strains under which air crew may be operating at the time of weapons release, such as avoiding surface-to-air missiles; the quality of intelligence about the identity of the proposed target [18 ] ; the technical problems of identifying and hitting targets on the ground from the air; the possible failures of guidance systems; the limitations of target recognition techniques [19 ] ; and ordinary human fallibility can all affect the outcome of an attack. Guidance and reconnaissance systems are usually affected by cloud, rain or smoke [20 ] . Furthermore, even if the target is accurately hit, that does not mean that there will be no collateral damage. [21 ]
In view of the cost of smart weapons [22 ] , they tend to be used sparingly. Only about 9% of the total tonnage of air delivered weapons during the Gulf War consisted of precision munitions and these were delivered mainly at strategic or operational level targets [23 ] .
Autonomous weapons [24 ] , such as cruise missiles, may seem to be the perfect answer to the problem of accurate targeting, but they generally provide no imagery of the attack and they also suffer from a serious drawback: they cannot adjust to last-minute changes of circumstances. Even against static targets that can be extremely problematic, for instance when a civilian refugee column crosses a bridge just as the autonomous weapon strikes. This is a scenario tha t can be avoided where the weapon is being guided by line of sight to the target. Against mobile targets, such as enemy troops on the ground, autonomous weapons are not easy to deploy effectively. Older systems like Paveway are more flexible in these situations. [25 ]
It seems that smart weapons are not as yet the solution to the problem of incidental damage nor are they as good as the public may have been led to believe. They are certainly more capable of accurate delivery than dumb bombs, but there are various other factors affecting possible incidental damage, not least the “circular error probable” which means that this risk, while much reduced, is not eliminated, especially in populated areas. Autonomous weapons bring with them a degree of inflexibility that means that the human element in guiding weapons on to the target will remain necessary in many cases in the foreseeable future.
The Incident of 14 April 1999
In the Kosovo conflict of 1999, pilots minimized danger to themselves by attacking from an altitude of 15,000 feet, where they were out of range of most hand-held surface-to-air missiles and anti-aircraft artillery [26 ] , until the Yugoslav air defences had been degraded to such an extent that they could safely fly at lower altitudes. [27 ]
On 14 April 1999, NATO pilots, flying at 15,000 feet and apparently looking for opportunity targets in Kosovo, attacked what they thought was the lead vehicle of a column of military vehicles. The following day there were numerous media reports that a refugee column had been attacked near Djakovica. Pictures on BBC World [28 ] showed scenes of carnage near the village of Meja. According to Serb reports, 64 civilians were killed, including three Serb policemen who were escorting them. The Serb-run Media Centre in Pristina claimed that a second, smaller column was hit on the roa d from Prizren to Djakovica in which six people were killed and 11 wounded. On 15 April, the NATO spokesman, Jamie Shea, expressed deep regret at the daily press briefing for loss of life to civilians from the attack the previous day on a convoy travelling between Prizren and Djakovica. [29 ]
There was some confusion about the incident in the days that followed. Survivors referred to a low-level attack by a MiG aircraft and reporters talked about machine gun or cannon fire. Robert Fisk, one of NATO's fiercest critics, visited three reported strike locations on the Prizren-Dakovica road: at Velika Krusa, Gradis and Terezifki [30 ] , he collected munitions parts and noted serial numbers. At Gradis, he said, there was evidence of strafing, similar to A10 strikes in the Gulf War, as well as bombing [31 ] . At the NATO press briefings reporters tried to ascertain precisely which attack Jamie Shea had been referring to, as there were at least two attacks on convoys in the area. A tape was played describing an attack on a column of three vehicles near Djakovica, but it was later stated that this did not relate to a refugee convoy.
General Marani answered press questions about attacks from an altitude of 15,000 feet. He said that target identification was more complex and was carried out correlating more information, not only what the pilot could see but other information he had before leaving and before arriving in the area, which he received from other aircraft, from other weapons systems not necessarily on board his aircraft. In other words, it was, he said, very much a team effort and not a case of a single pilot acting on his own initiative.
A full explanation was given on 19 April by Brigadier-General Dan Leaf [32 ] . He described two incidents on 14 April, both involving the use of 500lb laser-guided bombs. The first was north-west of Djakovica and involved an attack at 11:10 GMT on a vehicle whose crew were considered to be r esponsible for the burning of houses. This was followed at 11:48 GMT by an attack on a group of vehicles in a courtyard. That produced a secondary explosion consistent with the igniting of a gasoline store. The second incident involved a large convoy of more than 100 vehicles on a major road south-east of Djakovica. The first 20 vehicles were uniform in shape and colour and they maintained a set pace and spacing, indicative of military movement. The possibility of IDPs (internally displaced persons) being in the convoy was discussed but intelligence material indicated that this was a military convoy. The lead vehicles of the convoy were attacked by F-16 and Jaguar aircraft, starting at 12:19 GMT. Not all the attacks were successful. Some anti-aircraft fire was encountered. Doubts began to assail the military planners because it was unusual for the Serbian forces to travel in such large convoys, so A-10 aircraft, which are slower and more stable than F-16s, were called in to check and further attacks on the convoy were suspended at 13:00 GMT. On reports that the convoy consisted of both military and civilian vehicles, all attacks were cancelled and aircraft withdrawn at 13:20 GMT.
What are the legal requirements?
For States party to Geneva Protocol I of 1977 [33 ] , the duty, before carrying out an attack, is to “do everything feasible to verify that the objectives to be attacked are (…) military objectives” [34 ] . Feasible precautions may be defined as those which are “practicable or practically possible, taking into account all circumstances ruling at the time, including humanitarian and military considerations” [35 ] .
Humanitarian considerations would require a pilot to get close to the target to identify it properly; military considerations would require the pilot to fly at a safe height to be at reduced risk from anti-aircraft fire. This is a conflict that cannot be resolved easily.
Also relevant here is the rule of doubt. In case of doubt whether a person is a civilian or whether an object is normally dedicated to civilian purposes, the Protocol lays down a presumption of civilian status. [36 ]
For States not party to the Protocol, customary law nevertheless requires their forces to attack only military objectives and that means distinguishing them from civilian objects. The precise standard of care is not clear. It is certainly not higher than the “do everything feasible” standard imposed by Protocol I. At the very least, customary law would require those responsible for attacks not to attack persons or objects which they know or believe to be civilian.
Responsibilities in attack [37 ]
Whether or not a State is party to the Protocol, its armed forces are required to respect the customary rule of proportionality which attempts to balance military and humanitarian considerations. In applying this rule, the military decision-maker has to consider various factors when deciding what weapon or tactics to use, the object being to neutralize the military target with the least possible incidental damage or loss:
a. the importance of the target and urgency of the situation;
b. intelligence about the proposed target, i.e., what it is being, or will be, used for and when;
c. what weapons are available, their range, accuracy and radius of effect;
d. conditions affecting accuracy of targeting such as terrain, weather, night or day;
e. factors affecting incidental loss or damage, such as the proximity of civilians or civilian objects in the vicinity of the target or other protected objects or zone and whether they are inhabited, or the possible release of hazar dous substances as a result of the attack;
f. the risks to his own troops posed by the various options open to him. [38 ]
Smart weaponry, if available, has increased the options open to the attacker. From a legal point of view, he needs not only to assess what feasible precautions can be taken to minimize incidental loss, but also to make a comparison between different tactics or weapons so as to be able to choose the least damaging course of action compatible with military success.
Application of the principle of proportionality is more easily stated than applied in practice, as by adopting a method of attack that would reduce incidental damage, the risk to the attacking troops may be increased. The law is not clear as to the degree of care required of the attacker and the degree of risk that he must be prepared to take. [39 ]
As the author has written previously: “… Military necessity cannot always override humanity. In taking care to protect civilians, soldiers must accept some element of risk to themselves. The rule [of proportionality ] is unclear as to what degree of care is required of a soldier and what degree of risk he must take. Everything depends on the target, the urgency of the moment, the available technology and so on.” [40 ]
This approach is reflected in British defence doctrine:
“ Targeting . A key issue for commanders and planners is deciding what constitutes a legitimate target and how it may be attacked. This revolves around the principles of distinction and proportionality. Attacks should be limited to combatants and other military objectives. The civilian population and civilian objects must not be deliberately targeted; the morale of an enemy's civilian population is not a legitimate target and attacks designed to spread terror among the civilian population are expressly prohibited. Even military objectives should not be targeted if an attack is likely to cause (collateral) civilian casualties or damage which would be excessive in relation to the direct military advantage which the attack is expected to produce. The law stipulates that the military worth of the target needs to be considered in relation to the circumstances at the time. Therefore, a commander needs to have an up-to-date assessment of the significance of a target and the value of attacking it. If there is a choice of weapons or methods of attack available, a commander should select those which are most likely to avoid, or at least minimize, incidental civilian casualties or damage. However, he is entitled to take into account factors such as his stocks of different weapons and likely future demands, the timeliness of attack and risks to his own forces. Nevertheless, there may be occasions when a commander will have to accept a higher level of risk to his own forces in order to avoid or reduce collateral damage to the enemy's civil population.” [41 ]
Responsibilities in defence
It is not only the attackers who have a responsibility for protecting civilians from the effects of armed conflict. This is a general responsibility which falls not only on the armed forces but also on the civilian administration and political leadership. For States party to Protocol I this specifically requires, so far as possible, the removal of civilians from the vicinity of military objectives, avoiding locating military objectives in or near densely populated areas as well as taking steps like the provision of shelters to protect civilians [42 ] . Unfortunately, there is evidence to suggest that some States do not merely fail to comply with these requirements of international law, but deliberately put military assets close to protected objects or place civilians in military locations with the intention of protecting military objectives from attack or even in the hope of attracting international condemnation of the enemy if civilians are killed or civilian objects destroyed. Those carrying out attacks in such circumstances are not relieved of their obligation to attack military objectives only and reduce incidental damage as much as possible, but in considering the rule of proportionality, any tribunal dealing with the matter would be obliged to weigh in the balance in favour of the attackers any such illegal activity by the defenders. [43 ]
Applying the legal requirements in practice
It seems, at least in some cases of aerial attacks against targets on the ground, that the question of the risk to the attacker is not such an important consideration in practice.
If the target is sufficiently important, higher commanders may be prepared to accept a greater degree of risk to the aircraft crew to ensure that the target is properly identified and accurately attacked. No-risk warfare is unheard of. Risks may be taken, for example, to rescue pilots who have been shot down or in deploying forces on reconnaissance or target identification missions in enemy-held territory.
However, if the target is assessed as not being worth that risk and a minimum operational altitude is set for their protection, the aircrew involved in the operation will have to make their own assessment of the risks involved in verifying and attacking the assigned target. If their assessment is that (a) the risk to them of getting close enough to the target to identify it properly is too high, (b) that there is a real danger of incidental death, injury or damage to civilians or civilian objects because of lack of verification of the target, and (c) they or frie ndly forces are not in immediate danger if the attack is not carried out, there is no need for them to put themselves at risk to verify the target. Quite simply, the attack should not be carried out.
Rules of engagement could make this clear with a test like: “Are you sure that the target is a military objective? If you are in any doubt, would you or friendly forces be placed in danger if the attack were not carried out? If not, the attack is NOT to be carried out.”
Of course, cases will occur when members of the armed forces will not be able to make such cool calculations reminiscent of a jury in a criminal trial. In the heat of the moment, they may release their weapons even if not sure that the target is a military objective and even if not in danger themselves. Any tribunal considering the matter subsequently would be concerned with the proportionality principle. They would be assessing the importance of the target in relation to the incidental damage expected. However, the risk to attacking forces is an important factor to be taken into consideration when applying the proportionality principle. In dealing with the case, the tribunal would have to put themselves in the position of the armed forces member concerned, taking all the surrounding circumstances into account, so as to experience the situation as he experienced it at the time, before considering the question of culpability.
Operational reality is recognized in the Statute of the International Criminal Court. This makes the infliction of incidental loss or damage an offence only if the attack is launched intentionally in the knowledge that it will cause incidental civilian loss of life, injury or damage which would be clearly excessive in relation to the concrete and direct military advantage anticipated [44 ] . The use of the word “clearly” e nsures that the court would be involved only in cases where the excessiveness of incidental damage was obvious, not in cases that involved errors of judgement by commanders in the field. [45 ]
A member of the armed forces doing his level best in difficult circumstances to comply with the law of armed conflict would have nothing to fear from a subsequent inquiry, even if he made a mistake under pressure.
The law does not demand that there be no casualties in armed conflict. However, the law, political expediency and public
sentiment combine to demand that casualties, whether among members of the armed forces or among the civilian population, should be reduced to the maximum extent that the exigencies of armed conflict will allow. An important element of this endeavour is verification of the target, because attacking the wrong target is likely to lead to unnecessary casualties. Target verification requires reasonable care to be exercised. The precise degree of care required depends on the circumstances, especially the time available for making a decision. In the event of doubt about the nature of the target, an attack should not be carried out, with a possible exception where failure to prosecute the attack would put attacking forces in immediate danger.
1. Jamie Shea, at the NATO press briefing on 26 March 1999, website: www.nato.int.
2. Air Commodore David Wilbey, at the NATO press briefing on 1 April 1999.
3. Air Chief Marshal Sir Richard Johns, “Air power in a new era”, Royal Society of Arts Journal , 3-4 of 1999, p. 99.
4. As when a bomb fell 600 metres short of the target, at Aleksinac. There was also the incident on 12 April when missiles were fired at a railway bridge near Leskovac, just as a train from Belgrade to Thessaloniki was crossing it.
5. Indeed, the preamble to the Declaration of St. Petersburg of 1868 proclaimed that the only legitimate object of war was to weaken the military forces of the enemy and that, for this purpose, it was sufficient to disable the greatest possible number of men.
6. For the British Empire forces, 908,371 military battle deaths and 2,090,212 military wounded out of a mobilized military strength of 8,900,000. See the table in the British Army Review , August 1996, p. 79.
7. The Independent , 4 November 1999.
8. Op. cit. (note 6), p. 80.
9. Ibid . According to Erik Durschmied, The Hinge Factor, Hodder & Stoughton, 1999, Iraqi military losses were over 100,000 whereas Coalition forces suffered a total of 192 killed, of whom 35 died from “friendly fire” and two were killed dismantling a bomb. “In military terms, such a ratio is called the Zero Factor.”
10. “NATO won the Kosovo conflict without a single life lost in combat operations on its own side (…) for just two aircraft lost”, Nick Cook, “War of extremes”, Jane's Defence Weekly , 7 July 1999. Two Gurkha engineers, Lt.Gareth Evans and Sgt. Balaram Rai, were killed dealing with unexploded ordnance during the operation.
11. According to the Swiss Federal Office for Civilian Protection, the ratio of the First World War was 200 military: 1 civilian; in the Second World War nearly 1:1 and in the Vietnam War, 1 military: 20 civilians. See M.Sassoli/A. Bou vier, How Does Law Protect in War? , ICRC, Geneva, 1999, p. 145.
12. “While such errors were inevitable, observers contend, widespread public faith early on in the campaign in the ability of modern surveillance systems and smart weaponry to restrict collateral damage heightened reaction against the bombing campaign when civilian casualties did occur and led to intense pressure at times on politicians to call a halt to the attacks … If politicians insist on minimum -- and perhaps even zero -- attrition rates among their military personnel in future air campaigns, a range of new weapons, up to and including UCAVs, will need to be introduced as soon as the technology allows. The alternative for the military is to brace civilians -- politicians, media and public alike -- for the realities of attrition in parallel with an increased investment in equipment and tactics that support manned close air support and battlefield air interdiction missions at low as well as high altitudes.” Cook, op. cit. (note 10).
13. Johns, op. cit. (note 3), p. 96.
14. Charles Allen (ed.), Thunder and Lightening, HMSO, 1991, p. 74, 80.
15. Ibid . p. 81/2.
16. Ibid ., p. 137: “Aircrew usually had the comfort of being remote from the destruction caused by their bombs or rockets.”
17. Ibid ., pp. 113-114: “That really brought it home (…) we were actually killing people ”
18. According to Wing Commander Greg Bagwell, “Precision weapons”, Royal Air Force Air Power Review , Spring 1999, p. 7, the Al Firdos bunker in Baghdad was assessed as a communication node but, unknown to allied intelligence, it was also being used as a
civilian air raid shelter. The Chinese Embassy in Belgrade was hit by mistake on 7 May 1999 owing to the use of faulty intelligence at the planning stage which went unnoticed through subsequent checks. See the report in The Independent , 11 May 1999.
19. Where thermal, acoustic or radar “blueprints” of enemy military equipment are programmed into a weapon's memory. Op. cit. (note 18), p. 7.
20. Apart from the new US joint direct attack munition (JDAM) which can be deployed from altitudes above cloud cover. See Joint Statement on the Kosovo After Action Review , Department of Defense, 14 October 1999.
21. Op. cit. (note 14), p. 81: “… in reality you cannot have a surgical war. There is no such thing. For a bomb to go down an elevator shaft is just luck. It probably has a CEP -- a computer error probability -- of maybe 30-40ft. That can make the difference between hitting the place you're aiming for and the one next door which is full of civilians.” Incidentally, other authors define CEP as “circular error probable”, i.e. the radius of a circle centred on the target within which 50% of all bombs dropped will land. Bagwell, op. cit. (note 18), p. 14. Bagwell, p. 8, also comments that “the explosive force of a 500lb warhead does not constitute keyhole surgery!”
22. Ibid ., p. 13, quotes $9,000 for a simple GPS guided bomb and $500,000 for a Block III Tomahawk missile.
23. Ibid ., p. 3.
24. Defined by Bagwell, op. cit. (note 18), p. 14, as tho se requiring no human input during the final phase of flight.
25. At the NATO press briefing on 18 April 1999, an account was given of how a pilot launched an attack against a radar and, noticing that the site was close to a church, pulled his weapon off the target so that it exploded harmlessly in a wood.
26. Op. cit. (note 10).
27. Lord Robertson, Kosovo: An Account of the Crisis , Ministry of Defence, London, 1999, p. 13.
28. At 1800 hours BST on 17 April 1999.
29. NATO press briefing, 15 April 1999.
30. Or possibly Terzick Most Zrze.
31. The Independent , 17 April 1999.
32. To be found on www.fas.org/man/dod/ docs99/s990419b.htm.
33. At the time of the Kosovo conflict, for example, the USA and France were not party to Protocol I but Germany and the UK were.
34. Protocol I, Art. 57, para. 2(a)(i).
35. The United Kingdom made a declaration of understanding to this effect when ratifying Protocol I. See IRRC, No. 322, March 1998,
pp. 186 ff. -- The Conventional Weapons Convention of 1980 contains similar language in the Amended Mines Protocol of 3 May 1996, Art. 3, para. 10.
36. Protocol I, Arts 50, para. 1, and 52, para. 3. With regard to Art. 50, para. 1, the United Kingdom, on ratification of the Protocol, made a statement of understanding that the rule applied “only in cases of substantial doubt still remaining after the assessment (…) has been made, and not as overriding a commander's duty to protect the safety of troops under his command or to preserve his military situation, in conformi ty with other provisions of the Protocol.” Ibid .
37. This passage is adapted from the author's text in the Model Manual on the Law of Armed Conflict, ICRC, Geneva, 1999.
38. The United Kingdom, on ratification of Protocol I, declared that “Military commanders and others responsible for planning, deciding upon, or executing attacks necessarily have to reach decisions on the basis of their assessment of the information from all sources which is reasonably available to them at the relevant time.” Loc. cit. (note 35).
39. Bagwell, op. cit. (note 18), p. 9, thinks that this should be dealt with in rules of engagement: “… some targets could require such guaranteed accuracy that the aircraft/ platform might be placed at increased risk to enemy defences. In this case the ROE will need to make the desired identification criteria extremely clear so that crews, who are effectively judge and jury, will be able to take the right course of action.”
40. A. P. V. Rogers, “Conduct of combat and risks run by the civilian population”, Military Law & Law of War Review, 1982, p. 310.
41. British Defence Doctrine (JWP 0-01) issued by the British Minster of Defence in 1996.
42. Protocol I, Art. 58.
43. See A. P. V. Rogers, Law on the Battlefield , Manchester University Press, 1996, p. 79.
44. Statute of the International Criminal Court, Art. 8, para. 2(b)(iv).
45. Roy S. Lee (ed.), The International Criminal Court , Kluwer Law International, 1999, p. 111.
Abstract in French