Kosovo: Lessons From The Crisis

Chapter 7: Air Operations - Conduct and Lessons

Key lessons
 

The conduct of military, including air, operations must reflect political realities. Current doctrine should remain essentially unchanged, but its practical implication should always take account of these wider considerations.  The effective performance of the Tomahawk Land Attack Missile in UK service was proved.  There is a need for the UK and its Allies and partners to improve capabilities in the following areas:  Precision joint all-weather attack capability against both static and mobile ground targets;  Intelligence, Surveillance and Reconnaissance (ISR);  improved secure communications/data links, and better "sensor to shooter" links;  Electronic Warfare/Suppression of Enemy Air Defences;  Air to Air Refuelling;  Battle Damage Assessment.

Strategy and planning

7.1. NATO launched military operations only as a last resort when it was clear that the diplomatic track would not deliver a solution, whilst at the same time the humanitarian situation on the ground had deteriorated to such an extent that outside intervention became essential in order to avert a humanitarian catastrophe.

7.2. Targets for air strikes were selected by the NATO Military Authorities, acting in accordance with guidance agreed by the North Atlantic Council on broad sets of targets and the requirement to minimise collateral damage. The North Atlantic Council was not involved in the detailed process of target selection. Individual Allies were responsible for the clearance of the targets assigned to them by NATO.

7.3. The Allies hoped that the air operation would be short, and that Milosevic would again, as he had in the past, back down when confronted with the threat or use of force. But they did not take this for granted, and when he did not, operations were intensified to increase the pressure on Milosevic, his regime and his forces in order to achieve NATO’s objectives. It would have been possible at any time for Milosevic to bring the air operation to an end by withdrawing his forces from Kosovo. The initial phase of the air operation was designed to degrade the Yugoslav Integrated Air Defence System, the Yugoslav/Serbian Command and Control infrastructure, airfields and aircraft and deployed heavy weapons in Kosovo. The subsequent phase widened the operation to include targets of high military value across Yugoslavia.

7.4. There have been criticisms of the graduated approach adopted by NATO, because it did not conform to the classic doctrine of air power, which suggests early and heavy strikes against strategic targets. In the case of Kosovo, however, this overlooks three important factors. First, it was possible that Milosevic may have decided to concede after the initial strikes. An overly intensive and destructive start to the campaign might have precluded this possibility. Secondly, the launch of a major military campaign was in domestic, legal and political terms not an easy decision for a number of NATO Allies. Concern about public opinion was evident in some countries. A graduated approach to the start of the air operation made it possible to build support for the aims and objectives of NATO’s strategy. We should not draw the lesson from this that established doctrine needs to be amended, as in many future cases it may remain appropriate, but judgement will always be required on its practical application, taking account of the political realities of the situation. Finally, the operation was designed in phases (as described above), and targets throughout Yugoslavia were attacked in significant numbers from the outset.

7.5. The air operation was pursued on two axes, against strategic targets of high military value, for example headquarters buildings in Belgrade, and tactical targets in and around Kosovo such as military vehicles and heavy weapons. Both axes were essential in pursuing NATO’s military objective, to degrade the capabilities of Yugoslav/Serbian security forces and to limit their ability to continue their activities in Kosovo. We have been criticised for pursuing tactical as well as strategic targets, as this demanded the commitment of significant resources. Planning had always assumed that we would attack targets on both axes, but the level of attention we gave to fielded forces was increased in response to the shocking scale on which Milosevic’s forces pursued their policy of ethnic cleansing.

7.6. Attacks against tactical targets in Kosovo proved to be a significant challenge to the Alliance, given the difficulties in locating and positively identifying targets. The Yugoslav/Serbian security forces concealed their assets to a considerable extent, and were adept at deception techniques, including the use of decoys. But we were successful in that our efforts in Kosovo forced the security forces to conceal their tanks and heavy weapons from NATO attack, and thus limited their ability to use these weapons against Kosovo Albanians. Through attacks, and the threat of attack, NATO aircraft influenced the situation on the ground.

7.7. The Kosovo campaign was notable for the wide use of asymmetric (that is to say non-conventional) tactics by the Yugoslav/Serbian forces. Examples included: the location of tanks and other military equipment in the middle of villages and in other locations where the Yugoslav/Serbian forces knew that our concern to minimise collateral damage would prevent us from targeting them; at least one case of the use of human shields was documented by Human Rights Watch, and the OSCE suggest there may have been more; attacks against civilians; and extensive disinformation/propaganda. In future, if potential adversaries consider themselves unable to oppose us by conventional means, they may increasingly resort to other techniques such as those outlined above. We and our Allies will continue to prepare ourselves for this challenge.

Accuracy and the minimisation of collateral damage

7.8. Our experience in the Gulf War had demonstrated the need for precision attack capabilities, and the extent to which we have improved our capabilities in this field was proved in Operation DESERT FOX, the operation against Iraq in December 1998. Building on this success, Kosovo was one of the most accurate air operations ever mounted, and resulted in very few instances of collateral damage. (DESERT FOX was a much shorter operation, and used a far greater percentage of precision munitions than was the case in Kosovo.)

7.9. Collateral damage is the term used to describe the unintended loss of civilian life, or injury to civilians, or the damage to civilian property, which is caused by attacks on military objectives. Accuracy in attack, and taking of all feasible precautions with a view to avoiding, and in any event, minimising collateral damage, are important both politically and legally. We made the point repeatedly during the Kosovo campaign that our dispute was with the policies of Milosevic and his regime – we were not at war with the people of Yugoslavia. It is not possible to avoid some collateral damage in any significant armed conflict. This fact is recognised by international law, which outlaws attacks on military objectives which may be expected to cause collateral damage which would be excessive in relation to the concrete and direct military advantage anticipated. All pre-planned UK targets were assessed against the requirements of international law, and where UK pilots were authorised to attack opportunity targets, or needed to defend themselves, they did so with meticulous regard for the rules of engagement. The UK went to great lengths to minimise the risks of any civilian death, injury or damage. This was in marked contrast to the flagrant disregard for human life shown by the Yugoslav/Serbian forces, whose actions provided clear evidence of the need for NATO to have acted.

7.10. Although NATO has no independent means of verifying the number of civilian casualties, Human Rights Watch estimated in their report of February 2000 (available on the internet, address at Annex E) that there were only 90 incidents of collateral damage which involved fatalities. Moreover, about two thirds of the deaths they reported to be as a result of just 12 of these incidents. If this figure is correct, out of a total of some 10,500 strike sorties conducted by NATO forces, during the course of which some 23,600 pieces of ordnance were delivered, less than one percent of all missions led to unintended fatalities.

7.11. Human Rights Watch have estimated that between 488 and 527 Yugoslav civilians were killed as a result of NATO air strikes, far fewer than the (varying) estimates by the Yugoslav authorities, which range between 1200 and 5700.

7.12. The decision by NATO that operations should be flown at medium altitude (above 15,000 feet) was taken in recognition of the threat from Yugoslav air defences, which continued throughout the campaign. It has been alleged that this increased the risk to innocent civilians, but munitions were delivered from this altitude with great accuracy. If ever our pilots were in any doubt about the targets they had been tasked to attack, they did not drop their munitions. The 15,000 feet minimum operating altitude was introduced to ensure that NATO aircraft operated at an acceptable level of risk but, as the operation progressed, some NATO aircraft operated at lower altitudes when necessary to acquire and identify targets.

7.13. While we took every care to minimise the risk to our pilots, there was no requirement for "zero casualties", as has been alleged. Our decisions on tactics were based on a balance of risk against potential gains – we rightly took sensible precautions, and our tactics were successful. Our pilots faced a continual threat from ground-to-air missiles and Anti-Aircraft Artillery, including at medium altitude, and there was always a risk, despite the counter-measures deployed. It is a testimony to the exceptional skill of our aircrew, who were forced to manoeuvre aggressively and deploy counter-measures to defeat the threat from these weapons, that no UK aircraft were shot down. Only two Allied aircraft were lost over Yugoslavia (neither UK aircraft), and in both cases the aircrew were recovered safely. (See also paragraph 7.35 below on Combat Search and Rescue capability.)

7.14. Despite the emphasis on medium level operations during the Kosovo campaign, RAF aircraft were ready to fly at low level if required, and planned to do so in support of any opposed ground operation. Maintaining the skills required for flying at low altitude therefore remains important.

How much damage was done?

7.15. The short answer is enough. As discussed earlier in the paper, the success of the air operations, although essential, is unlikely to have been the sole reason why Milosevic eventually agreed to the international community’s demands, but it is certain to have had a substantial effect.

7.16. NATO attacked a total of 440 static targets during the air campaign, with more than three-quarters assessed to have suffered moderate to severe damage. NATO figures suggest that the targets destroyed or significantly damaged included:

• 14 command posts 
• 10 military airfields 
• over 100 military aircraft 
• 34 road bridges 
• 11 railway bridges 
• 29% of all Yugoslav/Serbian ammunition storage capacity 
• 57% of petroleum reserve capacity 
• all Yugoslav oil refineries.

7.17. General Clark (the Supreme Allied Commander Europe or SACEUR) made public NATO’s Battle Damage Assessment of attacks against mobile targets on 16 September 1999. The text, which includes a detailed explanation of the Battle Damage Assessment validation process, is available via the NATO internet site (address available in Annex E to this report). In short, he reported validated strikes on 93 tanks, 153 armoured personnel carriers, 389 artillery pieces and mortars and 339 military vehicles. It is not always possible to ascertain whether a target has been totally destroyed or what degree of damage it has sustained.

7.18. The UK and NATO have been accused of deliberately inflating the success of the campaign in public statements during the conflict. This is untrue. Public statements were always based on the best available information at the time. It would have taken some considerable time to ensure that all our assessments were 100% accurate. We judged it important to release as much information as possible as soon as we could.

7.19. The information which is often released in public shortly after an attack is based on initial reports, and its accuracy therefore cannot be guaranteed. Although these predictions are sometimes referred to as "Battle Damage Assessment", it is quite different from Military Battle Damage Assessment, which is a three phased assessment, and which usually takes some time to complete. Military Battle Damage Assessment examines:

(a) the effectiveness of an individual attack against a single target;
(b) the effectiveness of attacks in meeting the targeting objectives, based on a number of attacks against a target system; and
(c) the effectiveness of the overall campaign to date measured against the commander’s strategic objectives.

7.20. Various systems were employed to assess battlefield damage, all complementary to each other rather than being mutually exclusive. Reconnaissance aircraft, Unmanned Aerial Vehicles, satellite imagery and other intelligence gathering systems all provided data that could be used to assess the amount of damage caused by the NATO bombing campaign.

7.21. In the light of our experience during the air campaign, the UK national Battle Damage Assessment/targeting process has been fully reviewed and a series of recommendations for follow-up action highlighted. These include:

• the need to ensure the most appropriate mix of Intelligence, Surveillance and Reconnaissance assets is available to provide Battle Damage Assessment;
• sufficient background information should be compiled to enable the accurate assessment of the impact of operations and of an adversary’s remaining capability;
• and the limitations of the human eye and weapons systems video should be recognised and, where possible, supporting intelligence material obtained.

7.22. The recommendations of this review of the Battle Damage Assessment process will be included in future standing operating procedures.

Command arrangements

7.23. The overall NATO commander of the air operation was General Clark (SACEUR). Admiral Ellis, the Commander in Chief of Allied Forces South (CINCSOUTH) was the Operational Commander under General Clark, and had Operational Control (OPCON) of all assigned air forces. Tactical Control (TACON) was delegated by SACEUR through CINCSOUTH to Commander AIRSOUTH (Lieutenant General Short), the Combined Force Air Component Commander (CFACC), who operated from the Combined Air Operations Centre (CAOC) in Vicenza, Italy.

7.24. National Command of all UK forces in Italy involved in Operation ALLIED FORCE was delegated from MOD through the Chief of Joint Operations at Northwood to the Commander British Forces Italy (Air) (CBFI(Air)). HMS SPLENDID remained under national command throughout the operation, under the Operational Command of the Commander in Chief Fleet at Northwood.

The UK's contribution to the air operation

7.25. The contribution made by UK armed forces to the air operation is detailed in Annexes B, C and D.

7.26. Both Royal Navy and Royal Air Force assets were directly involved in air strikes, with Army assets such as the Phoenix Unmanned Aerial Vehicle operating in support. Operation ALLIED FORCE saw the first operational use of UK Tomahawk Land Attack Missiles (TLAM), a highly effective and highly accurate capability which was fully integrated into the wider operation. UK TLAM made a significant contribution throughout the operation, but especially in the early stages. The UK TLAM operational capability was developed and available particularly quickly, not long after trials were completed. UK TLAM operations were conducted in close co-ordination with US cruise missile operations. The significant utility of this capability has been proven, and is soon to be complemented by a RAF air-launched capability, Storm Shadow, a Conventionally Armed Stand Off Missile (CASOM). The Royal Navy has modified its TLAM operating procedures in the light of experience gained during the operation.

7.27. The UK’s initial commitment of aircraft for strike operations was eight RAF Harrier GR7s of No 1 Squadron from RAF Wittering deployed to Gioia del Colle. These were supplemented by a number of aircraft in vital support roles, including two TriStar tankers at Ancona, three E-3D Airborne Early Warning (AEW) aircraft from Nos 8 and 23 Squadrons and a Nimrod from No 51 Squadron. All of these aircraft were based in Italy during the operation.

7.28. An additional four GR7s moved to Gioia del Colle on 28 March. When NATO requested more strike aircraft, eight Tornado GR1s flew air operations from their base at RAF Bruggen, Germany, supported by three VC10s from the same location. These long range missions initially required significant amounts of in-flight refuelling and long hours in the cockpit for the crews. In late April, the UK responded to a NATO request for more aircraft and a further four Harrier GR7s together with an additional TriStar tanker aircraft were sent to the region. When NATO again requested an increase in the number of strike missions, the decision was taken to deploy forward a force of twelve Tornado GR1s to the Solenzara Air Base in Corsica. With the assistance of the French Government and in close co-operation with the French Air Force, the Tornados operated from Corsica from 1 June. The move also allowed a rationalisation of Air to Air Refuelling assets committed to the campaign, and the VC-10 tanker fleet was re-deployed to Italy to support the overall NATO refuelling effort, rather than just the Tornado missions.

7.29. The E-3D aircraft, although originally designed as AEW aircraft, were employed in the Airborne Warning and Control (AWACS) role as part of a complex command and control operation guiding hundreds of Alliance aircraft in what had become the busiest airspace in the world. This stretched the aircraft’s resources, and the long term sustainability of using the aircraft in this role is being addressed.

7.30. There was a very real threat to unarmed NATO aircraft such as tankers and E-3Ds, and two Serbian Mig 29s were in fact shot down over Bosnia by NATO fighters. In recognition of the dangers which these personnel faced, several tanker and E-3D pilots and mission crew were decorated, including the first Distinguished Flying Cross ever to be awarded to a tanker pilot.

7.31. Largely unnoticed by the media, but vital nonetheless, was the role of the Air Transport Force, whose Hercules C-130, VC10 and TriStar transport aircraft flew over 500 sorties into the Balkans theatre during the period of the conflict. These missions supported the UK’s humanitarian relief effort, the air operation and the build up and sustainment of ground forces. These efforts were supplemented by use of chartered civilian aircraft, including heavy lift aircraft which could carry loads too big for the RAF’s aircraft. By early June, the total number of Support Helicopters in theatre numbered eight Chinooks and six Pumas, the largest Support Helicopter deployment since the Gulf operation.

7.32. By far the single largest contribution of aircraft to the operation came from the US, but other members of the Alliance provided a similar proportion of their available aircraft. By the end of the operation over nine hundred aircraft from fourteen countries had taken part in the operation.

7.33. One of the major obstacles in pursuing the campaign was the weather. Only on 21 out of the 78 days of the campaign was the weather judged to be "favourable" for air operations. We had known that the weather at this time of the year would be bad, but the timing of the campaign was not in our hands – it was Milosevic’s actions which determined when we had to act. Inevitably, the weather had an effect on the tempo of the air campaign, but NATO commanders were able to continue operations at a lower intensity through the use of various assets including UK and US TLAM, other US cruise missiles, and the US Global Positioning System (GPS)-guided Joint Direct Attack Munition (JDAM). Following trials in the UK, which had confirmed the accuracy of such methods, UK pilots were cleared to use non-guided ("dumb") bombs through cloud against targets where the expected collateral damage (if any) was not excessive in relation to the direct and concrete military advantage anticipated, verifying their exact position by use of GPS equipment.

Weapons stockpiles

7.34. Sufficient numbers of weapons were available throughout the operation, with contingency plans in place to ensure that additional weapons would be available if needed, although these were not in the event required. Precision Guided Munitions were in particularly high demand during this operation, and this put our stockpile of this weapon under some pressure. Nevertheless, the majority of our stockpile of each of the weapon types used remained available at the end of the conflict, which would have enabled us to continue air operations for some time thereafter. Current stockpile guidance is being reviewed in the light of experience of the air operation and the planning for a ground operation, to ensure that we have the right quantity and balance of weaponry in stock.

Combat Search And Rescue

7.35. We relied on our Allies, particularly the US, for Combat Search and Rescue (CSAR) capability. The effect on the morale of Allied aircrew of the successful operations to rescue the US aircrew on two occasions was very considerable, and showed the professional competence of the Allies, while denying Milosevic propaganda opportunities. We are looking at the requirement for a UK or European capability.

Equipment

7.36. Our experiences of the campaign have highlighted areas of capability where we and our Allies and Partners can do better in the future. In taking forward action on these capability gaps, it is not necessary for the UK to look for enhancements in all these areas. We will act together with our Allies and partners through the Defence Capabilities Initiative and the European Defence Initiative, as well as nationally through the usual defence programme procedures.

Precision Attack

7.37. Precision Guided Munitions including the Royal Navy’s TLAM, the Royal Air Force’s Laser Guided Bombs (Paveway II and Paveway III) achieved good results. We had built up our precision attack capability as a consequence of lessons from the Gulf War, and this paid dividends, with our laser-guided systems performing well, and in accordance with expectations. Nevertheless, the limitations of our current laser guided systems in poor weather were highlighted, as cloud cover often prevented laser designation of targets, accurate delivery of ordnance became impossible, and was therefore not attempted.

7.38. The RAF already has a well-practised low-level all-weather attack capability against static targets, but not with precision. Given the strong possibility that future operations could again be from medium altitude, we are considering how best to be prepared for such an eventuality. The United States was very pleased with the performance of its GPS-guided JDAM, which was used accurately and effectively in all weathers. We are looking at the feasibility of developing a similar system for use by UK aircraft that may provide us with an effective means of engaging targets in all weathers. Decisions will be made in due course.

7.39. Despite the considerable utility of Precision Guided Munitions, there remains an effective role to be played by non-guided munitions. Particularly if the location of the aircraft can be verified by use of GPS systems, a non-guided bomb can be dropped with considerable accuracy. Targets for such bombs should obviously be considered carefully, but where the target is large (eg a munitions storage site) and there is a limited risk of collateral damage, it remains an alternative option. Ultimately, in this and other examples, weapon selection is always tailored to the selected target.

Attacks against mobile ground targets

7.40. As described above, NATO encountered significant difficulty in locating and positively identifying mobile ground targets. Given the long period over which the crisis had been developing, the Yugoslav/Serbian security forces had had time to disperse their personnel, equipment and logistics resources. Their war-fighting doctrine places great emphasis on dispersal, the use of camouflage, dummy targets, concealment and bunkers and they applied this doctrine competently. We knew that they would use these tactics, which are difficult to counter. We can expect future opponents to use these tactics also, and we will look at what we can to improve current detection methods.

7.41. In order to engage such targets more effectively, we and our Allies and partners need to look to acquire/develop:

• an improved Intelligence, Surveillance and Reconnaissance capability, to enable us to detect the right targets, in all weathers, and to distribute the intelligence to those who need it in timely fashion. Our planned all-weather Airborne Stand Off Radar (ASTOR) aircraft will greatly enhance our battlefield surveillance capability, alongside existing Allied capabilities such as the US JSTARS and Unmanned Aerial Vehicles, including the UK Army’s PHOENIX and planned successor systems such as SENDER and SPECTATOR. As technology advances, it should also become easier to detect where assets have been concealed; 
• improved real-time "sensor-to-shooter" communications/data links (to enable faster communication between intelligence gathering assets and the means of attacking the target eg a pilot in an aircraft on patrol);
•  in certain cases, additional types of weaponry to maximise the chance of success against a given target. We believe it remains appropriate to retain a range of capabilities, to tailor the response to the target. We were impressed by the performance of the Maverick anti-armour missile during the operation, and the Defence Secretary announced in March 2000 trials of this system on UK aircraft. Possession of such a missile during the Kosovo air operation would have enabled us to target tanks, other armoured vehicles and similar targets (eg air defence assets) more effectively, and with a low risk of collateral damage as the missile is locked onto the selected target before launch. Maverick would be complementary to the planned Brimstone missile, as the latter would be excellent for attacks on groups or columns of vehicles in war scenarios like the Gulf in 1991, as it is designed to seek out and destroy armoured targets autonomously.

Secure communications

7.42. The lack of a compatible secure air-to-air communications system with all Allies during the air operation caused problems, as we were unable to exchange freely some operationally sensitive information. Along with most Allies we used frequency-hopping technology and transmission security measures, which provided some degree of protection. Increasing advances in the technology available to countries such as Yugoslavia will make these procedures and systems more vulnerable and there is a need to enhance the security of communications in a combined and joint framework. Some NATO and US aircraft therefore are now making use of encrypted communications equipment. We have recognised this as a key lesson, and priority action on this issue was announced by the Defence Secretary in March 2000. We are now trialling enhanced secure air-to-air communications on a number of RAF aircraft, and if this proves successful, we will consider widening its fit.

Electronic Warfare and Suppression of Enemy Air Defence

7.43. Electronic Warfare (EW) and Suppression of Enemy Air Defence (SEAD) capabilities were vital force enablers during the air campaign. The first stage of the air operation was aimed at (and succeeded in) degrading the Yugoslav Integrated Air Defence System, but as a threat remained, force packages were escorted by EW and SEAD-capable aircraft to counter these threats as they arose. The bulk of this effort was provided by the United States, although other Allies, including the UK to a very limited extent (see below), played a role. We are looking at SEAD capability, and at whether a significantly increased capability could be achieved by improved Intelligence, Surveillance and Reconnaissance and better sensor-to-shooter links (as described in the section on attacks against mobile ground targets) rather than by buying more SEAD-capable aircraft to ensure that all sorties are escorted.

7.44. We are also looking at fitting improved missile approach warning systems to our aircraft, as well as improving countermeasures.

Air to Air Refuelling

7.45. Air-to-Air Refuelling (AAR) aircraft performed a vital role as force multipliers, but these assets were in short supply during the air campaign. Although some other European nations have limited numbers of AAR aircraft, the UK has by far the largest and best probe-and-drogue AAR capability. The RAF AAR aircraft committed to the operation spent much more time refuelling aircraft of other Allies, including US fighters, than our own aircraft. By the end of the operation 85% of their fuel had been spent refuelling aircraft from other countries. It is clear that more resources need to be invested in this area, but given our existing capability, and our plans to renew it through a major public private partnership, this is more of a priority for our European Allies and partners than ourselves.

Cluster Bombs

7.46. The use of cluster bombs by UK armed forces during the Kosovo conflict has attracted adverse media comment. UK armed forces will always use the weapons systems judged most effective against a given target, taking into account the need to minimise collateral damage. The bomblets are designed to detonate on impact but, as with any other similar munitions, a small percentage failed to do so. The manufacturer’s estimated failure rate for the RBL 755 cluster bomb used during the Kosovo conflict is approximately 5%. Contrary to a number of stories in the media, cluster bombs should not be confused with anti-personnel type weapons that are specifically designed to lie dormant and detonate once disturbed. Cluster bombs are an effective weapon against area targets such as a group of soft-skinned military vehicles. Nevertheless, we have learned from the Kosovo campaign that it would be useful to have a capability to strike single vehicles more accurately, hence the trial of the Maverick missile.

7.47. We have been criticised for continuing to use cluster bombs after the US had ceased to use them for safety reasons. The type of cluster bomb used by the US was discovered to have a fault and was temporarily withdrawn from service, being returned to service shortly thereafter for use until the end of the conflict. The UK uses a different type, which continued to function normally, and there was thus no reason for it to be withdrawn.

Depleted Uranium

7.48. Depleted Uranium armour-piercing ammunition was used by NATO aircraft during the Kosovo conflict, although not by the UK. Depleted Uranium based ammunition is used because of its unique capability as a kinetic penetrator against the most modern types of Main Battle Tank armour. At present, no satisfactory alternative material currently exists to achieve the levels of penetration necessary to defeat modern tanks. Depleted Uranium has a level of chemical toxicity that is similar to other heavy metals such as lead, and the health risks from exposure to it are assessed as low. Its use is not prohibited under any international agreements, and the International Committee on Radiation Protection does not list Depleted Uranium as a health hazard. As it is uniquely effective against certain targets, the use of Depleted Uranium ammunition by UK forces is retained as an option. Denying ourselves the most appropriate weapons would not help end conflicts quickly and could put our armed forces at greater risk of harm.

7.49. Public concerns about the use of Depleted Uranium are fully recognised, and the MOD is willing to consider carefully any peer-reviewed studies based on appropriately validated data on the environmental and health aspects of Depleted Uranium, and to examine the use of alternative materials in armour piercing munitions.

What is the cost of the improvements announced so far?