Playing a blinder

Posted on May 9, 2015 by Derek Gregory

A characteristically smart post from Larry Lewis at War on the Rocks about Obama’s promise to investigate the mistakes made in the CIA-directed drone strike that unwittingly killed two hostages in Pakistan in January 2015. ‘We’ve been on that path before, in Afghanistan,’ he writes, ‘and we know where it leads: more promises followed by a repeat of similar mistakes.’

Larry explains that the US military was causing an ‘unacceptable number’ of civilian casualties in Afghanistan between 2006 and 2009:

When an incident occurred, they investigated the incident, made changes to guidance, and promised to keep such an incident from happening again. But these incidents kept happening. So the military repeated this ineffective review process again and again. This “repeat” cycle was only broken when military leaders approved the Joint Civilian Casualty Study, a classified outside review requested by General Petraeus. This effort had two key differences from earlier efforts. First, it was independent, so it was able to overcome false assumptions held by operating forces that contributed to their challenges. And second, the study looked at all potential civilian casualty incidents over a period of years, not just the latest incident. This approach helped identify systemic issues with current tactics and policies as the analysis examined the forest and not just the nearest tree. This study also considered different sets of forces operating within Afghanistan and their relative propensity for causing civilian casualties.

You can access the unclassified Executive Summary – co-written by Larry with Sarah Sewell – here. I’ve noted Larry’s important work on civilian casualties before – here, here and here – but his short Op-Ed raises two issues that bear emphasis.

The first is that it is a mistake to abstract strikes carried out by a Predator or a Reaper from air strikes carried out from conventional platforms; the latter are often facilitated and even orchestrated by a UAV – as in the ‘signature’ case of the Uruzgan strike in 2010 – but, pace some drone activists, our central concern should surely be the wider matrix of military violence. This also implies the need to articulate any critique of CIA-directed drone strikes in Pakistan with the use of air power in Afghanistan (and not only because USAF pilots fly the ‘covert’ missions across the border). Here General Stanley McChrystal‘s Tactical Directive issued in July 2009 that directly addressed civilian casualties is a crucial divide. As Chris Woods emphasizes in Sudden Justice,

‘Radically different tactics were now being pursued on either side of the “AfPak” border…. Even as Stanley McChrystal was cutting back on airstrikes in Afghanistan, the CIA was escalating its secret air war in Pakistan’s tribal areas.’

The second issue is the extraordinary partitions – blinkers might be more accurate – that seem to be imposed on military operations and investigations. In the case of the Uruzgan attack, for example, a military lawyer was called in at the eleventh hour to monitor the video feeds from the Predator as it tracked a ‘convoy’ (a term surely as leading as ‘Military-Aged Male’) in the early morning. As the next two slides show, taken from my ‘Angry Eyes’ presentation, the JAG knew the Rules of Engagement (ROE) and the Tactical Directive; he obviously also knew the legal requirements of proportionality, distinction and the rest.

Knowing the ROE, the Tactical Directive and the formal obligations of international law is one thing (or several things): but what about ‘case law’, so to speak? What about knowledge of other, similar incidents that could have informed and even accelerated the decision-making process? In this case, before the alternative course of action could be put into effect and an ‘Aerial Vehicle Interdiction’ set in motion – using helicopters to halt the three vehicles and determine what they were up to – two attack helicopters struck the wholly innocent ‘convoy’ and killed 15-21 civilians. Fast forward to the subsequent, I think forensic Army investigation. This is the most detailed accounting of a ‘CIVCAS’ incident I have read (and you’ll be able to read my analysis of it shortly), and yet here too – even with senior military legal advisers and other ‘subject experts’ on the investigating team – there appears to be no reference to other, similar incidents that could have revealed more of the ‘systemic issues’ to which Larry so cogently refers.

This is made all the stranger because there is no doubt – to me, anyway – that the US military takes the issue of civilian casualties far more seriously than many of its critics allow.

‘That others may die’

Posted on January 10, 2015 by Derek Gregory

As I am (at last) moving into the finishing stages of my ‘Dirty Dancing’ essay on CIA-directed drone strikes in Pakistan’s Federally Administered Tribal Areas, it’s time to round up some of the latest work on drones and civilian casualties across multiple theatres.

First, Afghanistan: the principal theatre of US remote operations. I’ve noted Larry Lewis‘s remarkable work before (here and here), based on classified sources, and in particular this claim (see also here):

Drone strikes in Afghanistan were seen to have close to the same number of civilian casualties per incident as manned aircraft, and were an order of magnitude more likely to result in civilian casualties per engagement.

As I said at the time, the distinction between an ‘incident’ and an ‘engagement’ is crucial, though most commentators who have seized on Larry’s work have ignored it and focused on the dramatic difference in civilian casualties per engagement. Despite my best efforts, the Pentagon were unwilling to clarify the difference, so here is what Larry himself has told me:

An engagement is probably intuitively what you would expect – the use of force against a target. The distinction is the term incident, which is borrowed from ISAF definitions. I should have said “civilian casualty incident.” This refers to an engagement that results in civilian casualties.

This means that, if you look at the collection of civilian casualty incidents, the average number of civilian casualties is close to the same for manned and unmanned platforms. At the same time, the rate of civilian casualties for the two platforms is markedly different, with unmanned platforms being ten times more likely to cause civilian casualties than manned platforms. That doesn’t mean that drones caused more civilian casualties than manned aircraft, by the way, since the denominators (number of engagements of manned aircraft versus drones) can and in fact were very different. But it does suggest that the relative risk of civilian casualties was higher for one kind of platform versus the other.

And this is in the specific context of Afghanistan and for a specific time. I wouldn’t want to say that this specific rate would be repeated, necessarily. Yet there were certain risk factors I observed in the civilian casualty incidents that I would expect to continue to be factors unless steps were taken to mitigate them.

Larry’s most recent report, Improving lethal action: learning and adapting in US Counterterrorism Operations, is available here. It includes an analysis of the Uruzgan air strike that is central to my ‘Angry Eyes’ essay (next on my to-do list).

[The short clip above is from Baden Pailthorpe‘s stunning animation MQ-9 Reaper (That Others May Die) (2014) – you can find much more here]

You might think that all of this is now of historical interest since President Obama has declared the end of the Afghanistan war. Not so. Here is John Knefel writing in Rolling Stone this week:

Though many Americans may not have realized it, December 28th marked what the U.S. government called the official end of the war in Afghanistan. That war has been the longest in U.S. history – but despite the new announcement that the formal conflict is over, America’s war there is far from finished. In fact, the Obama administration still considers the Afghan theater an area of active hostilities, according to an email from a senior administration official – and therefore exempts it from the stricter drone and targeted killing guidelines the president announced at a major speech at the National Defense University in 2013.

“Afghanistan will continue to be considered an ‘area of active hostilities’ in 2015,” the official tells RS. “The PPG does not apply to areas of active hostilities.” (PPG stands for Presidential Policy Guidelines, the formal name for the heightened drone rules.)

That perplexing distinction – that formal combat operations are over but that the U.S. still remains in an armed conflict – in many ways exemplifies the lasting legacy of Obama’s foreign policy.

If you assume the situation in Pakistan is somehow less ambiguous, read Ryan Goodman on ‘areas of active hostilities’ over at Just Security here (I’m having to sort all this out for ‘Dirty Dancing’, of course).

Second, the Bureau of Investigative Journalism has released its end-of-year report on US drone strikes in Pakistan, Yemen and Somalia in 2014, which includes these tabulations of casualty rates for the first two countries:

The Bureau comments:

While there have been more strikes [in Pakistan] in the past six years, the casualty rate has been lower under Obama than under his predecessor. The CIA killed eight people, on average, per strike during the Bush years. Under Obama, it is less than six. The civilian casualty rate is lower too – more than three civilians were reported killed per strike during the past presidency. Under Obama, less than one. There were no confirmed civilian casualties in Pakistan in the past year, as in 2013….

The frequency of strikes [in Yemen] may have fallen in 2014 but more people were killed, on average, per strike than in any previous year. The casualty rate for last year even outstrips 2012 – the bloodiest year recorded in the US’s drone campaign in Yemen when at least 173 people were reported killed in 29 strikes. In 2014 at least 82 people were reported to have died in just 13 strikes.

You can find the Long War Journal‘s tabulations for Pakistan here and Yemen here.

Third, Israel. I’ve commented previously on an interview with an Israeli drone pilot, but it’s been difficult to put his observations in context (though see here and scroll down to the tabulations). Now Ann Rogers, who wrote Unmanned: drone warfare and global security (Pluto, 2014) with John Hill – as good an introduction to drone wars as you will find – has just released an essay on ‘Investigating the Relationship Between Drone Warfare and Civilian Casualties in Gaza‘. It’s in a special issue of the open-access Journal of Strategic Security 7 (4) (2014) on ‘Future challenges in drone geopolitics’. Here’s the abstract:

Unmanned aerial vehicles (UAVs), better known as drones, are increasingly touted as ‘humanitarian’ weapons that contribute positively to fighting just wars and saving innocent lives. At the same time, civilian casualties have become the most visible and criticized aspect of drone warfare. It is argued here that drones contribute to civilian casualties not in spite of, but because of, their unique attributes. They greatly extend war across time and space, pulling more potential threats and targets into play over long periods, and because they are low-risk and highly accurate, they are more likely to be used. The assumption that drones save lives obscures a new turn in strategic thinking that sees states such as Israel and the US rely on large numbers of small, highly discriminating attacks applied over time to achieve their objectives. This examination of Israel’s 2014 war in Gaza argues that civilian casualties are not an unexpected or unintended consequence of drone warfare, but an entirely predictable outcome.

It’s an interesting essay, but I fear that it takes the Israeli military at its word. Ann repeatedly refers to Israel’s ‘discriminating’ targeting:

‘The central point is that drones enabled the IDF to undertake detailed, extensive, and discriminating targeting of Gaza, before and during the actual fighting. The killing of civilians may be down to differing interpretations of military necessity, or in some cases, in how combatants and non-combatants are distinguished from one another. But it is the drone gaze that enables these targets to be ‘called into being’ (p. 102)…

‘As Israeli targeting of Gaza appears to have been highly discriminating, a more serious problem may lie in how its view of legitimate attacks differs from the global “norm.” (p. 104).’

I commented on Israeli attacks on hospitals and ambulances last summer here, here and here, and on the wholesale destruction of Gaza here and here, so I confess I am at a loss for words. But she is right to emphasise the operative power of international humanitarian law and its protocols of distinction (discrimination) and proportionality – though, as often as not, these seem to have been inoperative in anything other than a rhetorical sense. For much more on this, and the way in which military lawyers are incorporated into Israel’s kill-chains, you should click across to Craig Jones‘s War, Law and Space. All of which makes the Palestinian decision to seek membership of the International Criminal Court all the more important (there’s a good commentary on the wider legal issues by David Luban at Jus Security here and by a clutch of commentators at the Middle East Research and Information Project‘s blog here). Perhaps not surprisingly, Daniel Reisner, the former head of the Israeli military’s International Law Department, has condemned the Palestinian application as ‘a belligerent act within the framework of the non-physical and kinetic world of lawfare.’

Finally, the US-led air strikes on IS/ISIL targets in Iraq and Syria. Here we know much less than we should, not least because the Pentagon knows much less than it should. Here is Nancy Youssef reporting earlier this week:

In a war fought largely from the air and in places no one can safely go, the impact is as opaque as the war itself, making it difficult to measure whether the U.S. and coalition effort is working.

“We don’t have the ability to count the nose of every guy we schwack,” as Pentagon spokesman Adm. John Kirby told reporters Tuesday, using military jargon [sic] for killing. “That’s not the goal.”

Presumably, that also means the Pentagon can’t count how many civilians it has accidentally killed in the name of ridding the region of ISIS.

Drone Wars UK has an excellent survey of the logistics of air operations over Iraq and Syria from Chris Cole here, and the New York Times has produced a useful interactive map of US-led air strikes from which I’ve snipped this summary:

We don’t know how many of these were carried out by drones or even orchestrated by them, but as their limitations are becoming clearer it’s reasonable to assume that most involved conventional strike aircraft. We do know that targeting involves the analysis of video feeds from both remote and conventional platforms, and that CENTCOM has had considerable difficulty in juggling the competing demands for ISR from Afghanistan and from Iraq/Syria.

According to a report this week from W.J. Hennigan, who visited the USAF’s 480th Intelligence, Surveillance and Reconnaissance Wing at Langley Air Force Base in Virginia:

In a vast windowless room, several dozen intelligence analysts worked under the glow of more than 100 computer screens, quietly studying video streaming from U.S. drones and spy planes hunting for Islamic State militants in Iraq and Syria.

One team searched the incoming video to find a firefight underway between Iraqi security forces and militants somewhere south of the insurgent-held city of Mosul in northern Iraq.

For four hours, the analysts pored over the imagery before identifying 20 positions where the militants were dug in with machine guns and other weaponry. After the analysts called in the coordinates, 15 jets from five countries pounded the targets with more than two dozen bombs.

The Dec. 5 airstrike, one of 462 last month, underscores the Pentagon’s increased reliance on personnel far from the battlefield… Air Force analysts here stand — or rather sit — on the virtual front lines by tracking Islamic State fighters in a war zone some 6,000 miles away.

But here’s the rub:

Unlike in past wars, when U.S. troops on the ground helped provide targeting information and intelligence, commanders in the battle against Islamic State rely chiefly on airborne surveillance, captured communications chatter, signals intelligence and other material that is processed by analysts here.

U.S. officers said the video-watching analysts working half a world away are no match for spotters and other troops feeding intelligence from the front lines.

“We don’t have anywhere near the level of intelligence we used to,” Lt. Col. Marc Spinuzzi, a senior intelligence officer, wrote in an email from Baghdad. The analysts are under “a lot of pressure … to clearly distinguish friend from foe, and to pick out the enemy from the civilian population” on the battlefield.

That is precisely how mistakes are made and civilians killed. And, as Robert Naiman pointed out,

“There is a big danger here that U.S. air strikes in Syria are going to resemble the drone strikes in Pakistan and Yemen in the sense that there is no accountability for who is killed. We have reports of civilian casualties from people in the area and the U.S. government says, ‘No, they are bad guys.’ There has to be some public accountability for what happens when there are allegations of civilian casualties.”

At least the Pentagon has now gone some way towards recognising the problem. Previously it had insisted that it was unaware of any civilian casualties, which is disingenuous: it beggars belief that 1,000 air strikes could have resulted in no civilian casualties – but if your ISR is inadequate it’s scarcely surprising that you would be ‘unaware’ of the consequences. Even so, on 6 January the Pentagon announced that it had investigated 18 allegations of coalition airstrikes causing civilian casualties between 8 August and 30 December. It determined that 13 were ‘not credible’, but was continuing to review three others; a further two, one in Iraq and one in Syria, are now the subject of formal military investigations. But before you gold your breath, both Iraq and Syria are also exempt from the Presidential Policy Guidelines that require a ‘near certainty that no civilians will be killed or injured’. Here is Harold Koh (really):

‘They seem to be creating this grey zone… If we’re not applying the strict rules [to prevent civilian casualties] to Syria and Iraq, then they are of relatively limited value.’

Drones, militarized vision and civilian casualties

Posted on April 13, 2014 by Derek Gregory

I’m just back from the AAG Conference in Tampa, and there’s a lot to catch up on.

First, an art installation in Pakistan called #NotABlugSplat that reverses the paramilitary gaze and ‘targets Predator drone operators sitting thousands of miles away who refer to kills as BugSplats.’ Now they’ll see on their screens the face of a child who lost her parents and two young siblings in a drone strike.

It’s a collaboration between Pakistani and American artists, working with Reprieve and the Foundation for Fundamental Rights, who also designed it ‘to be captured by satellites in order to make it a permanent part of the landscape on online mapping sites.’

It’s an arresting project – but if you scroll through the comments that followed gizmodo‘s report you’ll see that ‘hope’ and ’empathy’ remain dismally distant for many people.

Perhaps some of them would benefit from a new report for the Center of Naval Analyses (CNA) by Larry Lewis, Drone strikes in Pakistan: reasons to assess civilian casualties. I’ve noted his (largely classified) work on Drone strikes and civilian casualties in Afghanistan before, and in this – unclassified – report he leverages what we know about US military drone strikes in Afghanistan to address the cross-border attacks directed by the CIA. Lewis makes two key points.

First, he notes that the US government’s claims about civilian casualties for its supposedly covert operations in Pakistan are significantly lower than ‘nearly every other estimate available’. (En route, he draws attention to something that is usually overlooked in these calculations: under International Humanitarian Law, ‘the burden of proof is to determine whether a casualty is a combatant’, and where in doubt the casualty must be regarded and recorded as a civilian). Based on his previous work in Afghanistan, Lewis suggests three overlapping reasons – apart from a disinclination to tell the truth – that ‘complicate the estimation process’:

An irregular enemy – it is exceptionally difficult to distinguish combatants from civilians in irregular warfare, and this is exacerbated by combatants ‘co-locating with the local population’;
Misidentifications – ‘US forces mistakenly believe civilians to be enemy combatants’; I’m not sure how this is different from the first, but Lewis provides two examples that suggest he has in mind specific rather than general characteristics: mistaking men digging drainage ditches for militants burying an IED, for example, or assuming all those in close proximity to an engagement were involved (‘guilt by association’);
Battle Damage Assessments (BDA) based on aerial surveillance – determining the consequences of an air strike without ‘boots on the ground’ is likely to be defective

Of all of these, Lewis suggests that it is misidentification that is likely to be ‘the basis for the majority of civilian casualty incidents’ and cites the case that I discussed in detail in Tampa: the strike carried out by two attack helicopters following persistent surveillance from a Predator of a ‘convoy’ of three vehicles in Uruzgan province in Afghanistan in February 2010. I’ll post my version of events shortly, since I think it is a mistake to collapse this episode into a monotonic ‘Predator vision’; there were other eyes in the sky [see the image below], and – still more significantly – military vision is not a uniquely technical process (which is why the concept of visuality is so important) and in this case involved different interpretations of the Full Motion Video Feed from the drone by different people at different locations. In short, there was a de-centralized, distributed and dispersed geography of militarized vision that was never resolved into a plenary (still less totalizing) frame.

That said, Lewis’s second point is about process not platform. He has no truck with claims like Avery Plaw’s – ‘Where civilian casualties cannot be avoided they must be minimized. This is what drone strikes do’ – because they mistake ‘platform precision for a comprehensive process that minimizes civilian casualties’ and are in fact ‘contradicted by operational data’. He cites his earlier analysis of 2010-2011 data from Afghanistan, which ‘showed that several forms of attack, including engagements by manned air platforms, were less likely to cause civilian casualties than drone strikes’ (my emphasis; see my earlier discussion here). In his view, then, ‘minimizing civilian casualties is less a matter of platform or ordnance selection as it is using an approach that considers factors that lead to civilian casualties and then effectively takes them into account.’

The point is sharpened by Mark Gubrud in a response to a report from Charli Carpenter at the Duck of Minerva:

‘…drones use the same targeting pods and precision-guided weapons as the manned platforms they replace; in fact, the quality of imagery from the drones is degraded by the limited bandwidth and frequent interruptions of satellite links, as well as the transmission delay which can frustrate last-moment aborts. On top of the “soda straw” vision as compared with low-flying aircraft in close air support, these factors mean that, if anything, the drones are actually inherently less discriminate.’

Again, all of these factors were in play in the Uruzgan attack: degradations and interruptions of both video and audio transmissions were of critical importance – again, see the image from my Tampa presentation above – but even more significant was the way in which the military field of view expanded, contracted and even closed at different locations as the episode unfolded. More to come.

Drones and civilian casualties in Afghanistan

Posted on July 3, 2013 by Derek Gregory

I’ve been urging for some time now that the debate over drone strikes must not neglect what has been happening in Afghanistan: hence “From a view to a kill” (DOWNLOADS tab) and the much longer version to appear in The everywhere war. According to a report on the Guardian website posted late on Tuesday by the resourceful Spencer Ackerman, ex-Danger Room (now sadly itself an ex-site), ‘A study conducted by a US military adviser has found that drone strikes in Afghanistan during a year of the protracted conflict caused 10 times more civilian casualties than strikes by manned fighter aircraft.’ The period under analysis (from mid-2010 to mid-2011) followed a series of measures announced by General Stanley McChrystal to reduce civilian casualties from air operations, and coincided with a dramatic increase in the number of drone strikes (which continued to increase through 2012).

The study in question was carried out by Larry Lewis for the Joint Coalition and Operational Analysis (JCOA), a division of the Joint Staff J7 whose work is summarised in the following slide (which comes not from Edward Snowden‘s cache but from here):

Notice ‘Civilian casualties in Afghanistan’ is #6 on the list of major studies. The completed study is called Drone Strikes: Civilian Casualty Considerations and, apart from the Executive Summary that was published on 18 June, remains classified. The opening paragraph is the primary source for Spencer’s story, and here it is:

‘The US government has described drone airstrikes in operations outside declared theaters of armed conflict as surgical and causing minimal civilian casualties. Analysis of air operations in Afghanistan, combined with a review of open-source reports for drone strikes in Pakistan, suggest that these fell short of intended goals. Specifically, drone strikes in Afghanistan were seen to have close to the same number of civilian casualties per incident as manned aircraft, and were an order of magnitude more likely to result in civilian casualties per engagement. Specific causal factors were identified that contributed to the relative propensity of drones to cause civilian casualties. Tailored training that addresses these causal factors could aid in reducing civilian casualties in engagements involving drones. While processes and operating forces in Afghanistan can differ from those in operations outside declared theaters of armed conflict, the factors above suggest that a dedicated analysis of civilian casualties in such operations would be worthwhile.’

The key sentence is in bold, and a version of it reappears in an essay, ‘Changing of the Guard: civilian protection for an evolving military’, that Lewis wrote with Sarah Holewinski, Executive Director of the Center for Civilians in Conflict and which appears in the latest issue of PRISM 4 (2) (2013), a journal of the Center for Complex Operations at the National Defense University (the press release from Civilians in Conflict is here). The authors’ central concern is that ‘as Washington shifts its focus from counter-insurgency to counterterrorism, and from large-scale ground operations to more discrete and oftentimes-unmanned operations, the progress U.S. forces have made on preventing and mitigating civilian harm may soon be lost.’

They consider the risks attendant upon the increased reliance on Special Operations Forces – something that readers of Jeremy Scahill‘s Dirty Wars will need no warning of – but also those that are likely to flow from an increased use of Unmanned Aerial Systems (UAS):

‘America’s use of force will increasing rely on new technologies, including air force capabilities to penetrate enemy defenses and strike over long distances. Unmanned Aerial Systems, sometimes referred to as “drones,” saw major use in Iraq and Afghanistan, and are slated for a big leap in funding. The Pentagon called for a nearly one-third increase in its fleet in the years ahead.

‘The use of UAS can have military advantages for avoiding civilian casualties in armed conflict, if used with that intent in mind. Their systems feature precision weapons, their sensors have increasingly high-resolution imagery to assess the ground situation, and back in the control room, trained imagery analysts scrutinize a target area prior to engaging, which isn’t always possible in a full ground operation.’

They list a series of familiar objections to the use of these remote platforms in areas outside ‘traditional combat theaters’, but they are also critical of claims about their forensic capacity inside war-zones:

‘The assumption that UAS strikes are surgical in nature is … belied by research on recent combat operations in Afghanistan. There, UAS operations were statistically more likely to cause civilian casualties than were operations conducted by manned air platforms. One reason was limited training for UAS operators and analysts in how to minimize civilian harm. adding or improving training on civilian casualty prevention is a resource decision in direct tension with the increasing demand for more uaS and more operations, since additional training on civilian protection means time must be taken from somewhere else including the mission itself.’

They don’t say much about the reasons for this, except that they then criticise the ‘clandestine use of UAS by the US government’ because it raises ‘significant concerns that civilian casualties will not be properly monitored or investigated and thus called into question’ and then, several paragraphs later, they note that non-covert operations in Afghanistan ‘are replete with examples where all the engaged individuals were believed to be combatants, but a later investigation found many or all were civilians misidentified as combatants.’

A key issue, then appears to be misidentification, and in my examination of one hideous incident in Uruzgan province in Afghanistan in February 2010 (in “From a view to a kill”; see also the image above) – an analysis I develop in my more detail in The everywhere war – I suggested that the high-resolution video feeds from these remote platforms engender an intimacy with ground troops that belies the physical remoteness of the drone operators. They routinely claim to be not 7,500 miles from the battlespace but just 18 inches – the distance from eye to screen – and this immersive capacity (which these feeds do indeed share with videogames) predisposes them to view virtually every proximate Afghan action as hostile:

‘… the greater incidence of civilian casualties when close air support is provided to ‘troops in contact’ may result not only from time-critical targeting and its correspondingly ‘fewer checks to determine if there is a civilian presence’ … but also from the persistent presence of the [UAS] and its video feeds immersing its remote operators in, and to some substantial degree rendering them responsible for the evolving situation on the ground. This predicament, in which proximity not distance becomes the problem, cannot be resolved by tinkering with the Rules of Engagement; high-resolution imagery is not a uniquely technical capacity but part of a techno-cultural system that renders ‘our’ space familiar even in ‘their’ space – which remains obdurately Other.’

I don’t know if this forms part of the classified report, of course, but in an earlier report, Reducing and mitigating civilian casualties: enduring lessons (dated April 2013) Lewis emphasised the importance of using ‘discrimination tools’ in ‘situations where forces need to discern whether an individual is demonstrating hostile intent’ – but what he seemed to have in mind was another technological fix, ‘higher-resolution imagery or night vision devices’, whereas the root of the problem may well not be the power to see but the capacity to make sense of what is seen.

In any event two other questions remain.

First, during the period under analysis drone strikes accounted for around 5 – 6 per cent of total weapon releases by all aircraft in Afghanistan, but many of the conventional strikes nevertheless relied on persistent surveillance of targets from Predators or Reapers and then attacks by helicopters or fighter-bombers (which was the case in the Uruzgan attack). Does Lewis’s statistical analysis shed any light on the difference (I assume there is one) between a UAS acting as a ‘hunter-killer’ and a UAS providing only real-time ISR as part of a networked operation?

Second, what is the difference between ‘an incident’ and an ‘engagement’ in the first extract I’ve quoted? This is a substantive issue of considerable moment: if drone strikes produce a roughly similar number of civilian casualties as conventional strike aircraft ‘per incident’ but ten times the number ‘per engagement’, it’s vital to know the difference. Protagonists of remote operations will undoubtedly seize on the first, critics on the second.

I’m trying to chase down the difference. To be continued…