Grim Reaper: an RAF drone arrives at Kabul airport. Fg Off Owen Cheverton/MOD / Wikimedia. UK Open Government licence.In late 2001, Afghanistan was the site of the first armed drone strike and 13 years later it is the most heavily drone-bombed country in the world. But remotely-piloted operations there are shrouded in mystery: there is more knowledge of the secret US drone wars—the campaigns waged in Pakistan, Yemen and Somalia by the Central Intelligence Agency and its similarly mysterious military cousin, the Joint Special Operations Command—than of the drone campaign in Afghanistan.

Figures released to the Bureau of Investigative Journalism showed that by November 2012 more than 1,000 drone strikes had occurred in the country—more than in Pakistan, Yemen and Somalia combined. The UK is a major partner in Afghanistan’s drone war: despite having a tiny fleet of Reaper drones, albeit recently doubled in size to ten, it has carried out more than 300 strikes.

Yet a new study by the bureau shows that there is no systematic, public accounting of how remotely piloted aircraft are being used in the conflict. The little information that is available comes in scraps which must be pieced together.

Increasingly important

Drones are an increasingly important element of the Afghan air campaign. Data for the UK recently showed that almost 80% of its missiles came from remotely piloted aircraft, while figures released to the bureau of air strikes by members of the International Security Assistance Force (ISAF) showed that in 2012 drones accounted for 18% of missiles fired from aircraft—up from 11% in 2009.

The detail of those strikes—when and where they occurred and who they killed—remains largely a mystery.

The US has repeatedly sought to present remotely piloted aircraft as a highly accurate weapons platform. In April 2012, the president’s then counter-terrorism adviser, John Brennan, said: ‘‘With the unprecedented ability of remotely piloted aircraft to precisely target a military objective while minimising collateral damage, one could argue that never before has there been a weapon that allows us to distinguish more effectively between an al-Qaeda terrorist and innocent civilians.”

But there are troubling indications that these promises may not be borne out on the battlefield. The United Nations Assistance Mission in Afghanistan (UNAMA) investigates all claims of civilian deaths in the country. In 2013, it reported that civilian drone fatalities had tripled to 45 and accounted for 38% of all deaths in air strikes—a higher proportion than those associated with helicopter or fixed-wing attacks.

Separately, last year Dr Larry Lewis, senior research scientist at the non-profit research centre CNA, analysed classified ISAF data for Afghan air strikes and found that remotely piloted aircraft were “an order of magnitude” more likely to cause non-combatant casualties on each engagement than other aircraft—which questions the conventional presentation of drones as highly accurate.

Afghanistan differs from the covert drone wars in that two air forces—those of the US and the UK—operate armed drones alongside one another. But there is a failure of accountability. The UK has consistently insisted that it is aware of only one incident in which British-piloted drones killed non-combatants—a strike in which four farmers died in March 2011. But there has been no public acknowledgment from the US of any responsibility for the 45 civilians UNAMA believes were killed last year.

Accountability gap

Air strikes are usually described as “NATO air strikes”, no matter which force conducted them, but for drones the accountability gap is particularly stark: either of two states might have launched the strikes that killed civilians, yet one specifically and repeatedly denies responsibility while the other remains silent.

In the covert wars, particularly in Pakistan and Yemen, there is also official silence. But the bureau and others have been able to develop a detailed, strike-by-strike record of how drones and other air power are used. This has been done primarily using open sources. Media reports, court affidavits and field investigations by NGOs, journalists and the bureau have all combined to offer a nuanced and revealing picture of the secret campaigns.

Over the past couple of months, the bureau has carried out a study examining whether it could adapt the research techniques it developed for monitoring those covert conflicts to track drone strikes in Afghanistan. Researchers gathered all the open-source information they could relating to reported air and drone strikes occurring in September 2013—a month selected partly because UNAMA had identified one strike as being of particular concern. They also contacted ISAF and research organisations working in the country to see what records they held and what they were prepared to release.

The team identified 34 reported air strikes that month—more than one a day. Ten were described as drone strikes. ISAF did not release casualty data but it did provide figures showing that 232 missiles had been launched by aircraft (including drones) in the month, suggesting that many attacks may be going unreported.

Of the strikes that were reported, half were reported by a single source and several more were corroborated only by a source which the bureau had previously found to be unreliable, the Iranian outlet Press TV, so that 60% of reported strikes were effectively single-sourced.

There is simply not enough robust evidence to draw any reliable conclusions but here, again, there is room for concern about civilian harm. Attacks reported as drone strikes were more than three times as likely to include claims of civilian casualties as those described simply as air strikes. But without further research it is impossible to know whether this reflects reality on the ground.

The small glimpses we have seen of official recording also raise the prospect that civilian deaths are being under-reported by the military authorities. This would not only affect the scope for accountability and acknowledgement but also, more practically, compensation and condolence payments for the bereaved and wounded. It would also impede the military’s ability to learn lessons where strikes have gone wrong.

The bureau obtained ISAF’s casualty records for September 2013. These showed that the authorities recorded one “close air support” incident in which civilians were killed. This appears to refer to a strike in Kunar province which, after exhaustive investigation, UNAMA found had killed ten civilians. Yet ISAF’s data record just three civilian deaths.

Previous reporting by John Bohannon of Science on ISAF’s casualty-recording processes indicates that this was not an isolated incident of disagreement between the UN and ISAF. Bohannon found that for reports of air strikes in 2009 and 2010 ISAF’s casualty estimate was 74% below that of UNAMA.

Independent scrutiny

In the absence of a comprehensive public record, it becomes all the more important that those operating armed drones become more open about their use of the platform, particularly in releasing data capable of independent scrutiny, in so far as this can be done without jeopardising current operations.

Previous ISAF commanders have recognised the vital importance of being as transparent as possible about civilian casualties in particular. In 2010, shortly after taking over command of ISAF operations, General David Petraeus published guidelines on how to turn around the counterinsurgency campaign, dogged at the time by persistent reports of civilian harm.

He urged troops: “Be first with the truth … Pre-empt rumours. Get accurate information to the chain of command, to the Afghan people and to the press as soon as possible. Acknowledge setbacks and failures, including civilian casualties, and then state how we’ll respond and what we’ve learned.”

Four years on, the lack of official transparency sustains a void of comprehensive casualty recording. This must be addressed if a key frontier of drone warfare is to be understood.

The bureau’s study of drones in Afghanistan is supported by the Remote Control project, an initiative of the Network for Social Change.