People hold their belongings as they stand in front of a house damaged by the airstrikes in Azizabad.

© 2008 Reuters

January 15, 2009

Robert M. Gates
Secretary of Defense
100 Defense Pentagon
Washington, DC 20301-1000

Re: Investigation into US Airstrikes in Azizabad, Shindand district, Afghanistan

Dear Secretary Gates,

We write to you concerning the October 1, 2008 Executive Summary of the investigation by USAF Brig. Gen. Michael W. Callan into civilian casualties resulting from the US and Afghan engagement in Azizabad in Afghanistan on August 21-22, 2008 (Callan Report Summary), and, more generally, efforts to reduce civilian casualties from US airstrikes in Afghanistan.

As you may know, Human Rights Watch in August 2008 published “Troops in Contact: Airstrikes and Civilian Casualties in Afghanistan”, which documented civilian deaths and injuries from US and NATO airstrikes, as well as Taliban “shielding,” and made recommendations aimed at limiting civilian casualties. We appreciate the strong comments you made on September 17, 2008, in Afghanistan about the need for the US to do more to prevent civilian casualties from airstrikes. We also welcome the personal regrets that you expressed for the victims—even before the investigation was completed—and your promise to compensate victims quickly and to conduct timely investigations.

We recognize that positive operational changes have been announced, including the Tactical Directive issued on September 2, 2008, by the commander of International Security Assistance Force (ISAF). These include partnering ISAF forces with Afghan armed forces in all ISAF operations; demonstrating “proportionality, restraint, and utmost discrimination” in the use of firepower, including by making the greatest possible use of precision systems; having on-scene commanders make every effort to confirm that targeted houses are not sheltering civilians; minimizing the use of deadly force in “escalation of force” procedures against civilians through “tactics, techniques, procedures and training”; and, acknowledging civilian casualties immediately and rapidly conducting transparent investigations.

In a subsequent statement, Brig. Gen. Richard Blanchette, chief spokesperson for ISAF, has said that in addition to this directive, NATO commanders are under orders to consider a “tactical withdrawal” during an engagement with insurgent forces, rather than using close air support when civilians are present.

On November 12, 2008, Maj. Gen. Jeffrey Schloesser, the commander of the 101st Airborne Division and Regional Command East, told the media: “I’ve given direct guidance, and so has my boss to me, that if there’s any doubt at all that the enemy is firing from a house or building where there might be women and children, that we’ll just back off.”

On December 8, 2008, further additions were made to the tactical directive, including new reporting for escalation of force, as well as a rewording to emphasize proportionality and the need to justify the use of close air support.

Human Rights Watch welcomes these operational changes and statements from high-ranking military officers and looks forward to their implementation in Afghanistan, including by non-ISAF US forces.

We are, therefore, deeply dismayed by the Callan investigation and ensuing Report Summary on the events in Azizabad. Given the high numbers of civilian deaths, criticism from Afghan President Hamid Karzai, and concerns raised by various Afghan and international organizations, the Callan investigation was an important opportunity for the United States to demonstrate that it would accept responsibility for its mistakes. Instead of being an exemplary US investigation derived from a new operational mandate, the Callan Report Summary appears to be little more than a return to the discredited inquiries of recent years. It simply and summarily dismisses the methodology used in the investigations by the United Nations, the government of Afghanistan, and the Afghan Independent Human Rights Commission (AIHRC); rejects information provided by villagers by arbitrarily calling into question their motivations; effectively places responsibility for preventable civilian deaths on Taliban forces; and exonerates US forces of any wrongdoing.

Perhaps General Callan’s findings would have greater legitimacy if a declassified version of the full report were made public; however, since the US military has stated that the full report will not be released, we and others have no choice but to judge the findings based on the limited information provided. As such—despite this on-the-ground investigation into civilian casualties being more comprehensive than previous investigations—it represents a disappointing failure of accountability.

Callan Report Summary

According to the Callan Report Summary, on the night of August 21-22, 2008, US and Afghan forces entered Azizabad, Shindand district, Herat province, in order to kill or capture a “High Value Individual” named by the US as Mullah Sadiq. US and Afghan forces approaching the village came under fire. According to villagers from the area who spoke to Human Rights Watch, and the investigations by the UN and AIHRC, there was a firefight lasting 20 or 30 minutes. The US then called in close air support, which a public information official of the US forces in Afghanistan informed Human Rights Watch involved an AC 130H gunship and an MQ-9 Reaper UAV. Airstrikes lasted for two to three hours, and reportedly entailed the dropping of a 500-pound bomb from a drone and shellfire from the gunship’s M102 105mm howitzer and 40mm grenades.

For several hours after the airstrikes, the US conducted “site exploitation,” during which time evidence was gathered, two wounded civilians were removed for medical assistance, and five men were taken into custody. The detained men included three members of the Afghan National Police and two local residents. They were described as being Taliban members in the Bagram press release of September 2 (“Coalition forces complete Shindand investigation”). On August 29 four of the five were released without charge; the fifth spent three months in the US detention facility at Bagram, but has since been released without charge.

In a press release and media statements during the hours immediately following the operation, the US military denied that there had been any loss of civilian life in Azizabad. By the end of August 22 a press statement was issued by the US media center at Bagram air base admitting that five civilians had been killed and two wounded. Three separate investigations conducted in the days following by the UN Human Rights Office, the provincial and central government of Afghanistan, and the AIHRC concluded that from 78 to 92 civilians were killed during the operation, the majority of whom were children.

For several weeks afterwards, US military officials robustly rejected all three alternative accounts. An initial US military inquiry carried out by the Combined Joint Task Force 101 between August 23 and August 29 concluded that no more than five to seven civilians and between 30 to 35 Taliban had been killed. In various media interviews US officials suggested that the villagers were spreading Taliban propaganda. Human Rights Watch interviews with US military personnel, villagers, and non-military investigating officers suggest that this investigation may have been flawed by the lack of US access to the village, as well as a failure to recognize that large numbers of bodies were buried beneath the rubble of the 12 to 14 destroyed and damaged homes.

After the deaths were reported, there was wide public attention given to the discrepancies in civilian casualties, including strong criticism from President Karzai and Kai Eide, the head of the United Nations Assistance Mission in Afghanistan. The US did not appear to shift ground until video evidence of the casualties was released on various media outlets in early September. On September 7, 2008, General McKiernan ordered a follow-up investigation. Because this investigation was to be conducted independently of the forces directly involved in the operation and would include field inquires in Azizabad, it was hoped that it would provide a credible and detailed analysis of the tragic events of August 21, place blame where it fell, lead to appropriate disciplinary action, and result in operational changes that would avoid such events in the future.

Unfortunately, this has not happened. The Callan Report Summary accepts a larger US figure for the number of dead—33 civilians—but dismisses out of hand the much higher civilian death tolls reported by UN, AIHRC, and Afghan authorities. It addresses methodological issues—such as testimony from villagers and graveyard statistics—in a cursory and inaccurate manner, and fails to address the flaws in the earlier US investigations that rejected reports of significant loss of civilian life. The summary fails to acknowledge any possible mistakes in US intelligence gathering that preceded airstrikes. Finally, it suggests without presenting evidence that Taliban forces deliberately used civilians as “shields,” apparently to reach an unsubstantiated conclusion that the actions taken by US and Afghan forces were “in self defense, necessary and proportional”—and thus lawful under the laws of armed conflict.

We address each of these points in turn below. Our conclusions are based on interviews in Afghanistan with Azizabad village elders, members of the investigating teams of the Afghan Independent Human Commission, the UN, and the government of Afghanistan, as well as interviews and email communication with members of ISAF and the US military; our analysis of the reports by UN, Afghan, and US authorities; and the requirements of the laws of armed conflict, as set out in the Protocol Additional to the 1949 Geneva Conventions (Protocol I), which is broadly accepted as reflective of customary law:

1. Methodology

The Callan Report Summary does not recognize that the initial US investigations into the Azizabad incident were inadequate and inaccurate and that official press statements in which the United States denied any or significant loss of civilian life only exacerbated the outrage among Afghans. Nothing in the report summary suggests that any lessons have been learned from the flaws in the initial two US investigations and similar mistakes in previous cases—or that the same mistakes will not be repeated.

The Callan Report Summary calls into question the methodology of the investigations into the Azizabad incident carried out by the UN, the government of Afghanistan, and the AIHRC. It says: “Due to primarily relying on villager statements, limited forensics... their report lacks independent evidence to support the allegations of higher numbers of civilian casualties.” The summary does not adequately justify its dismissal of villager statements as independent evidence—implying that statements of local residents are inherently unreliable—despite receiving information that those statements had been cross-checked numerous times by experienced human rights officers. It also says that “the evidence was tainted by alleged witnesses’ interests in seeking financial, political and/or survival agendas,” yet provides no supporting evidence for this claim.

The summary states that the investigations by the UN, the government, and the AIHRC compiled civilian casualty lists that “were judged invalid.” Participants in these investigations told Human Rights Watch that all three lists were based on visits to up to nine grave sites, dozens of interviews and re-interviews with primary and secondary sources, and considerable video and photographic evidence of the casualties and the destroyed buildings.

The wide discrepancy in civilian deaths between the US death toll and those of the three other investigations may be partly explained by the counting of the dead at gravesites. The Callan Report Summary states that according to the “General Director of the Haj, in the GIRoA [Afghan government], all casualties in this incident would have been buried one body per grave”—an assumption affecting US calculation of civilian death tolls. However, representatives from the Ministry of Haj have told Human Rights Watch that this is not the case. Villagers and human rights investigators who visited the area consistently reported that there were multiple graves where more than one person had been buried; Human Rights Watch confirmed this in interviews with villagers and human rights officials involved in the investigations.

These reports were backed by officials in the Ministry of the Haj who were involved in the investigation and with Islamic scholars in Afghanistan who told Human Rights Watch that there is no prohibition in Islam on burying several bodies per grave, particularly when many people have been killed, and when there are body parts that must be buried.

The Callan Report Summary provided limited evidence for defining 22 of the deceased as “Anti Government Militia.” The AIHRC concluded that there were 13 men of fighting age who may have been involved in the firefight. Their investigators could not establish a link with insurgent groups. As the US investigation admits, several of the dead men were employed by the British company Armor Group, working as security guards for the US military in Shindand. Government officials and villagers have alleged that the operation was the result of misinformation resulting from tribal rivalries in the area. The Callan Summary dismisses these allegations, but military sources have told Human Rights Watch that they are credible. If so the US should urgently review its use of intelligence, particularly when calling in close air support. The tactical directive of December 8, 2008, stressed the need for the use of two corroborating intelligence sources—this should have been the case for the Shindand operation.

Only 15 weapons were confiscated during the US “site exploitation,” and of those, at least five were legally registered to the Armor Group contractors. The Callan Report Summary says that the presence of small numbers of rifles, a box of mobile phones, and some mines were evidence that all 22 men killed were “Anti Government Militia.” Such items are commonly possessed by individuals other than “Anti Government Militia.” Particularly given the occupations of several villagers and that all five men arrested for being members of the Taliban were released without charge, the evidence provided thus far on the presence of Taliban forces is problematic.

2. Legal Findings

The Callan Report Summary suggests that US forces at Azizabad appropriately used existing information regarding civilians in the vicinity and that civilian casualties resulted solely from Taliban actions, and thus summarily concludes that all US actions were proper.

The Callan Report Summary states that, “[u]nfortunately and unknown to the US and Afghan forces,” the Taliban insurgents chose positions near civilians [emphasis added]. Protocol I requires that commanders as a precautionary measure to spare civilians “do everything feasible to verify that the objectives to be attacked are neither civilians nor civilian objects… but are military objectives” (article 57(2)(a)(i)). The laws of armed conflict do not prohibit military forces from deploying or fighting in residential areas, but they are required to take all necessary steps to minimize harm to civilians in areas under their control.

In addition, had the US had intelligence from sufficient sources, they would have been aware that there was about to be a memorial ceremony in the village where the operation was taking place. The ceremony had drawn many civilians, including persons from outside the village. Reasonable precautions should have uncovered a large civilian presence in the village. It is, therefore, questionable that the close proximity of insurgent forces to civilians was “unknown” to US and Afghan forces; if it was unknown, then the quality of US intelligence was shockingly poor.

Protocol I prohibits parties to a conflict from using civilians to shield military targets from attack (article 51(7)). “Shielding” refers to intentionally using the presence of civilians to render areas or forces immune from military attack, and is a war crime. The Callan Report Summary states that insurgents “chose fighting positions in close proximity to civilians” [emphasis added]. The accompanying Central Command press statement about the Azizabad incident (“CENTCOM concludes Shindand investigation,” release number 080810-01, October 8, 2008) also noted that “this ruthless enemy routinely surround themselves with innocents.” The strong implication is that Taliban forces in Azizabad deliberately used civilians to shield themselves from attack.

Human Rights Watch’s research in Afghanistan has found a number of instances in which Taliban forces purposefully kept civilians from leaving their homes, using them as shields from coalition attack (‘Troops in Contact’, chapter IV). However, the Callan Report Summary provides no evidence whatsoever to support a claim that Taliban forces in Azizabad deliberately used civilians as “shields.” The mere presence of civilians “in close proximity to combatants” does not demonstrate that they were forced to be there.

The Report Summary’s implication is that the Taliban bear full responsibility for the loss of civilian life in Azizabad. Human Rights Watch has criticized the Taliban for using human shields and for deploying their forces in densely populated areas that placed civilians at unnecessary risk. However, unlawful actions by a defending force do not permit the attacker to ignore the civilian presence; constant care must be taken to minimize harm to civilians. Attacks must not be indiscriminate nor cause disproportionate civilian loss.

The Callan Report Summary determined that the US attack on insurgent forces in Azizabad were “necessary” and “proportional.” Protocol I requires that “in the choice of means and methods” of attack, “all feasible precautions” be taken to minimize incidental loss of civilian life (article 57(2)(a)(ii)). Protocol I additionally prohibits attacks in which the expected loss of civilian life and property are “excessive in relation to the concrete and direct military advantage anticipated” (article 57(2)(a)(iiI)). An attack should be cancelled when it becomes apparent that that the expected civilian loss will be disproportionate (article 57(2)(b)). Given what could be expected to have been known about the large civilian population in the village at the time, conducting airstrikes over several hours that destroy or damage 12 to 14 houses in the middle of the night makes high civilian casualties almost inevitable.

There is admittedly no simple method for determining whether a particular attack would likely cause disproportionate civilian loss or should have been halted for being disproportionate. However, an attack designed to target a mid-level enemy commander and causing (at most) 22 insurgent deaths, but coming at the expense of high numbers of civilian deaths, raises serious proportionality concerns. The length of the encounter— two to three hours—heightens concerns that the attack could have been stopped once the high civilian presence became clear. Simply concluding that the military’s actions were “proportional” without providing any basis whatsoever for reaching that conclusion does nothing to demonstrate that this conclusion is supported by the facts in this case.

Protocol I also prohibits attacks in which the weapon used is likely to strike combatants and civilians without distinction (article 51(4)). Human Rights Watch is thus concerned to learn that the AC 130H gunship’s 105mm howitzer was used in Azizabad, as confirmed to Human Rights Watch by Col. Greg Julian, head of media for US forces in Afghanistan. Its use under the circumstances should not be assumed to have been “necessary.” Human Rights Watch believes that area-effect weapons such as howitzers and other heavy artillery should never be used against targets in populated villages. These weapons have indiscriminate effects when used in populated areas because their blast and fragmentation radius is so large it puts civilians as well as combatants at risk. In Human Rights Watch’s investigation of Israeli artillery fire in Gaza in 2006 we demonstrated that reducing the safety margin from 300 meters from civilian homes to 100 meters was one reason for a spike in civilian deaths there (see “Indiscriminate Fire: Palestinian Rocket Attacks on Israel and Israeli Artillery Shelling in the Gaza Strip” at https://www.hrw.org/en/reports/2007/06/30/indiscriminate-fire). We would appreciate information on what if any safety margin US forces employ when using aerial artillery.

Recommendations

Given the discrepancies in the various public statements by US officials on the Azizabad incident and evident problems with previous US investigations into civilian casualties in Afghanistan, there is a clear need to review the methodology of the US military’s own investigations. We urge you to initiate a comprehensive review of the methods used in your post-incident investigations. This is important not just to obtain the facts in any given case, but to regain the lost credibility of the US in Afghanistan when it comes to civilian casualties caused by US airstrikes.

The Callan Report Summary recommends that “solatia payments” be made in Azizabad and that in the future such payments should be made to survivors and victim families prior to the completion of investigations. In our report “Troops in Contact,” we criticized the ad hoc nature of military compensation in Afghanistan and recommended that a unitary approach involving all foreign military powers be adopted. While we are pleased that payments are being made earlier than in the past, we are disappointed that there is no sign of a new joint policy among all international forces operating in Afghanistan.
In “Troops in Contact,” Human Rights Watch called on the US and NATO to take the following steps to reduce civilian casualties during military operations:

  • Ensure air attacks comply with the legal obligation to take all feasible precautions to minimize harm to the civilian population.
  • Stop using airstrikes in densely populated areas unless the intelligence is highly reliable, and the target has been visually identified. It is critical that US forces improve their assessments on the ground before they employ close air support, taking into account the risk of misinformation or disinformation from sources.
  • Refrain from using 105mm howitzers or similar area effect weapons against targets in densely populated areas.
  • Do not carry out airstrikes without an adequate Collateral Damage Estimate (CDE). Precision-guided low collateral-damage munitions should be used whenever possible, especially on targets in populated areas.
  • Thoroughly investigate the collateral damage and battle damage assessment processes to determine how they can be improved to reduce civilian casualties, and implement appropriate changes.
  • Provide accurate and timely information on civilian casualties in military operations.
  • Create an independent group of experts in partnership with the Afghan government to investigate civilian deaths in combat operations.
  • Take responsibility for civilian casualties when that is warranted and take appropriate disciplinary or criminal action against those responsible.
  • Provide timely and adequate compensation to victims of airstrikes. Create a unitary system that includes both OEF and NATO forces. Consider a system administered by UNAMA and monitored by the AIHRC.
  • The weaknesses in the Callan Report Summary call into question the depth of the Defense Department’s commitment to institute reforms that would reduce civilian casualties. We are concerned that unless such reforms are urgently implemented, more unnecessary civilian deaths and injuries will result, generating greater public outrage and hostility towards the presence of international forces. This will continue to undercut support for military efforts to provide civilian protection, which should be the primary goal of ISAF. Despite the commitments at high levels to reduce civilian casualties, unless proper procedures are put in place there is a high risk that additional US troops on the ground calling in more close air support will lead to a repeat of Azizabad and other errors, causing more and more civilian casualties.

    As noted, we deeply regret the Defense Department decision not to declassify and publish the full report of the Azizabad investigation. In the interests of bringing to public attention the investigation’s methodology, analysis, and findings, we urge you to reconsider that decision.

    Thank you for your attention to this matter. We look forward to your response and request a meeting with appropriate officials to discuss these matters further.

    Yours sincerely,

    Brad Adams
    Executive Director
    Asia division