(Editor’s note: This article is part of a mini-forum on U.S. military operations in Somalia and civilian casualties. Other articles include analysis by Daphne Eviatar and by Luke Hartig.)
March feels like a decade and a thousand COVID-19 news cycles ago, so you could be forgiven for having missed the news that on March 31, U.S. Africa Command (AFRICOM) publicly committed to issuing a quarterly public report on civilian casualties alleged to have resulted from U.S. military operations. According to the press release, the report will include “any new allegations and updating the status of assessments that have been closed or remain open.”
According to AFRICOM Commander Steven Townsend, the change resulted from a year-long review of the Command’s civilian casualties procedures. Alone, a quarterly report is unlikely to resolve longstanding concerns held by human rights groups with AFRICOM’s investigations of alleged civilian casualties, nor even longer-standing concerns with the U.S. military’s expansive view of who can be targeted. Even so, the announcement should be welcomed as a small but important step toward greater transparency.
This article explores the circumstances under which civilian casualties may occur in Somalia, the site of the majority of civilian casualty incidents in AFRICOM’s area of responsibility; why they may be under-reported; and what to do about it.
The War on Terror: Going Local
That a quarterly report from AFRICOM is even necessary provides a reminder that the U.S. military is, for all intents and purposes, at war in Somalia, using force against both Al Shabaab and ISIS. New America reports that the U.S. has conducted 49 airstrikes in Somalia in the last six months. Al Shabaab is one of the few groups that the White House publicly acknowledges it includes among the “associated forces” of Al Qaeda, and against which it uses lethal force on the basis of the 2001 Authorization for the Use of Military Force (AUMF). At least parts of Somalia are also reportedly designated as “areas of active hostilities” by the administration.
But over the last few years, the U.S. campaign against Al Shabaab in Somalia has been increasingly framed as part of a political strategy to support the Somali Federal government in its attempt to consolidate power and legitimacy. In Somalia, much like Afghanistan, what started as a series of “targeted killings” of “high value targets” as part of a “global war on terror” has become an intensified effort that employs lethal force to buy time and space for local governance to take hold. (Although AFRICOM still periodically expresses its concerns about Al Shabaab’s intention to attack the U.S. homeland.)
Within that context, the lethal operations in which U.S. forces may be involved, and thus the circumstances under which the U.S. could commit civilian harm in Somalia, take three distinct forms.
First, the U.S. continues to conduct its own targeted strikes against Al Shabaab members, such as the strike on April 2 against Al Shabaab senior leader Yusuf Jiis. Second, the U.S. also conducts strikes under what it deems “collective self- defense” of Somali security forces, such as the most recent strike on April 10, in cases when Somali forces anticipate or enter into hostilities with Al Shabaab or other “affiliated” armed groups. (Concerns about “collective self- defense” related to the risk of exposing U.S. forces to hostilities in circumstances or against groups not authorized by Congress have been addressed in earlier Just Security posts.)
Finally, the U.S. conducts “advise, assist, and accompany” activities, primarily with the Somali “Danab Advance Infantry Brigade” through some combination of Special Operations Forces and U.S. State Department-funded contractors.
Each kind of activity affects the prospects of civilian harm, the control of US forces over the situation, and the U.S. military’s reporting of harm, in different ways.
“Targeted Killings”: The Fog of Precision Warfare
The government’s public acknowledgement of civilian casualties in the first category — the so-called “targeted killings,” i.e. pre-planned strikes against specific “high-value” targets — have been very rare in any of the places where the United States conducts this type of operation, which includes in Libya, Pakistan, Somalia, and Yemen. The risk of under-reporting civilian casualties in these kinds of operations stems from a combination of government secrecy, the government’s faith in “precision warfare,” positive identification errors, and a significant burden of proof for establishing civilian casualties. Each of these variables apply in Somalia.
The government’s confidence in the precision of planned operations is not entirely misplaced. Although reported policy changes to targeting parameters may now allow for the targeting of “foot soldiers,” and not just senior leaders in Al Shabaab (a change described by Luke Hartig on this site), the resources needed for pre-planned operations normally limit their use to circumstances where the United States has a higher degree of confidence, based on its own intelligence, that it has correctly identified a member of an armed group that it may target under domestic law and the laws of armed conflict. These operations incidentally afford more time for greater precaution in the targeting process itself, such as the choice of ordnance, the trajectory of a projectile, and assuring that civilians such as family members are not present.
During a planned strike, the military personnel who are involved are expected to report any unplanned civilian casualties they directly observe, for example, if a civilian on a motorbike suddenly rides past a targeted vehicle at the last moment or the body of a young child is seen removed from a bombed building. At present, the only civilian casualties admitted by AFRICOM came as the result of an internal review led by the former Commander, General Waldhauser, which surfaced an incident of harm during a planned operation that had never been properly elevated.
The Center for Civilians in Conflict (CIVIC), alongside many human rights organizations and other experts, has long cautioned against overestimating the precision of planned operations or assuming that the military’s own self-reporting will capture any and all errors. For one thing, while the U.S. military takes more precautions against civilian harm than many other countries in its planned operations, this very care taken in planning can actually introduce the risk of confirmation bias in the results, making it less likely that the military will be later willing to second-guess its work when challenged with evidence by outside groups. For example, while intelligence, surveillance, and reconnaissance (ISR) can be critical to avoiding civilian casualties, overhead imagery is still subject to the limitation that it does not capture anything beneath the roof of a building.
The military’s scorecard on checking its own work has improved thanks to processes put in place in both AFRICOM and for Operation Inherent Resolve in Iraq and Syria to fully assess any external reports and to cross reference them with documented military activity (a process described in full in our recent report.) In other theaters, the U.S. military has conceded “collateral” harm from its operations retrospectively when outside organizations, like Human Rights Watch and Amnesty International, have brought credible information to its attention, especially when this information includes documentary evidence such as photographs or video. But by the military’s own admission, the assessment process imposes an high standard for determining that an external report is “credible.” (The effect that this unduly high burden of proof has had on overall credibility assessments and casualty estimates has been covered by Ryan Goodman here and elsewhere.) According to Chris Woods of Airwars who weighed in by e-mail, “Until now there’s been little confidence in an opaque AFRICOM assessment process, which has rejected hundreds of locally claimed civilian deaths from US actions since 2007.”
Positive Identification Errors: Intelligence and Status Determinations
More complicated than the cases of collateral death or injury the military “misses” during planned strikes are the cases in which there are questions about the identity and combatant status of the dead and injured, which introduces two longstanding but unresolved sources of tension between outside groups and the military that persist in Somalia. First, outside groups and experts have raised alarms about the reliability of the intelligence used to make targeting decisions, whether human sources with political motivations, or through geolocational tracking of the “signature” of individuals suspected of certain affiliations. Outside groups and lawyers have also called into question whether the military uses an overly expansive definition of “combatant” in its targeting decisions or overly broad indicia for who counts as an enemy fighter in post-strike assessments; and they’ve raised concerns that public accounts suggest the military may make assumptions about the combatant status of individuals who are in very close proximity to a target (e.g. in the same car).
The fact that AFRICOM does not interview witnesses in its civilian casualty credibility assessment process further aggravates persistent doubts about the definitions the U.S. military uses in its targeting standards. Its critics allege that if the military does, in fact, make its targeting decisions on suitably defined assessments of who constitutes a combatant, and who is participating in hostilities, it should be more willing to entertain, if not actively seek, information that could exculpate the dead or injured from association with a terrorist organization. For its part, and from what I can glean from three years of speaking with military and civilian personnel in the Department of Defense, the military sees no compelling reason to talk about, or to, people who they confidently suspect are affiliated with a terrorist organization even if their suspicion rests on a pattern of behavior, and not specific information.
Civilian Harm and Partnered Operations: Collective Self Defense and Advise/Accompany
In cases where the United States is providing air support to Somali forces who come in contact with armed groups, the “luxury” of careful and time-intensive planning in the targeting may be lost, even if the military still adheres to the principles of distinction and proportionality in its strikes. In these types of operations, the military may target somebody it can directly observe is participating in hostilities (e.g. shooting at a partner Somali unit), perhaps making so-called “positive identification errors” less likely. Under similar circumstances in other theaters, however, military forces have mistaken the intent or behavior of civilians, which introduces the possibility of similar errors in Somalia. And the possibility that bystanders could be caught in the cross-fire cannot be dismissed, especially if Al Shabaab fighters embed themselves with civilians – a risk that global experience tells us probably grows with every battlefield victory won by the Somali government and its backers.
If Somali forces kill or injure civilians in partnered operations, U.S. military personnel may report the incident internally, or take note of public allegations; but they have no obligation, according to AFRICOM’s governing framework, to include these cases in the US military’s own public reporting on allegations of civilian casualties. The posture is to defer to the sovereignty and responsibility of the host government involved. According AFRICOM’s Feb 2019 Civilian Casualties Instruction, with respect to allegations involving partner forces:
“Considerations weighing in favor of utilizing the CIVCAS review process include, but are not limited to: the scope of U.S. force involvement in the partner force mission that resulted in an allegation of CIVCAS, the strategic partnership with the partner force against whom allegations are made, the relative capacity of the partner force to conduct an independent assessment, and the overall severity of the allegation itself.”
The same is true of cases where U.S. forces accompany Somali forces on patrols or raids, during which civilians may be harmed, as has been periodically alleged in Somalia. In these cases, the government of Somalia does have the primary state responsibility to prevent and account for the civilian casualties caused by its forces, but if it does neither (due to capacity or resolve), then, as NGOs have argued, the United States, as a significant benefactor and partner, should take steps to help the Somali government address gaps. Some contend that international law itself requires the US military to engage in such efforts when the actions of those partner forces may violate international humanitarian law, including failure to take feasible precautions to minimize the loss of civilian life.
Why now, and where do we go from here?
The nature of American involvement in Somalia, as in other parts of the continent, suggests that a new quarterly report from AFRICOM is unlikely to tell the whole story of civilian harm that has occurred either as a result of its unilateral actions or those that result from its role in partnered operations. Rather, the pronouncement probably reflects a strategic decision by leadership, aware of the benefits of early public disclosure, to counter false allegations and take hold of the public narrative around civilian casualties rather than letting others – from Al Shabaab to human rights organizations – shape it for them. The Command may also feel emboldened by a general belief that the U.S. government has a good story to tell overall, not only due to the care its forces take to avoid civilian casualties, but also because of the comparative moral legitimacy that the U.S. military itself assigns to its role in Somalia. To wit, the military often frames its lethal operations in Somalia not in terms of protecting the homeland against global terrorist threats, but in the language of protecting local civilians from terrorist groups like al Shabaab, much as it does in Afghanistan and Iraq.
Regardless of the motivation, observers should welcome this announcement as a change that could, and should, portend more meaningful measures designed to directly respond to the needs and concerns of Somali civilians themselves. As a start, AFRICOM should reconsider its practice of not seeking witnesses to interview during its assessments. The Departments of Defense and State should work together to ensure that Somalis can report civilian harm locally and safely, as well as to improve international coordination and response to civilian harm by working with AMISOM, the UN, and local NGOs to determine how best to overcome the very real security and access challenges Somalia presents. The U.S. military and State Department should also take steps to assist the Danab and the Somali government with their own capacity to prevent, track, investigate, respond to, and learn from civilian harm, as they should in other countries where the U.S. military is involved in “advise, assist, and accompany” activities. For its part, Congress, which has recently enacted several measures to improve transparency and oversight on civilian casualties generally, can set its attention to ensuring the Departments of State and Defense are well-coordinated, resourced, and expected to better prevent and respond to civilian casualties specifically in Somalia.
Finally, leaders from U.S. military and State Department can ensure that the full range of U.S. security cooperation activities are accompanied by consistent political messaging that without greater accountability on the part of U.S., Somali, and other international forces, civilian harm will pose an obstacle to the shared political goals of establishing a legitimate government capable of protecting the Somali people. In so doing, they can refer to the national experience of the United States. As former Secretary of Defense Robert Gates noted in his memoirs in 2014, “I don’t believe any military force ever worked harder to avoid innocent victims, but it seemed like every incident was a strategic defeat”.