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 Daniel Mahanty.

Once every few days for the past 15 months, New York Times reporter Charlie Savage has tweeted “Meanwhile in Somalia,” along with a press release or news account of a U.S. airstrike in the Horn of Africa. Coming amid a chaotic news period, the now-70-plus-tweet thread is a documentary record of how U.S. Africa Command (AFRICOM) has quietly ratcheted up its campaign in Somalia. According to Savage’s reporting in the Times, AFRICOM conducted 32 strikes through the first quarter of 2020, putting it on a pace to more than double the 63 strikes it conducted last year. But this surge in strikes has been coupled with a growing number of civilian casualty allegations from the media and outside groups.

Later this month, AFRICOM will try to account for the civilian death toll, when it releases the first of what it says will be a quarterly accounting of “the status of ongoing civilian casualty allegations and assessments.” In announcing the new report, AFRICOM commander Gen. Stephen Townsend touted his commitment to transparency and “protecting civilians from unnecessary harm.” And the report is a very positive move toward transparency. Nonetheless, it is unlikely to resolve many of the longstanding disagreements between AFRICOM and outside groups on how civilian casualties are assessed, the policies and standards that govern these operations, or the overall trajectory for U.S. military operations in Somalia. Indeed, much more transparency is needed – well beyond what AFRICOM can offer or Savage can tweet – if we’re to adequately understand what’s happening in Somalia and the toll it’s taken on civilian lives. And, indeed, if we hope to judge whether the military mission is worth it.

Slow Marching toward Transparency

AFRICOM’s report is the latest step in what has been a steady, if at times frustratingly slow, march toward transparency in Somalia. The first U.S. acknowledgement of a specific targeted counterterrorism airstrike outside of the hot war zones of Afghanistan, Iraq, or Syria came in Somalia. That was when AFRICOM announced the September 2014 killing of al-Shabaab’s top leader, Ahmed Abdi Godane. Since then, AFRICOM has led the way among the combatant commands in acknowledging strikes and providing context from its commanders on America’s shadow wars. The AFRICOM public affairs site is regularly updated with accounts of U.S. airstrikes, at times disclosing the targets of strikes, and usually stating whether the strike was a pre-planned operation against alleged high-value terrorists or a strike conducted in the collective self-defense of U.S. and partner forces. AFRICOM has also engaged in open dialogue with Amnesty International and other outside groups assessing its operations.

Reconcilable Differences?

Major disagreements remain between DOD and outside groups over just how many civilian casualties have resulted from U.S. operations in Somalia. These differences are unlikely to be resolved anytime soon. In April 2019, AFRICOM assessed that it had caused just two civilian casualties in 110 airstrikes over the previous two years; it has not acknowledged any confirmed civilian casualties in the 60-plus strikes reportedly conducted since then, including a strike earlier this month from which allegations of civilian casualties emerged. In sharp contrast, Amnesty International concluded that the U.S. military killed 21 civilians and injured 11 more in nine of the strikes in southern Somalia that the group investigated (AFRICOM’s lone acknowledged civilian casualties occurred in central Somalia). Beyond human rights investigators like Amnesty, groups that aggregate and analyze local and media reports of U.S. counterterrorism strikes – like the Bureau of Investigative Journalism (until 2019), New America, and Airwars – assess substantial numbers of civilian casualties since 2017. The estimates from these groups vary widely, and none of them conduct the rigorous case-by-case investigations that Amnesty does. Airwars, which consistently produces the highest estimates, says that the U.S. government may have killed up to 156 civilians in Somalia since 2017, while at the low bound, New America assesses a minimum of 15 civilian casualties since 2017.

The numbers on both ends of the spectrum strain credulity. To believe AFRICOM has killed only two civilians in more than 175 strikes in three years is to completely refute the rigorous investigations conducted by Amnesty, an organization with a long track record of credible work. But to believe the high bound estimates that strike-tracking groups report, one would have to believe that more than 10 percent of all those killed are civilians. This is highly implausible to anybody who has worked with the U.S. military in recent years and observed their professionalism and commitment to civilian protection. Indeed, while commanding forces in Iraq and Syria, Gen. Townsend blasted the high-bound estimates put forth by Airwars in a blistering op-ed in Foreign Policy magazine.

The fundamental tension in assessing civilian casualties hinges on how much to rely on witness reports versus classified sources and methods. Although specific procedures may vary by theater, according to the Director of National Intelligence (DNI), “U.S. Government post-strike reviews involve the collection and analysis of multiple sources of intelligence before, during, and after a strike, including video observations, human sources and assets, signals intelligence, geospatial intelligence, accounts from local officials on the ground, and open source reporting.”

An Obama-era Executive Order requires DOD also to consider credible reporting from non-governmental organizations. But at least in Somalia, AFRICOM does not conduct interviews of witnesses to the strikes (reportedly not ever), citing the infeasibility of visiting strike sites in Shabaab territory. And DOD is critical of open source reporting on strikes in Somalia, which often comes from local sites like somalimemo.net, which even Amnesty acknowledges has Shabaab sympathies. (Human rights groups note that such sites can still provide valuable information, like photographs.) The DNI states that this broad range of information sources means that, “the U.S. Government may have reliable information that certain individuals are combatants, but are being counted as non-combatants by non-governmental organizations.”

Amnesty and other human rights investigators, by contrast, rely heavily on interviews, and although they are not able to visit strike sites in Somalia due to security concerns, they still leverage an intensive methodology to corroborate witness accounts. Interviews are central to nearly any credible investigation, and AFRICOM’s lack of interaction with witnesses or the families of the deceased reveals a large hole in its methodology. (One study by Larry Lewis reportedly found that when Battle Damage Assessments were conducted by aerial video only, they missed civilian casualties in 19 out of 21 cases that were subsequently discovered by ground-led investigations.)

Yet the complexities of conducting interviews in certain theaters is a longstanding challenge. Human rights groups have suggested that AFRICOM establish an office in Mogadishu that could receive and evaluate claims of civilian casualties. AFRICOM has reportedly resisted this solution, noting that such an office could be a target for attacks and could lead to a flood of false allegations. AFRICOM instead updated its website to provide a mechanism for locals to report civilian casualties. Human rights groups noted the infeasibility of producing complaints in English from areas where civilians not only lack access to the internet but where al-Shabaab has been known to target anybody with a smart phone.

The best solution to producing accurate assessments of civilian casualties would of course be witness accounts coupled with classified information, with each source corroborating or refuting information from others until the most accurate conclusion possible under the circumstances is found. But DOD cannot share with human rights organizations sensitive information that could reveal sources and methods, and it is understandably wary of relying on interviews that it did not conduct itself.

Further deepening the gulf is confusion about just who counts as a combatant. In 2013, the White House pointedly refuted the prevalent media narrative that the U.S. government defaults to counting all military-age males in the vicinity of U.S. air strikes as combatants. Yet just last year, a former senior U.S. special operations officer with experience in Somalia seemed to suggest that this is exactly how combatants are identified. According to Amnesty, “Brigadier-General Donald Bolduc told Amnesty International that all military-aged males observed with known Al-Shabaab members, inside specific areas – areas in which the US military has deemed the population to be supporting or sympathetic to Al-Shabaab – are now considered legitimate military targets.” AFRICOM refuted Bolduc’s characterization, but such a statement from a former senior officer suggests that there remains deep confusion about either the policy or operations.

An Opaque Rulebook

Counting the dead in Somalia has become a methodological mess. But let’s set that aside for a moment and imagine that DOD and outside groups could agree to a reasonable range of potential civilian casualties. Even then, we would have no basis for assessing the significance of these figures and whether they’re consistent with U.S. policy, because we don’t know the rules that govern U.S. counterterrorism operations in the first place.

Depending on how one reads press reporting, it’s possible that one of three or four different policy frameworks might govern strikes in the country. Most accounts of Somalia counterterrorism operations in the Trump era note the President’s reported decision in March 2017 to suspend the more rigorous Obama-era drone strike rules for Somalia and Yemen. However, those same accounts rarely note that the rules were reportedly suspended for only 180 days, nor that shortly thereafter, AFRICOM’s then-commander, Gen. Thomas Waldhauser, affirmed he would still abide by high standards of civilian protection, including the requirement that commanders assess with “near certainty” that civilians will not be harmed prior to conducting a strike.

Then, in September 2017, President Donald Trump reportedly approved an updated policy framework governing targeted strikes, the Principles, Standards, and Procedures (PSP) which relaxed some of the Obama-era rules and policy oversight of operations but retained the “near certainty” rule. While the Obama administration publicized its new set of rules, coupling it with a major presidential address, the Trump administration has not made any public statements on the PSP. Although news accounts indicated that the new rules applied to Somalia, it’s unclear if there were any exceptions.

A December 2017 New York Times report said that DOD had developed an aggressive operational plan for Somalia that would include fewer checks on military action from civilian departments and a reduction in the “near certainty” standard for preventing civilian casualties to “reasonable certainty.” The Times story also raised the prospect that parts of Somalia were totally exempt from the PSP guidance. This is not without precedent; the Obama administration designated “certain portions of Libya” as “areas of active hostilities,” meaning that its rigorous Presidential Policy Guidance (PPG) on lethal action did not apply. One can imagine a similar carve-out being applied to Somalia, particularly in places where U.S. advisers are actively supporting Somali and African Union forces or, simply put, when the operational tempo does amount to active hostilities.

In short, any of these frameworks could currently govern U.S. operations in Somalia. Depending on which is in place, a given amount of civilian casualties could be assessed as being well within or dramatically outside of U.S. policy bounds.

Beyond these specific rules, it’s hard to make sense of U.S. military action in Somalia because there are just so many different types of operations taking place. Each of these, in turn, may produce different risks of civilian harm. Over the past dozen years, news accounts have reported on unilateral drone strikes, attack helicopter strikes, commando raids, and hostage rescue missions in support of U.S. citizens held by Somali pirate rings.

Since President Trump took office, the United States has conducted partnered raids with Somali and African Union partners, including a May 2017 operation that killed a U.S. Navy SEAL and an August 2017 operation that produced multiple allegations of civilian casualties and AFRICOM investigations. The U.S. military has an advise and assist mission supporting Somali and African Union forces that has reportedly grown to nearly 600 U.S. personnel (plus additional supporting forces in the region, like the aerial surveillance team that al-Shabaab attacked in Kenya in January). On many occasions, U.S. military forces have conducted collective self-defense strikes to defeat threats to U.S. and partner forces. (Members of Congress have asked for additional details on the nature of collective self-defense strikes and who might legally be struck in such operations.) DOD has also come under increasing scrutiny for its use of a fiscal authority (Section 127e), reportedly including in Somalia, that allows it to build up partners and proxy forces to conduct operations on behalf of U.S. forces. This has raised questions about these shadow forces and U.S. responsibility for the range of operations they conduct. And while we have no specific indication of the United States conducting offensive air strikes in direct support of its partners’ offensive operations (as opposed to collective self-defense strikes), such operations are at least conceivable under the reported December 2017 plan for more aggressive action in Somalia.

Each of these operations, if currently underway in Somalia, would entail different levels of danger to U.S. forces, involve partner forces at various levels, and produce scenarios with varying levels of risk of civilian casualties.

What Other U.S. Interests Weigh in the Balance

None of this is necessarily to say that we should change any of these operations. The U.S. military and its partners have made substantial gains against al-Shabaab, a gruesome terrorist organization, over the past decade. Al-Shabaab has been driven from Mogadishu and forced into the rural parts of the country. They have increasingly resorted to hit-and-run terrorist attacks in Somalia and Kenya, probably because they have not been able to control the amount of territory they once did. U.S. strikes have removed top operational figures, and others have been captured or surrendered. Although the Somali Federal Government is still very fragile and the Somali security forces have a long way to go, both have developed substantially out of the failed state the international community encountered a decade ago. Much more work is needed, and it’s not clear that Somalia will be able to stand on its own, without African Union and U.S. support, any time soon. At a time of bipartisan calls to draw down the “forever war,” we should scrutinize our operations in Somalia and whether they advance our core national interests. But without knowing the full range of operations currently ongoing in Somalia, and how each contributes to our end goal, we have no way of assessing the risks that we or Somali civilians face and whether it’s all worth it.

Conditions for Moving Forward

The solution is more transparency and dialogue, though likely beyond what Gen. Townsend is capable of providing due to matters outside his control. At a minimum, the AFRICOM report or Gen. Townsend’s engagement around it should explain with greater specificity than DOD’s annual report why AFRICOM’s assessments so often differ from those of outside groups. Ideally, AFRICOM would also address specific allegations that it disagrees with, but if that’s not possible, a broader discussion of the differences in methodology and where it thinks outside groups fall short would be helpful. AFRICOM should provide an update on its nascent effort to collect witness allegations on its website and what steps it is taking to better integrate witness accounts into its post-strike assessments. The Command should address head on the lingering confusion over military-age males and affirm that they are not presumed to be combatants, both in pre-strike targeting decisions and post-strike assessments. Gen. Townsend should clarify, as his predecessor did, what standard of civilian protection applies to U.S. strikes. AFRICOM should explain the range of operations underway and the risk profiles – to both U.S. forces and Somalis – for each kind of operation. And AFRICOM should provide a more thorough explanation of the gains the United States has made over the past several years, both in degrading al-Shabaab and putting Somalia on a path to provide for its own security. To be sure, there will be some details that AFRICOM cannot disclose due to the risk of compromising sources and methods or harming the operational security of our forces in the region, but it is well past time to bring more of this mission out of the shadows.

At a level above Gen. Townsend’s pay grade, top Pentagon and White House officials should explain the policy guidance in place for Somalia, particularly the thresholds for conducting strikes and standards for civilian protection. To the extent that the PSP is in place, DOD should clarify how that guidance applies to specific conflicts and under what circumstances exceptions could be and are made. They should publicly articulate the legal and policy frameworks governing operations in Somalia and elsewhere. And they should announce longer term efforts to improve oversight of strike operations and enhance mutual understanding with outside groups. I have previously recommended an outside advisory board that would include experts from a range of backgrounds, including the human rights community, to review U.S. direct action policy. Notably the 2020 National Defense Authorization Act directs DOD to contract with a federally funded research and development center (e.g., RAND Corporation) to produce an analysis of “standards, processes, procedures, and policies relating to civilian casualties resulting from United States military operations.” Among other things, this report must include “analysis of general reasons for any disparity between third party public estimates and official United States Government estimates of civilian casualties resulting from United States or joint military operations.” These would be important steps toward mutual understanding with outside groups and producing accurate assessments of civilian harm.

Unfortunately, however, this is a level of policy information that AFRICOM is probably not authorized and the Trump administration will almost certainly decline to release. After steady moves toward transparency across the Obama administration, the Trump administration has made clear that transparency is not a priority. The administration appears to be ignoring a statutory requirement to produce a report to Congress on the legal and policy frameworks for its use of force around the world. The White House has not even acknowledged the existence of the PSP, and last year, the administration did away with a provision in the Obama-era civilian casualty executive order requiring an aggregated report on global U.S. counterterrorism strikes. DOD’s general counsel used one of his major public speeches to knock down strawman criticisms from outside groups of DOD’s activities, instead of addressing legitimate critiques in good faith. His counterparts at the State Department and Justice Department have largely shied away from commenting on the legal frameworks governing counterterrorism operations, unlike their predecessors. Policy officials have steered clear of engaging publicly altogether.

Gen. Townsend seems to bring a sincere commitment to transparency and accountability, and his new report appears to be a very good development. But AFRICOM has a growing credibility problem on civilian casualties, and it will take a rich dialogue with outside groups, members of Congress, and AFRICOM and Pentagon officials to resolve it.

Photo credit: U.S. Army Gen. Stephen Townsend, commander, U.S. Africa Command, and Maj. Gen. Mpho Mophuting, director of general support services, Botswana Defence Force, meet Aug. 14, 2019. Townsend met with Chiefs of Defense from Liberia, Ghana, and Malawi Aug. 13-14. in Garmisch-Partenkirchen, Germany (Takisha Miller, U.S. Africa Command)