The U.S. military’s collateral damage estimation (CDE) process could incorporate real-world data to lessen civilian casualties (CIVCAS), according to analysts who have studied ways to reduce such harm.
In 2010, the Joint Staff’s Joint Civilian Casualty Study called for using real-world data to validate CDE tools and processes, a call made again in 2018 in a follow-on study, Civilian Casualties: We Need Better Estimates—Not Just Better Numbers, co-authored by Larry Lewis, one of the authors of the original Joint Staff study. Lewis is the director of the Center for Autonomy and Artificial Intelligence at CNA.
“The formal collateral damage estimation process, as rigorous as it is, has never been calibrated with real world data to test its accuracy in predicting operational outcomes,” per the 2018 study. “This could be remedied through a study that examines how well estimates match up with actual operational data. Are there particular kinds of situations where these estimates could be improved through a refined model? This can go both ways – such an assessment could show that there are some cases where the model is overestimating anticipated civilian casualties. Improving the model would then improve freedom of action in those cases, without a false concern about civilian casualties. At the same time, if there are cases where the U.S. is systematically underestimating likely civilian casualties, a recalibration of the process could remedy that error.”
Under CDE, targeting planners examine the possible effects of various weapons to propose recommendations to lessen collateral damage to comply with rules of engagement (ROE) and the law of armed conflict (LOAC).
Combatant commands use the guidance in the classified Chairman of the Joint Chiefs of Staff Instruction 3160.01C—No Strike and the Collateral Damage Estimation Methodology, a process modified since its adoption more than a decade ago.
U.S. Central Command has used the guidance of CJCSI 3160.01C during Operation Inherent Resolve (OIR) against ISIS targets in Iraq and Syria. Analysts perform civilian pattern of life (CIVPOL) analyses and use algorithms to inform the targeting process.
“Neither CJCSI 3160.01C nor the methodology it prescribes is an algorithm or computer program,” CENTCOM wrote in an email. “The methodology is a deliberate process employed by trained analysts. USCENTCOM analysts who are certified to use the methodology may use a variety of tools and resources to assist in the implementation of this methodology, including the Digital Precision Strike Suite (DPSS) Collateral Damage Estimation (DCiDE) tool. Such computer systems require direct input from trained analysts to produce CDE assessments.”
In 2019, a U.S. Air Force document said that “the joint community has mandated use” of the DCiDE tool “for collateral damage analysis of kinetic weapons.”
DoD has not released any information on the effectiveness of DCiDE in helping to lessen civilian casualties.
“I think it’s a great question, and we don’t really have an answer,” Lewis replied in an email when asked about DCiDE’s effectiveness.
Anthony Blinken, President-elect Joe Biden’s nominee to become Secretary of State, has said that he favors re-examining sales of U.S. arms to countries in the Middle East that have indiscriminately killed civilians. In addition, Avril Haines, Biden’s nominee to be the director of national intelligence (DNI), said in written answers to questions from the Senate Select Committee on Intelligence before her Jan. 19 confirmation hearing that she favors an executive order (EO) establishing a reporting requirement by the U.S. government on all civilian casualties caused by its agencies. That EO would fall along the lines of Executive Order 13732, United States Policy
and Pre- and Post-Strike Measures to Address Civilian Casualties in U.S. Operations Involving
the Use of Force, issued by former Pres. Barack Obama on July 1, 2016–an order which did not mandate annual reporting for all civilian casualties, but rather those in areas “outside of active hostilities,” such as Somalia.
EO 13732 called on federal agencies to “develop, acquire, and field intelligence, surveillance, and reconnaissance systems that, by enabling more accurate battlespace awareness, contribute to the protection of civilians” and to “develop, acquire, and field weapon systems and other technological capabilities that further enable the discriminate use of force in different operational contexts.”
There is no clear indication yet that Biden’s national security team will try to improve CDE, however.
“On why the formal CDE process hasn’t been calibrated with real world data, I would reply: why haven’t the scores of recommendations that have been made over the last ten years from all the work on civilian casualties been acted upon?” Lewis wrote in his email. “I think the answer is because there is no proponency for this kind of work, and no dedicated resources. The 2018 CJCS CIVCAS review…was approved by then [Defense] Secretary [James] Mattis, and yet very little has been done. The CDE tool issue is just one example of a larger pattern, the U.S. military saying it does ‘everything possible’ to reduce civilian harm but then not actually acting on known problems to tangibly reduce that risk.”
Calibration of DCiDE “with real world examples would be helpful and informative,” Lewis wrote. “Right now, we see a systematic problem where strikes on structures cause civilian casualties that were unanticipated in the strike approval process.”
One such case is the March 17, 2017 U.S. airstrike against two snipers on the second floor of a building in the al-Jadidah neighborhood of West Mosul, Iraq in which a Boeing [BA] 558-pound GBU-38 Joint Direct Attack Munition (JDAM) with a delayed fuse killed the snipers but also set off explosives planted by ISIS, thus causing the deaths of more than 100 civilians, DoD said in a 2017 after-action report. Inclement weather had prevented the use of full-motion video monitoring of the building for two days before the strike, and information provided to CENTCOM before the strike conflicts on whether there was, or was not, a high-risk of CIVCAS.
U.S. Air Force-Brig. Gen. Matthew Isler, the investigating officer for the March 17, 2017 incident and the then-director of OIR’s Joint Air Component Coordination Element, said in documents released last year by CENTCOM under the Freedom of Information Act (FOIA) that then-Army Brig. Gen. John Richardson, the Target Engagement Authority (TEA) at the time at Camp Erbil, Iraq “was unaware that over 100 civilians were in the structure.”
“Had the TEA known of the large number of civilians, his proportionality assessment would have been drastically different, and he would not have approved the engagement,” per Isler, who said that he advised a change in tactics, techniques and procedures to reduce possible “CIVCAS entrapment” casualties after the March 17 airstrike.
Targeting planners did not conduct an informal CDE of the building just before the airstrike because of the immediacy of the threat of the two snipers to Iraqi Counterterrorism Services (CTS) forces within 65 meters of the building and because CIVPOL had indicated that there was a low probability of civilians in the building, Richardson said in the FOIA-released documents.
The immediacy of the sniper threat “would have favored” a close air support (CAS) response, rather than artillery surface to surface munitions, like the Raytheon [RTX] M982 Excalibur or Lockheed Martin’s [LMT] M142 High Mobility Artillery Rocket System (HIMARS), according to the documents.
Because of the proximity of CTS forces to the building and the inclement weather, the use of direct fire weapons was also not possible, per the documents. Such direct fire options would have included the Lockheed Martin AGM-114 Hellfire, the Raytheon AGM-65 Maverick, the BAE Systems‘ Advanced Precision Kill Weapon Systems (APKWS), the MBDA UK Brimstone, and strafing using A-10 aircraft.
In the documents, the chief of operations/strike director for the Combined Joint Forces Land Component Command (CJFLCC)-OIR at Camp Erbil, said that the only options available for the airstrike would have been the GBU-38 or the Boeing GBU-54 Laser Joint Direct Attack Munition (LJDAM) with delayed fuses, the 285-pound Boeing GBU-39 Small Diameter Bomb, or low-collateral damage (LOCO) variants of the GBU-38 and GBU-54.
The strike director said that the targeting cell would typically reserve such LOCO munitions for any needed strikes on Category I protected structures, such as schools, mosques, and hospitals.
Dense urban environments will continue to pose challenges for U.S. military targeting, as in the March 17, 2017 incident.
“In this particular case, the U.S. negligently killed a massive number of civilians through not conducting proper CDE, assuming away what is a high probability of CIVCAS both by their own prior assessment and logic of urban warfare, and they conducted a disproportionate strike that caused foreseeable, predictable, and yet unintentional harm to civilians.” John Emery, the author of Probabilities Towards Death: Bugsplat, Algorithmic Assassinations, and Ethical Due Care, wrote in an email. “Due care would indicate the lowest possible munition size to minimize risk to civilians even if target destruction is not assured, as well as perhaps putting pilots at some reasonable low-level of risk to aid in CIVCAS mitigation.”