Air strikesAir Rules of Engagement

United States’ airstrike rules of engagement reviewed

In 2017, Action on Armed Violence (AOAV) recorded an alarming rise in civilian casualties from airstrikes. A large percentage of these were due to the US-led Coalition’s airstrikes in Syria and Iraq. Civilian casualties (deaths and injuries) from these airstrikes alone amounted to 2,864, accounting for 20% of civilian casualties from global airstrikes in 2017. Such levels of civilian harm from US-led Coalition airstrikes constituted a 228% rise from the previous year, when 873 civilian casualties were recorded from such strikes. Though levels of civilian harm decreased in 2018, it remained higher than in previous years, with 952 civilian casualties recorded. In total, AOAV has recorded at least 5,110 civilian casualties from US-led coalition airstrikes in Syria and Iraq between 2014 and 2018. Other sources suggest the toll could be even higher.

There are a number of estimates that seek to gauge the impact of the US-led Coalition’s strikes on civilians.  These range from the Coalition’s own admission of at least 1,190 civilian deaths in Iraq and Syria between 2014 and February 2019, [1] to local allegations of over 28,000 civilian fatalities.[2]  Perhaps the most comprehensive reporting comes from an organisation called, that monitors and assesses civilian harm from airpower-dominated international military actions. In addition to Syria and Iraq, civilian casualties from US airstrikes have also been reported in Afghanistan, Pakistan, Somalia, Libya and Yemen in recent years.

Given the levels of civilian harm these airstrikes have been shown to cause, AOAV has published this paper to help throw more light on the Rules of Engagement (RoE) governing the US’s use of airpower, and the way that US RoE impacts on civilian harm.

Airframes, airstrikes and casualties

In Syria and Iraq

In recent years, coalitions have become a preferred method of fighting. Yet whilst coalitions can increase the resources, insights and power of their members, they can also create hindrances in terms of minimising civilian casualties and the transparency of civilian casualty reporting. Members of coalitions, for instance, have different RoE as well as differing methods on how to assess, minimise and measure the civilian casualties that may have occurred during their attacks. Coalitions can also make it difficult to assess which state is responsible for which strike, and how many civilian casualties have resulted from each member. Such realities generate problems surrounding transparency, but also for providing an apology or ex gratia (condolence payments) to civilian communities impacted by coalition airstrikes.

These challenges are not inconsiderable. US, for instance, is just one of 79 partners in a Global Coalition to defeat ISIS in Iraq and Syria. The military element of this is known as the Combined Joint Task Force-Operation Inherent Resolve (CJTF-OIR), or simply Operation Inherent Resolve (OIR). As part of this operation, the US and its allies began conducting airstrikes against ISIS in Iraq and Syria in 2014 – a combined airforce that was to include the US, the UK, France, Iraq (in Syria), the Netherlands, Canada, Australia, Belgium, Denmark, Turkey, Saudi Arabia, the United Arab Emirates, Bahrain and Jordan.

As part of the Coalition, one of the US’s main contributions has been conducting airstrikes and air support; carrying out more airstrikes than any other member. The US has also admitted to far greater levels of civilian casualties, though it is likely that all members of the Coalition conducting airstrikes, including the US, have perpetrated more civilian casualties than they claim.

Casualties resulting from US strikes are not published by the Department of Defense (DoD), but instead are listed on the CJTF-OIR website with all US-led Coalition casualties, a process that however obscures the responsible state party. Since October 2014, the US Central Command (CENTCOM) has even stopped identifying which nations were involved in strikes in its daily summaries, and by March 2015 it had stopped distinguishing between US actions and those of its partners. This ability to ‘hide in the crowd’ means that states are unlikely to come forward to acknowledge civilian casualties.

Unlike the UK, US reports rarely detail the aircraft and munitions used during airstrikes. It appears that the popular aircrafts used by the US in recent years have been MQ-9 Reaper drones, the A-10 Thunderbolt II, the B-52 Stratofortress, B-1 Lancers, F15 Strike Eagles and F22 Raptors, among others. The B-52 took over the B-1’s heavy bombing role in OIR after the B-1s were rotated out of the region in January 2016.

The most common munitions employed by the US Air Force, particularly during 2017, when the Raqqa and Mosul campaigns occurred, were said to be various types of GBU-38s, GBU-31s and 1,000lb class GBU-32/B JDAMs. Aircrafts are reported to routinely carry 500lb GBU-12/B Paveway laser-guided bombs and GBU-58/B dual-mode Laser JDAM, as well. However, after two years of engagement in Iraq and Syria, the US were thought to be ‘digging into war reserves’ as stockpiles reduced – it has been indicated that weapons with a greater destructive radius were used in Mosul for this reason. The US’s engagement in Iraq and Syria has also seen an increased US use of GBU-39/B Small Diameter Bombs, a 250lb GPS-guided glide bomb and Hellfire missiles. However, the failure of the US military to identify bomb-types and aircrafts used in each strike reduces accountability and blurs responsibility for civilian casualties – it also provides less transparency in regard to the proportionality of the strikes carried out.

One case, from April 2017, saw the US drop a GBU-43/B Massive Ordnance Air Blast Bomb, the first time a conventional bomb of this size was used in combat. The MOAB (a 20,000lb bomb, named the Mother of All Bombs) was dropped on an ISIS cave complex in Achin District of Nangarhar Province, Afghanistan. The bomb had not been used in the past because of the payload and the danger its use could pose to civilians.  Such use of forced seemed to echo the strategy of President Donald Trump who, in his election campaigning, said he would bomb “the shit” out of ISIS.

Casualty numbers: Raqqa and Mosul

Since 2014, the UK based monitoring group Airwars estimates that the Coalition is likely to be responsible for between 7,465 to 11,837 civilian deaths overall in the war against ISIS,[3] out of more than 28,000 civilian fatalities alleged locally.[4] By 2018, more than half of the deaths recorded by Airwars had taken place in the vicinity of either Mosul or Raqqa. The civilian casualty figures produced by the Coalition are significantly lower.

The US-led Coalition has delivered 110,000 munitions from the air (split equally between Iraq and Syria), with most airstrikes – over 90% – dropped or launched during ‘dynamic’ situations – strikes that were not wholly pre-planned. Additionally, most airstrikes were launched at targets situated in urban areas.

During the Mosul campaign, the United States and Australia – the only two Coalition members to admit to having caused civilian deaths there – reported responsibility for 400 civilian casualties. In reality, the Coalition is likely to have caused far greater levels of harm; a Johns Hopkins University study, which surveyed 800 households, found that there had been 2.09 deaths per 100 people during the Mosul offensive.[5]

AOAV recorded casualties – combining both fatalities and injuries – from 2016 and 2017 in Raqqa city and Raqqa Governorate as well as Mosul and its surrounding Governorate (Nineveh). The figures show a drastic increase in civilian casualties in this period, which coincides with an increase in airstrikes by Coalition forces in and around both cities.

AOAV recorded seven air-launched ‘events’ on Raqqa city in 2016, resulting in 300 civilian casualties. In 2017, AOAV recorded a total of 195 air-launched events in Raqqa city, resulting in 2,200 civilian casualties. Raqqa Governorate witnessed 22 air-launched events in 2016, resulting in 497 civilian casualties. In 2017, the number of air-launched events jumped to 339 and resulted in 3,481 civilian casualties.

Mosul and Nineveh Governorate saw a similar increase in airstrikes and civilian casualties in 2017. In 2016, Mosul city was recorded as being a target for air-launched weapons on 35 occasions, resulting in an almost equal number of civilian and armed-actor casualties – 469 and 396, respectively. However, in 2017, as the number of casualty-causing air-launched attacks increased to 145, the number and percentage of civilian casualties increased alarmingly, with 2,944 civilian casualties compared to 331 armed-actor casualties.

This pattern is also discernible for the Governorate as a whole: AOAV recorded Nineveh being targeted 62 times by air-launched weapons in 2016, resulting in almost twice as many armed-actor casualties as civilians – 980 and 511, respectively. In 2017, this backdrop again changed drastically: Nineveh Governorate witnessed 173 air-launched attacks that resulted in 3,103 civilian casualties and 649 armed-actor casualties.[6] This data also shows that when airstrikes were launched outside the city it was less likely that civilians would be amongst the casualties. When airstrikes were launched outside of Mosul city in 2017, 159 civilian casualties were recorded killed or injured compared to 318 armed-actors. Civilians accounted for 33% of reported casualties, whilst in the city they accounted for 90%. In 2016, the disparity had been even greater, with just 42 civilian casualties recorded in Nineveh Governorate, excluding Mosul city, compared to 584 armed-actors, civilians accounted for 7% of the casualties recorded outside of the city.

The intensity of the campaigns to re-take Mosul and Raqqa in the months leading up to ‘liberation’ in July 2017 and October 2017, is reflected in the number of civilian casualties recorded by AOAV by month (due to all explosive weapons). In Raqqa city in July, August, September and October 2017, there were 457, 695, 280 and 289 civilian casualties, respectively. In November 2017, the number of civilian casualties dropped to 25, and in December 2017 there were 40. A similar pattern is discernible for Mosul city: In April, May, June and July (2017), AOAV recorded 904, 316, 392 and 56 civilian casualties, respectively, due to air-launched, ground-launched or Improvised Explosive Device (IED)-related attacks. The number of civilian casualties fell to one for August and three for September.

The number of civilian casualties from these campaigns gives rise to serious concerns surrounding US RoE for conducting airstrikes. Some particular incidents have highlighted the failings on the part of the US-led Coalition in assessing and accounting for civilian casualties as a result of strikes.

For example, in an airstrike in March 2017 on Mosul’s al-Jadidah neighbourhood, Iraqi forces called in an airstrike on an area where ISIS snipers were shooting from the roofs of houses. The US-led Coalition conducted airstrikes on these buildings using GBU-38 munition, 500llb bombs. The basements of these buildings were full of civilians sheltering from the strikes, and as many as 240 civilians were killed. The Coalition conducted an investigation into the strikes, and the Pentagon admitted the strikes killed at least 105 civilians – a lower figure than other estimates. The US CENTCOM iterated that they believed ISIS explosives were also present in the building and caused their collapse, though witnesses made no mention of secondary explosions. Instead they claimed such damage from the airstrikes was normal, indicating the other buildings in the neighbourhood that had also been destroyed by the airstrikes.

Whilst the US stated that they had been monitoring the area, the weather on the days just prior to the strike had prevented overhead surveillance. Civilians are likely to shelter from the raids for long periods of time and may therefore go unmonitored. This strike demonstrates the failure of the Coalition to adequately assess the levels of civilians present and monitor casualties caused post-strike.

Accountability and RoE

Casualty counting methods

Casualty recording can be initiated by self-reporting (internal reporting), where a rapid report outlines the basic details of the incidents, including estimated civilian casualty numbers, or it can be initiated through external reporting by NGOs.

Pilots and operators of the United States Air Force are encouraged to report proactively on individual missions, particularly when there are concerns that an airstrike has gone awry. However, there are challenges to detecting civilian casualties with the available military capabilities, particularly given the urban setting of many operations in Syria and Iraq and the lack of boots on the ground. An over-reliance on airborne sensors means that the US and other Coalition members are not able to reliably identify when civilian casualties occur or accurately assess levels of civilian casualties – those buried under rubble are highly likely to be missed.

When there are reports of civilian casualties, the next step is to corroborate the report with US strike logs. Assessment has been a formal part of the process conducted by the CJTF-OIR Civilian Casualty Report (CIVCAS) cell since roughly March 2017. Prior to that, the Initial Assessment was part of the CIVCAS credibility assessment report. If corroborated in the strike log, the report then moves to a credibility assessment that determines whether civilian casualties likely occurred.

There are some issues that arise in regard to casualty recording methods. For example, in some instances, the Target Engagement Authority who approved a strike that led to a CIVCAS allegation will be the same person who conducts or oversees the credibility assessment, a practice that can lead to bias.

In December 2016, the US established the permanent Coalition CIVCAS cell. Since then, better engagement with NGOs, such as Airwars, has allowed for greater levels of CIVCAS reports to be considered, with such organisations becoming the dominant source for claims. In 2018, Airwars referred just under 90% of all assessed civilian harm events. Overall, of the 64 incidents judged ‘credible,’ 23 were Airwars referrals.

However, issues continue to arise in regard to the treatment of reported CIVCAS incidents, particularly surrounding the isolation of an incident to a specific time and location. Several criticisms highlight the inability for some local populations to report an incident in a timely manner due to wartime contexts and cultural constraints. Some criteria on date, time and location is also said to overly filter CIVCAS allegations. Standards are also said to vary significantly across military operations.

Some NGOs report that that they often do not receive a response to their reports of CIVCAS incidents. Of further concern, the Coalition has consistently failed to provide reasoning behind the decisions for incidents declared non-credible. Without feedback, NGOs are unable to rectify reporting errors or provide information that the Coalition may have thought was lacking.

It is therefore considered highly likely that external allegations are ruled out, which may have caused civilian casualties.[7] Such external allegations are vital to better monitor civilian harm. A recent study on CIVCAS reporting during recent US military operations from the US’s own Pentagon, released in February 2019, reviewed 191 credible incidents from the CJTF-OIR CIVCAS database, which occurred between January 2015 and December 2017. They found that external allegations were the source for just 23 of these 191 incidents, yet accounted for 58% of the total number of civilian casualties. Overall, that report sought to identify if civilian casualties recorded by the Coalition were an accurate reflection of the harm caused. It was found that the current numbers are far lower than the true figure is likely to be.[8]

Role of CENTCOM and transparency of reporting

Beyond the criticisms mentioned earlier, in regard to the difficulties in distinguishing between Coalition members’ responsibility for strikes, there are further issues concerning transparency and accountability. A recent Pentagon study found significant flaws in recording and reporting civilian casualties, flaws that are likely to lead to lower numbers of civilian casualties recorded. Amongst some of the most notable concerns were:

  • Feedback to subordinate commands of the reasons for/or lessons learned from a civilian casualty incident is inconsistent;
  • There was decreasing transparency in US government reporting, as CENTCOM’s public release of assessment and investigation findings offers little detail as to why a CIVCAS allegation is considered ‘not credible.’ Since the autumn of 2017, strike data from Operation Freedom’s Sentinel (operating in Afghanistan) is no longer shared. The US-led Coalition also stopped reporting strike data for Iraq and Syria in December 2018;
  • US military standards for verifying third party allegations vary significantly and some may be construed as restrictive.

The report recommends greater efforts to monitor civilian casualties and greater transparency. Amongst other recommendations, it encourages the US to: clarify guidance and doctrine to address the increased risk of casualties when operating as part of a coalition; systematically seek additional sources of information on potential civilian casualties; include individuals in the CIVCAS cells tasked with reconciling external and US military reports on CIVCAS, as well as coordinating with relevant units to declassify or appropriately release relevant information; and institutionalize the CIVCAS investigation process, including sharing best practices and closer engagement with NGOs.

Beyond recording civilian casualties, the study did not recommend changes to tactics or RoE. However, this seems to have produced disagreement amongst the authors of the paper, and understandably so, given the levels of civilian harm. On the positive identification (PID) process, the report believes there is ‘sufficient guidance and structure and therefore does not increase the risk for civilian casualties’. Yet, the footnote outlines considerable disagreement on this from the authors, some of whom disputed this claim and instead characterise the PID process as a significant cause of civilian casualties and ‘a driver in the major civilian casualty incidents that led to significant DoD reviews’.

It is perhaps worth noting that in 2018 the Coalition began to share multiple Military Grid Reference System (MGRS) coordinates, for both ‘credible’ and ‘non credible’ assessments, with Airwars. This improvement was welcomed but, unfortunately, at the end of 2018 the Coalition stopped reporting the locations and dates for each air and artillery strike, drastically decreasing transparency. Between mid-December 2018 and the end of January 2019, the military reported more than 1,600 strikes in Iraq and Syria, accounting for 5% of US strikes since 2014. For all 1,600 airstrikes, the Pentagon omitted details on strike dates and locations.

US Rules of Engagement

The Law of Armed Conflict’s standard for civilian casualties provides the foundation of the United States’ Standing RoE. These govern the actions of all US forces, alongside mission-specific RoE, which generally establish what groups and individuals may be targeted – those that are hostile – and what groups may be defended, such as US forces, partner forces and designated non-military personnel and groups.

The United States Air Force identifies two general types of targeting selections: deliberate and dynamic. Deliberate targeting selection is a process in which strategic locations are targeted, typically 36-40 hours in advance, that serve the particular campaign’s strategy. Dynamic targeting is typically reflective of a more fluid operating environment. Dynamic targeting can include strikes that are unplanned, unanticipated or self-defensive (unit/collective). This type of targeting is often more time-sensitive and involves the opportunistic selection of enemy combatants. The latter is more likely to result in the targeting of individuals or groups moving within a populated area.

The DoD provides that a citizen or national of a state involved or engaged in a conflict may be subject to the hardships of war. However, if a civilian is not actively engaged in resistance, the civilian is not to be made the object of an attack. Furthermore, if a target is chosen that will likely lead to civilian casualties, the operational command is to carefully consider the cost-benefit ratio of the designated military objective of the strike and the likely level of civilian casualties.[9] The use of the term object appears to allow for incidental civilian casualties.

When determining if civilian casualties are likely, the non-combatant value (NCV) will also be considered – this is the highest number of civilians that can be put at risk by strike planners without seeking further approval. This is context specific and changes to this number can occur according to country, region, environment, target and own forces at risk. For example, during the surge in Iraq in 2007, the NCV for some targets could be as high as 26. In Iraq in 2003 the value had been 30. During the surge in Afghanistan in 2010, the NCV was nine. After anti-government protests and outrage in Afghanistan, the NCV decreased to zero. In urban areas the value is often higher, meaning a greater risk of civilian casualties is accepted. In Ramadi in December 2015, the value was 10. In Syria and Iraq in 2017, the value was reported to be 10.

However, this allows far more civilian deaths than are likely to actually be perpetrated by the Coalition. In 2016, if the NCV remained at 10 throughout the year, this would have allowed a maximum of 45,890 incidental civilian deaths. Whilst it is clear this number has not been killed, the higher the NCV, the more civilian casualties are likely to be caused.

When circumstances permit, sufficient warning is to be given to a civilian population that may be impacted by an airstrike. Circumstances that may circumvent the need to provide clear warning can include strategic military maneuvers necessitating the element of surprise.

Dynamic strikes are said to account for the vast majority of civilian casualties, with dynamic strikes under troops-in-contact conditions in an urban environment especially challenging. Such strikes are often executed in a reduced timeframe, which may account for the higher civilian casualty rates. Dynamic strikes account for nine of every ten US-led Coalition airstrikes in Iraq and Syria.

The Hague IV regulations prohibit the aerial bombardment of civilian populated areas, including cities, towns and other dwellings where there is not an active defense present. However, in conflicts where non-state actors, such as insurgent or terrorist groups, operate amongst civilian populations, such group practices have risen as a defense to the use of air power in these areas. The US also identifies religious, cultural and charitable buildings and monuments as ‘no-strike’ targets. Nevertheless, as these facilities are often used as a shield in urban environments by non-state militant actors, with relevant commander approval, a strike may be carried out in close proximity to these sites. Strikes on such buildings have been carried out numerous times by the Coalition in Iraq and Syria – some have resulted in high numbers of civilian casualties; for example, see the strike on a mosque in Aleppo in 2017 mentioned below.

The US’s general approach to RoE has been guided by a series of principles intended to minimise civilian casualties whilst still retaining the right to use airpower on aggressors and in self-defence. Recent alterations to RoE have led to a relaxation of measures that could prevent civilian casualties and a broadening of permissions for the use of firepower.

Changes to RoE

Changes to RoE have the potential to seriously change the levels of civilian casualties that result from airstrikes. When casualties are effectively monitored and examined, changes to RoE based on such analyses have the potential to further mitigate civilian harm. When operating in populated areas, such monitoring becomes even more essential, and measures to mitigate harm should be implemented quickly. Dr Larry Lewis, an expert on civilian casualties and an author on the recent Pentagon study examining casualty monitoring, highlights the virtue of ‘operational adjustments’ based on civilian casualty tracking:

‘After the completion of the Joint Civilian Casualty Study in August 2010, I continued to receive updated civilian casualty data from ISAF [International Security Assistance Force] and used that data to determine trends and potential areas of concern. In early 2011, several types of operations showed increased risk of civilian casualties. After this finding was forwarded to ISAF, international forces made operational adjustments in those areas of operations to address those concerns. As a result, the civilian casualty trends were quickly reversed.’

Dr Lewis also highlights how the failure to implement monitoring efforts and accompanying operational adjustments in Iraq and Syria saw civilian casualties continue to rise over time.

On the other hand, frustration at a lack of ‘success’ in conflict or losses from allied forces can see RoE loosened. Despite the rises in civilian harm from US airstrikes in Iraq and Syria, RoE have been relaxed at points through the conflict. Some of these instances are discussed below.

In December 2016, Barack Obama’s administration loosened RoE, allowing personnel further down the command chain to call in airstrikes, and it is likely that this exacerbated civilian harm. In the weeks following this decision, the numbers of civilian casualties significantly increased.

The US administration under Obama also increased the non-combatant value of some targets – allowing a greater number of civilians to be put at risk by strike-planners without further approval. This change occurred just months before Obama released the 2016 Executive Order on Pre- and Post-Strike Measures to Address Civilian Casualties, which sought to better mitigate civilian harm.

During the battle for East Mosul, Iraqi soldiers suffered many casualties on the ground fighting against ISIS, and commanders complained about the lack of and delays in receiving air assistance. The relaxation of the rules, allowing more junior officers to call instant air support, is thought to account for the far greater levels of civilian casualties seen in West Mosul and subsequently in Raqqa.

Under US President Donald Trump and former Secretary of Defence Jim Mattis, some RoE relaxed even further. For example, the need for ‘proximity’ was loosened. This was discussed in relation to the conflict in Afghanistan, where they also sought to extend authorisation to order airstrikes. Under Trump there appears to have been a cultural shift to RoE from the point of view of the President, who advocated ‘total authorisation’ when it came to warfare.

Previous civilian casualty-causing incidents in Afghanistan, notably the accidental bombing of an MSF hospital in Kunduz, had led in 2015 to an initial tightening of the US RoE. Such tightening is likely to have been negated under these new rules from Trump and Mattis. In fact, over 2017 and 2018, the number of airstrikes that resulted in civilian casualties rose from two incidents in 2016 – albeit these left at least 97 casualties, including those in the Kunduz hospital strike which prompted changes to RoE – to eight incidents in 2017, causing 132 civilian casualties, and nine in 2018, causing 108 civilian casualties. Considering changes to RoE brought in after the Kunduz bombing in 2016 sought to mitigate civilian harm, President Trump’s changes appear to have reversed any potential progress.

The loosening of RoE in regard to who can call in an airstrike, made under both Obama and Trump, are particularly troubling. This has the potential to see those in lower positions of command call in an airstrike. There are dangers that these individuals may have other priorities or bias that could put civilians at risk – a concern raised at a roundtable at the end of 2018 on the RAF and RoE organised by AOAV and Chatham House.

In the first year of the Trump presidency, AOAV and others monitored an increase in civilian casualties from US or US-led Coalition airstrikes across Syria, Iraq, Yemen and Afghanistan, though only slight in the case of latter two.

In response to the escalation in civilian deaths and the Pentagon study already mentioned, David Trachtenberg was selected for the new position of Pentagon coordinator to ensure greater transparency. Improvements are said to be underway, with Joint Staff beginning to update internal manuals and casualty investigators increasing their use of external information. Though, as already mentioned, civil society has noted that the administration’s transparency on strikes since then has decreased.

International Humanitarian Law (IHL) and civilian protection

Having signed and ratified the four Geneva conventions and its Additional Protocols, the US are bound to IHL’s central principles of distinction, precaution and proportionality. These principles holistically maximise the protection of civilians and property during war, limiting the use of means and methods of warfare, and minimise the harm to civilians during conflict.[10]

The US’s RoE are the additional directives within the boundaries of IHL that provide the circumstances and limitations for the military to initiate and/or continue engaging in combat.[11]  It is the principal mechanism that ensures that US military forces comply within obligations under both domestic and international law, as well as reinforce IHL. However, there is evidence of incongruences between IHL and the US’s RoE.

For example, the US airstrike on a mosque in Aleppo in April 2017 appears to violate IHL, yet was in line with US RoE. Intel from the Pentagon claimed the mosque was where an al-Qaeda gathering was taking place.  According to Army Brig. Gen. Paul Bontrager: ‘any structure at all, if it’s being used for a military purpose, can be struck, can be a legal target to strike.’ This argument is based on the principle of distinction. However, the airstrike killed at least 38 civilians attending evening prayers.

Although the US DoD’s internal investigation deemed the airstrike to be legal, they admitted to failing to understand that the targeted building was a mosque, that prayer was about to begin, and that a religious lecture was taking place at the time of the attack.  As such, the UN Commission of Inquiry found the strike to be unlawful and in violation of IHL, because the US failed to take all feasible precautions to minimise loss of civilian life.

Investigations by NGOs throughout the period of the US-led Coalition’s involvement in Syria and Iraq have persistently highlighted instances, and campaigns, in which breaches of IHL appear to have taken place.

A 2017 report by Human Rights Watch (HRW) investigated some of the airstrikes that caused the most civilian casualties, including airstrikes that hit a school housing the displaced in Mansourah on March 20th 2017, and a strike on a market and bakery in Tabqa on March 22nd 2017. They found that, despite the presence of ISIS fighters, the civilian death tolls suggested the US-led Coalition failed to take the necessary precautions to avoid and minimise civilian casualties, as required by IHL.

The CJTF claimed they had believed the strike in Mansourah was targeting an ISIS headquarters and weapons storage facility. Residents spoken to by HRW denied the locations were military bases and explained that, whilst ISIS members and their families were present, civilians made up the vast majority of those at each location at the time of the strikes. Additionally, whilst ISIS families were staying at the school housing the displaced, they, as well as ISIS members who carry out purely non-combatant functions, are considered civilians under IHL. In response to IHL questions over the strikes, the CJTF informed HRW that they had observed the school location and determined there was no civilian activity. However, civilians on the ground assessed that both sites had long been used by civilians, and that ‘any person with local knowledge would likely have been able to identify the substantial risk that the two sites contained significant numbers of civilians’.

Given these circumstances, it appears that the US-led Coalition failed to detect dozens, if not hundreds, of civilians at these sites, failing to take all precautions to minimise harm or detect the civilian presence, and yet attacked anyway. In information provided to HRW, there is no indication that the Coalition used any human intelligence to verify information on the Mansourah strike. Witnesses suggest that any aerial monitoring of either site would have revealed civilian activity, including children playing at the school. HRW identified further similar attacks in which civilians were killed in US-led Coalition airstrikes on locations such as mosques and residential buildings.

A 2018 study by Amnesty International examined the conduct of the US-led Coalition in Raqqa. Whilst they recognise that ISIS exacerbated the challenges to conducting urban combat, such tactics were previously known and the Coalition still failed to adequately account for civilians and take precautions to minimise harm to civilians and civilian objects. Amnesty International concluded that ‘the cases documented provide prima facie evidence that several Coalition attacks… violated international humanitarian law’.

In Raqqa, the US forces carried out 90% of the airstrikes (and 100% of the artillery) – the only other US-led Coalition members to conduct airstrikes were the UK and France. The strikes detailed in the report were described as ‘disproportionate or indiscriminate or both and as such unlawful and potential war crimes,’ and are emblematic of the wider patterns of harm seen throughout the Coalition’s operations in Syria and Iraq.

Despite such criticisms, the US-led Coalition restated in January 2019 that the Coalition and partner forces ‘continue to make the safety of civilians in the area a top priority’.


Whilst the US has admitted to civilian casualties during the conflict in Iraq and Syria and appears more transparent than their European counterparts, they lack the transparency and engagement in monitoring civilian casualties that they could have, particularly considering that they are also conducting the vast majority of the airstrikes in Syria and Iraq.

In recent years, as air campaigns mainly in Iraq, Syria and Afghanistan have intensified, the US’s RoE appear to have loosened. Some of these measures have been acknowledged, and as a result, the campaigns engaged in by the US appear to have seen rises in casualties, as documented earlier. In some recent and concerning instances, despite progress over the past few years, transparency has also worsened.

The lack of transparency is concerning given the potential breaches of IHL documented by NGOs in airstrikes that have caused thousands of civilian casualties across Syria and Iraq and left even more civilians impoverished and displaced. Such levels of civilian harm should be of considerable concern. Beyond just avoiding IHL violations, the US should seek to employ all measures to mitigate civilian harm, and it is clear that RoE could be strengthened to achieve this. In order to better prevent such harm, the US must commit to better monitoring and recording of civilian casualties, and from such records implement changes to RoE to ensure civilian harm is mitigated as far as possible.

As highlighted by Dr Larry Lewis, this has been done in the past in response to the rate of civilian casualties seen from US engagement in Afghanistan. The US needs to ensure that this is a continual process that allows RoE to respond to particular contexts and constantly seek to monitor and mitigate civilian harm. That this process has not been seen in Iraq and Syria is a considerable failure of the US and one that has cost civilians dearly.

[1] As of February 19th 2019.

[2] Chris Woods, Airwars, APPG on Drones meeting, ‘Airstrike and Civilian Casualties’, House of Commons, UK, 26th February 2019.

[3] As of February 19th 2019.

[4] Chris Woods, Airwars, APPG on Drones meeting, ‘Airstrike and Civilian Casualties’, House of Commons, UK, 26th February 2019.

[5] Lafta R, Al-Nuaimi MA, Burnham G (2018) Injury and death during the ISIS occupation of Mosul and its liberation: Results from a 40-cluster household survey. PLoS Med 15(5): e1002567.

[6] Figures are available from AOAV upon request

[7] Dr Lewis, at the APPG on Drones meeting, ‘Airstrike and Civilian Casualties’, House of Commons, UK, 26th February 2019.

[8] Dr Lewis, at the APPG on Drones meeting, ‘Airstrike and Civilian Casualties’, House of Commons, UK, 26th February 2019.

[9] Blank, L. R. (2018). Muddying the Waters: The need for precision-guided terminology in the DOD Law of War Manual. In the United States Department of Defense Law of War Manual Commentary and Critique, forthcoming 2018). Retrieved from

[10] Blank, Laurie. (2018) Muddying the Waters: The Need for Precision-Guided Terminology in the DOD Law of War Manual. The United States Department of Defense Law of War Manual Commentary and Critique (Michael Newton, ed. 2018, Forthcoming); Emory Legal Studies Research Paper.

[11] Joint Chiefs of Staff. 2007. Legal Support to Military Operations, no. 1-04.