On May 6 2025, a ceasefire between Yemen’s Houthis and the United States ended the most extensive military campaign of President Trump’s second term to date. But what was the full human cost of Trump’s eight-week bombing campaign, and how does it compare to the history of U.S. military action in Yemen?

Airwars analysed every public allegation of civilian harm during the Trump campaign against the Iran-allied Houthis - dubbed Operation Rough Rider - and compared it to previous harm allegations from U.S. campaigns in Yemen, both targeting the Houthis under President Joe Biden and against Al-Qaeda in the decades before.

Key findings reveal:

• In the period between the first recorded U.S. strike in Yemen to the beginning of Trump’s campaign in March, Airwars tracked at least 258 civilians allegedly killed by U.S. actions. In less than two months of Operation Rough Rider, Airwars documented at least 224 civilians in Yemen killed by U.S. airstrikes - nearly doubling the civilian casualty toll in Yemen by U.S. actions since 2002.
• The two deadliest civilian harm incidents publicly recorded in the history of U.S. military operations in Yemen occurred during Trump's campaign. Strikes on Ras Isa Port and Saada’s Remand Detention Prison allegedly killed at least 152 civilians and injured almost 200 others. Human Rights Watch and Amnesty International have now questioned the legality of both strikes.
• The scale of the campaign resulted in an unprecedented level of civilian casualties per incident. During Operation Rough Rider, Airwars documented more incidents with higher numbers of casualties per strike than in any other U.S. campaign.
• Some of the most advanced munitions in the U.S. military arsenal were deployed, including the first documented use of the StormBreaker in combat - a new precision-guided U.S bomb.
• Civilian harm incidents were concentrated in the heavily populated cities of Sana’a and Saada. This differed from President Biden’s campaign against the Houthis, where civilian harm was typically dispersed across less heavily populated areas in western Yemen.

23 years of strikes

The United States has been bombing Yemen for more than 20 years. The first known drone strike was conducted by the U.S. in 2002, targeting Al-Qaeda in Yemen. Full scale counterterrorism strikes began in 2009, when the Obama administration led a targeted drone campaign against Al-Qaeda in the Arabian Peninsula (AQAP). In December 2009, the first instance of civilian harm was recorded, with at least 47 civilians reportedly killed by a U.S. strike. This incident remained the deadliest single U.S. action in Yemen until earlier this year. The sporadic campaign against AQAP and other Islamist groups has continued from 2009 until today.

In October 2023, the pro-Iranian Houthi political and militant group - which controls the Yemeni capital Sana'a - began attacking vessels in the Red Sea, demanding an end to Israeli attacks in Gaza. The attacks disrupted international trade, prompting a campaign led by the U.S., in partnership with the United Kingdom, to attack Houthi military capabilities in January 2024. Under President Biden hundreds of strikes were conducted, though Houthi attacks on both shipping and Israel largely continued.

In March 2025, Trump launched another campaign to restore freedom of navigation and deter Houthi attacks, calling Biden’s campaign “pathetically weak”. The debate on when to launch the first wave of strikes was revealed in a leaked Signal group chat between top U.S. national security leaders. The strikes authorised in the group chat marked the beginning of Operation Rough Rider.

The campaign reportedly cost more than $1 billion in just its first month. In May 2025 a ceasefire agreement saw both sides claim victory, with the Houthis committing to no longer attacking U.S. ships.

Strikes more deadly for civilians than any other U.S. campaign

Gaining a full understanding of civilian harm in Yemen is challenging, particularly through open sources. The Houthis have sought to control the information environment, and often impose strict measures to prevent access to strike sites. With photography prohibited and restrictions placed on health centres, hospitals and medical teams in tracking civilian harm, casualty recording groups have little documentation to rely on. During Operation Rough Rider the Houthis also launched a campaign online urging Yemenis not to post details or images from strikes, or the names of those killed.

Despite this, Airwars’ documentation teams were able to identify 33 cases of civilian harm under the second Trump administration, and have drawn upon the work of other organisations such as the Yemen Data Project and Mwatana for Human Rights in piecing together accounts of the violence.

By examining publicly available incidents in the Airwars archive and comparing the Trump campaign to previous U.S. military operations in Yemen, this report identifies patterns of harm between incidents recorded under the same or similar restrictions.

The rate of civilian harm over eight weeks of Operation Rough Rider was unparalleled in the 23 years of U.S. action in Yemen.

Over a 23-year period, from November 2002 to January 2025, U.S. military actions allegedly killed at least 258 civilians, according to Airwars monitoring. In eight weeks, at least 224 civilians were allegedly killed.

Strikes in which civilians were killed were also more deadly than previous U.S. campaigns. By the time the operation came to an end in May 2025, Airwars had documented at least five incidents in which more than 10 civilians were reported killed.

Incidents with more than 10 fatalities were comparatively rare in previous campaigns. Airwars had documented just six such cases in the previous 23 years.

Included in the Trump campaign were the two deadliest incidents ever recorded in the history of U.S. action in the country. In a strike now labelled an apparent war crime by Human Rights Watch, Airwars documented 84 civilians killed at Ras Isa port in April. Less than two weeks later, Airwars recorded a further 68 civilians killed in a strike on a detention prison. Amnesty International has said the strike may have also violated international humanitarian law.

The pattern of strikes resulting in civilian harm reflected the more residential nature of the targets. Of the 33 incidents of civilian harm allegations under the Trump administration, 16 involved strikes on residential buildings. 14 of these occurred in the cities of Sana’a or Saada. In these incidents alone, a minimum of 39 civilians were allegedly killed - exceeding the number of civilians allegedly killed in the entirety of Biden’s campaign.

Under Biden, Airwars documented at least 30 civilians allegedly killed across 15 incidents in the campaign against the Houthis. These strikes were largely dispersed across western Yemen in less densely populated areas, with five incidents of civilian harm in Sana’a, two of which were fatal, reportedly killing four civilians.

This eight-week period also saw a larger range of targets struck. Incidents of civilian harm involving strikes by the U.S. military on communications, legal and water infrastructure were documented for the first time.

These higher levels of civilian harm raise questions about the administration’s commitment to existing civilian harm mitigation mechanisms enshrined within U.S. policy and law.

Wes Bryant, a U.S. Special Operations veteran, told Airwars it was unlikely the administration had undertaken sufficient civilian harm mitigation measures during this campaign.

“[For] each and every strike, there should be an equally weighted effort in civilian harm mitigation fitting with all of the doctrine and standards that we already have in place in the targeting process - from planning characterisation of the civilian environment before execution, to execution of the actual strike while in process.” Bryant, who is also the former branch chief of civilian harm assessments at the Pentagon, told Airwars.

“I’d say [it is] a near impossibility to do the level of civilian harm mitigation we’ve seen in other campaigns with this high intensity of a strike campaign in such densely populated urban areas with, again, no actual partner forces on the ground and very limited source intelligence on the ground.”

Yemen Data Project found similar trends to Airwars. “Overall, there were clear indications, both in the data and from statements made by President Trump and his administration, that the threshold for civilian harm had been reduced and the tolerance for civilian casualties raised,” Iona Craig, the organisation’s representative, told Airwars.

As with all casualty recording efforts relating to alleged U.S. action in the Middle East, Airwars submitted each of these cases to United States Central Command (CENTCOM) via their Civilian Harm reporting portal. Though reporting on the outcome of civilian harm assessments is required under U.S. policy, no public determinations have been released to date.

Testifying before a House Armed Services Committee hearing on June 10, CENTCOM Commander General Michael Kurilla said CENTCOM would "absolutely" publish their findings on civilian casualties in Yemen.

Airwars reached out to CENTCOM for comment on the 33 strikes documented to have caused civilian harm under Trump. In response, CENTCOM stated, “[We are] aware of the claims of civilian casualties related to the U.S. strikes in Yemen, and we take those claims very seriously. We are currently conducting our battle-damage assessment and inquiry into those claims.”

‘This strike was not intended to harm the people of Yemen’

On April 17 the U.S. struck Ras Isa Port, allegedly killing at least 84 civilians and injuring 150 civilians, including first responders from the Yemen Red Crescent Society and Civil Defense of Yemen. At least three children were amongst those killed.

The port, a major oil terminal which also plays a key role in supplying humanitarian aid, was targeted throughout the night. In a statement released the same day, CENTCOM said the aim was to “eliminate this source of fuel for the Iran-backed Houthi terrorists” and that the strike “was not intended to harm the people of Yemen.” So why was this strike so deadly?

The strike on Ras Isa was highlighted during a hearing of the House Armed Services Committee on June 10, during which CENTCOM Commander General Kurilla was asked about the steps taken to mitigate civilian harm during Operation Rough Rider.

“We did messaging to the individuals there,” Kurilla said, when commenting on the Ras Isa Port strike. “We did messaging to the ships.”

“If the purpose of the attack was to destroy a fixed target, then it was feasible for the U.S. military to provide warning to civilians before the attack,” Dearbhla Minogue, senior lawyer at Global Legal Action Network (GLAN), told Airwars. “If they did not, and they could have, this breaches the fundamental IHL obligation to take all feasible precautions to avoid and minimise civilian harm in attacks.”

Multiple sources, including a Civil Defence official who was stationed at the port, told Airwars that there was no effort to warn civilians ahead of time or to facilitate an evacuation before the port was targeted. In the absence of any evacuation warnings large numbers of civilians were killed in the strike, many whilst working at the port.

Airwars asked CENTCOM to clarify whether Kurilla’s statement before the committee indicated that the U.S. issued warnings prior to the strike, but CENTCOM stated that it had “nothing additional to provide outside of Gen. Kurilla’s testimony.”

On June 4 2025, Human Rights Watch declared that the strike should be "investigated as a war crime."

“I don’t believe there is any way this attack could have been lawful,” Minogue said.

At the hearing, Kurilla said CENTCOM was investigating the strike to determine the “actual” number of civilians killed, claiming that the widely circulated death toll was referring to “Houthi casualty numbers”.

However, through open source research, Airwars has identified the 84 people allegedly killed in the strike, including four families whose members were killed alongside each other. These names were identified based on testimonies and tributes posted online by loved ones and colleagues of those killed.

22-year-old Salim lost five members of his family in the strike on Ras Isa - all of whom were working at the port when it was attacked.

“All of them were young and all of them were married and had children. They were all drivers - just truck drivers. No involvement in any party. Just civilians doing their jobs,” he told Airwars.

“We haven’t found anything left of them.” Salim said.

The U.S. used some of its most advanced munitions to strike Yemen

Open source evidence suggests that in addition to standard air-delivered bombs, the U.S. employed some of the most advanced munitions in its arsenal during Operation Rough Rider.

This campaign also marked the first time that GBU-53 StormBreaker glide bombs were used in a conflict zone. The StormBreaker, one of the U.S.’ newest air-delivered munitions, has a range of up to 110 kilometres and the ability to be redirected to other targets after launch. Airwars documented one incident of civilian harm involving a StormBreaker - a civilian was allegedly injured in a strike on a residential building in Sana’a on March 28.

These munitions ranged from Tomahawk land-attack cruise missiles launched by warships in the Red Sea, to air-launched cruise missiles fired from fighter aircraft. The use of highly sought-after munitions suggests that U.S. forces were concerned by the strength of Houthi air defences.

Remnants of less advanced munition types than the StormBreaker were also identified in civilian harm cases in Yemen, including the apparent use of the GBU-39 in the April 28 strike on the migrant detention prison in Saada - one of the deadliest incidents recorded during this campaign.

Tallying the human cost of Trump’s campaign continues despite ceasefire

The ceasefire provides opportunities for civil society organisations to further investigate casualties that occurred under Operation Rough Rider. But, despite cases being submitted to CENTCOM through its own civilian harm assessment system, the military justification for these strikes, and whether the U.S. will accept responsibility for the strikes that have harmed civilians, remains in question.

Airwars reached out to CENTCOM for comment on the tools it is using to assess civilian harm allegations. CENTCOM stated that battle-damage assessments (BDA) are one tool, but for “operational security purposes” it could not disclose more information on the specifics of the investigation process.

Battle-damage assessments have historically led to unreliable numbers of civilian harm. In 2010 during U.S. military operations in Afghanistan, one study identified that BDAs missed civilian casualties later identified during ground-led investigations in 19 of 21 cases - more than 90 percent of the time.

In the absence of military admissions, casualty recording groups are continuing to work to illustrate a fuller picture of civilian harm.

Mwatana for Human Rights, told Airwars: