SAN10006 - Jaref Spa Bombing

May 8, 2019

Incident ID: SAN10006
Location: Near Jaref Spa, Bilad ar-Rus, Sana’a
Coordinates: 14.998248, 44.155014
Incident Grading: Confirmed
Date: 2016/01/12
Time (AST): Between 0800-1100 am AST


- Open source information, including news reports and images taken at the location confirm that a building approximately 1 km from the Jaref Spa was hit by an airstrike on 2016/01/12. Satellite imagery from the site taken the following month supports this conclusion.

- Contextual evidence and witness accounts of planes flying overhead indicate that this was an airstrike. Reports and open source imagery indicate that many of the victims of the attack were women and children.

- Open source information suggest that this area is a tourist destination, and that it is visited for natural health purposes. Other information indicate that the actual building hit may have been a mosque.


Searches were carried out using English and Arabic variations of the location names “Sana’a”, “Jarif”, “ar Rus” (الروس) and the words “resort”, “spa”, “bath” and “mosque”. We also searched across these terms using the keyword “airstrike”. These search terms were run through Google with site searches for specific sites of relevance, such as YouTube and the Saudi Press Agency. The search terms were also used in Tweetdeck and for Facebook. The results included searches timebound to focus on the days immediately following the alleged date of the strike: 2016/01/12. A full list of results from these searches can be found below this report.

Example search string: site: حمام جارف


Where did this incident take place?

Some of the reports indicated the incident took place at the Jaref Spa Resort along Wadi Jaref Road, located at coordinates 15.001994, 44.164332. The resort is located in the district of Bilad Ar Rus south of Sana’a, which was identified using this map from ReliefWeb. Having narrowed down the search area, the resort’s pool was identified in satellite imagery captured prior to 2015/05/12. After finding the location on satellite imagery, it was found that the suspected site was indexed by Google as “Jaref Spa,” which led to this Google Maps webpage.

Left: reference imagery from The Intercept depicting Jaref Spa. Right: Google Maps satellite image (courtesy of Google/Airbus)

However, although this location matched the site of previous raids, it did not match images and footage reportedly showing the aftermath of the 2016/01/12 raid which had been posted on social media. We therefore widened the search area to buildings outside the immediate area of the spa, and discovered a location at 14.99825, 44.15502 where a large building had been destroyed during the correct period of time.

Image of suspect location. Left: 2016/01/04, Right: 2016/02/27 (courtesy of Google/Airbus)

The topography facing southeast from this location matches that seen in this video posted to Twitter on 2016/01/12.

Top: skyline from video, Bottom: Skyline from Google Earth Pro landscape function (courtesy of Google/Airbus/Landsat/Copernicus)

Other images and videos purporting to depict the aftermath of the strike on 2016/01/12 can be confidently geolocated to this location.

Left: image of strike location, note the bend in the dry riverbed to the left. Right: satellite imagery of the destroyed building from 2016/01/03 (courtesy of Google/Airbus)
Top: still of strike location from the News Yemen video, Bottom: satellite imagery of the destroyed building from 2016/01/04 (courtesy of Google/Airbus)
Top: image of strike location posted to Facebook. Note the construction of the hut outlined in green. Bottom: satellite imagery of the destroyed building from 2016/01/03 (courtesy of Google/Airbus)

Some of the images and videos posted to social media depicting casualties have details which match this location. In the image below the bodies are laid out on a dried-up river bed. The building that was bombed was immediately next to a dried-up riverbed.

Image showing bodies on riverbed

The same set of bodies can be seen in the image below. We can clearly identify the riverbed, the blue tarpaulin, and a track running into the riverbed. Note that the same orange and green blanket, motorcycle and blue tarpaulin can be seen in both the image above and below.

Top: Image depicting casualties, Bottom: Satellite image from this location taken on 2016/01/03 (courtesy of Google/DigitalGlobe)

In the image below a man can be seen holding a severely injured or dead child in front of a small stone hut, which matches the huts that are located around the bombed building and that can be seen at 01:47 in the News Yemen video.

Image of man in front of hut. We have obscured a graphic injury.

We can therefore say with a high level of confidence that most of the news reports and user generated content (UGC) claiming to show the scene of the 2016/01/13 incident were taken in the vicinity of a large building at 14.99824, 44.15501.

Although some reports appear to claim that the Jaref Spa itself was also hit, there does not appear to be any visible damage caused to the spa building on satellite imagery between the period 2016/01/04 - 2016/02/27. Some images posted in connection with the strike on 2016/01/13 that depict the spa location appear to have been recycled from the strikes that took place in September 2015 (1, 2).

Locations marked on the crowd-sourced mapping application Wikimapia indicate that the sulphuric baths extend further down the wadi, beyond the purpose-built spa and past the location of the strike. It seems likely that many users were not aware that the natural hot springs cover a large area and so confused the actual strike location with Jaref Spa.

What was the location being used for?

Much of the user generated content and news reports we found (see Discovery) claims that a mosque was hit in this strike. The building we identified is certainly large enough for this purpose, and does externally resemble some kind of hall. The large pillars seen in many of the images and videos of this location may support this.

Image posted to Facebook of this location. Note the large pillars.

It also appears that some kind of red cloth, possibly a carpet, can be seen strewn about this site.

Red cloth seen in many images and videos of this location. (Sources, clockwise from top left: 1, 2, 3, 4)

In several of the images and videos of this location a pile of books can be seen at this site, as well as what appear to be parts of the red carpet.

Image posted to Facebook showing pile of books. This same pile can be seen from different angles in other images and videos.

A close-up shot of this pile shows at least one of these books to be the Quran. It should be noted that the Quran would be a common book to find in many buildings.

Still from this Yemen News video at 02:08 showing a passage from the Quran

Taking into consideration the size, construction and some of the contents of the building, as well as multiple reports that a mosque was hit, it appears probable that this building was a mosque, although further investigation is required.

There also appears to be a number of smaller, stone buildings in the immediate vicinity of the building that was destroyed. There do not appear to be any clear images inside these buildings, however, a bicycle can be seen inside one, and men can be seen tugging at a brightly coloured blanket in another. This indicates these buildings were in use, although their exact purpose is not clear.

Still from this Yemen Today video showing men pulling a blanket from the rubble of what appears to be one of the smaller buildings

Although there are many claims that a mosque was hit, the number of claims that “baths” were hit are interesting. Although clearly some reports have confused the Jaref Spa with this location, other comments appear to indicate that the mosque and baths were connected. On Wikimapia the site of the attack is within an area marked as ​​”a large sulfuric water bath”.

Note that the building sits within an area marked as a “sulfuric water bath” on Wikimapia.

Historic imagery from Google Earth, taken in 2011, also indicates that this location was the source of a flow of water.

Historic satellite imagery taken on 20110412 which appears to show a flow of water originating from this location (courtesy of Google/DigitalGlobe)

Military activity?

There is no obvious military activity in any of the images or videos of this location that could be found. However, 400 meters to the south of the strike location there appear to earth-works along the top of a ridge approximately 1500 meters long. These earthworks had a second layer added between 2013/09/24 and 2015/12/30 and resemble an earth bank. It is possible that these are defensive earthworks of some kind, however without further investigation it is not possible to reach a firm conclusion.

Satellite image taken on 2015/12/30 showing earthworks in lower half of image. The strike location is marked in red in the top right (courtesy of Google/Airbus)

One witness interviewed after the incident claimed that the site as “a tourist place, where there are no camps or gatherings for the militants.”



All sources, including news and social media posts, agree this attack took place on 2016/01/12.


The earliest tweet we identified about an attack in this region was made at 0844 AST on 2016/01/12 by an account called @SahafahyemenNet.

Earliest Tweet referencing an airstrike in the area. Note that Twitter displays the time of the viewer’s Twitter settings - this Tweet was made at 0644 BST, so 0844 AST.

The first mention we could find of an attack on the baths themselves was a Tweet made at 0909 AST. No mention was made of a mosque.

Earliest Tweet referencing this airstrike. Note that Twitter displays the time of the viewer’s Twitter settings - this Tweet was made at 0709 BST, so 0909 AST.

These Tweets are roughly consistent with reports that this strike took place at around 0800 AST.

There appear to be two images that depict the strike as it happened, both of which were posted in this Tweet. These images can be geolocated to a place approximately 3.7 km to the south of the strike location. Due to the camera being higher than the highlighted buildings the cameraperson was likely standing slightly to the south on the ridge-line at 14.96441, 44.15459.

Left: image showing moment of strike, Right: location depicted in the foreground of the image in the vicinity of 14.96695, 44.15455

The location of the impact cloud lines up perfectly with the location of the strike on 12/01/2016, as portrayed by the red line in the image below. This image cannot depict the September strikes on the Jaref Sap resort, as that location is too far to the east, as shown by the blue line in the image below. Note the massif immediately behind the strike location in both the image of the strike and Google Earth Landscape render. This image cannot portray any strikes after 2016/01/12 as that was the date it was posted on Twitter. Therefore these images almost certainly depict the strike on 12/01/2016.

Google Earth Landscape view from approximate camera location. The red line points to the location of the 2016/01/12 strike, the blue line points to the location of the Jaref Spa resort, and the yellow box highlights the buildings seen in the geolocation of the image (courtesy of Google Earth/DigitalGlobe/Airbus/ Copernicus)
Comparison of image of strike to Google Earth Landscape view. The light area on the right image indicates what is shown in the image of the strike. Note the orientation of the massif in the background. (courtesy of Google Earth/DigitalGlobe/Airbus/ Copernicus)

Due to a lack of reference points in the image, we cannot use to identify exactly what time these images were taken due to the inability to accurately align the shadows. However, we can say that it took place some time during the morning.

In the below image we have placed 300 m lines to represent shadows running from the strike location along the axis that a shadow would lie upon during various times. We established the correct bearing using the Sun’s azimuth as given by and then drew these lines using Google Earth and a compass tool. Although the length of these lines is a very rough approximation, they do allow us to examine the angle at which the shadows lay, which vary based solely on time.  

For example, at 0800 AST on 2016/01/12 the azimuth at this location was 120°. Therefore the angle of the shadow would be exactly the opposite direction. 120° + 180° = 300°.

- 0700 - turquoise: 294°
- 0800 - orange: 300°
- 0900 - black: 307°
- 1000 - purple: 318°
- 1100 - yellow: 334°

Image of simulated “shadows” and compass tool used to measure them (imagery courtesy of Google/Airbus)

When we compare the image of the strike to the simulated “shadows” we can see that the shadows appears to best match around 0700-0900 (turquoise to black). However, this technique is very far from precise due to a lack of reference points, and can only accurately establish that the strike took place sometime in the morning.

Left: simulated “shadows” (courtesy of Google/Airbus), Right: image of strike


To Whom?

Media reports and open source imagery indicate that many of the victims of the attack were women and children. This video, posted by the Tribune of Yemen, shows several casualties that are not seen in other media, including two children with severe head-wounds, one of whom appears in other images of this attack.

The same casualties are consistently seen in images and videos of this attack, including the same wounds, the same clothing and the same improvised shrouds.

Note the colours of the improvised shrouds. Left: Image of bodies in shrouds lined up immediately to the south of the site of the attack. Right: bodies being placed into matching shrouds from the Yemen Today video.
A blanket with the same pattern can be seen in both this image (left) the video from Yemen Now (top right) and the Tribune of Yemen video (bottom right)

Within these sources, which all seem to show the same group of fatalities, there are several men, at least one woman, and at least two children. The videos or images appear to focus primarily on fatalities and do not seem to depict many seriously wounded people. None of the videos or images show fatalities who were obviously combatants, although it should be noted that it is unlikely that every fatality is portrayed in open sources.

Is there evidence of military activity?

While we could not find evidence of military activity at this exact location prior to the strike, there are armed individuals pictured in imagery of the aftermath, most of whom are armed with rifles. It should be noted that in Yemen being armed can be regarded as a mark of status and does not necessarily indicate the person has a military role.

Still from Yemen Today video showing armed man behind the interviewee

However, there was at least one person in the Yemen Today video who seems to have been armed with a light machine gun, possibly an RPD or RPK.

Image from Yemen Today video at 00:24. Note the man in front of the pillar, who appears to be carrying some kind of light machine gun, possibly an RPD or RPK.

There was also at least one claim on Facebook that Houthi’s were at this location and were the reason for the bombing, made by a user named Ahmed Rashid claiming to be a lawyer living in Sana’a.

What kind of vehicles are present in the image and videos?

At least three vehicles could be identified at the site of this bombing. From the sources available, none of the vehicles were obviously military or indeed ambulances.

Composite panorama of the scene in this Yemen Today video looking south. Two vehicles can be seen highlighted in red on the west side of the building, and one vehicle can be seen highlighted in yellow on the east side.
The two vehicles on the west side of the building (source)
The vehicle on the east side of the building (source)


The casualty count for this strike varies between two and 20 fatalities:

Yemen Press - Two fatalities

Dhamar News -  “Dozens of martyrs” and wounded - 20 fatalities, “dozens wounded”

Tweet stating that seven women were among the fatalities

Tweet claiming 20 fatalities, 21 wounded

Tweet claiming 14 fatalities: four children, five women, two elderly people, and three men. 14 wounded


Neither the Joint Incidents Assessment Team (JIAT) nor representatives of the Saudi-led Coalition (SLC) appear to have made statements on this strike. Some reports point to information from “Houthi media”, but no official statement from the Houthis could be found.



There is no obvious impact crater visible in UGC or satellite imagery.


In this video, a man being interviewed is holding what may be remnants of a munition, however the video is too low quality to assess exactly what this object might be.



There are some claims on social media that the area was targeted by multiple strikes, and this report claims that paramedics tending to the victims were targeted by further strikes. However, there is no open source imagery that directly supports these claims. None of the vehicles we identified as having been destroyed in this strike appear to be ambulances.

The images which appear to show the strike itself in this Tweet may indicate that two strikes took place. In one of the two images there is what appears to be a dissipated cloud of dust to the left of the obvious impact cloud. There is the possibility that this is the remnants of the cloud caused by a previous strike that happened several minutes before. Without having a video or clear series of images showing the strike itself it is not possible to be sure of this point.

Image showing strike on building. Note what may possibly be remnants of a dust cloud in the orange box.


The level of damage caused to the building, contextual evidence, reports of planes flying overhead and the images of at least one explosion indicate that there was an airstrike at the site on 2016/01/12. This, combined with reports from The Intercept about an airstrike at the site “several days” before 2016/01/17 and open source information shared from the scene on 2016/01/12, confirm that an airstrike did take place at this site on that day.


Previous strikes

The nearby Jaref Spa resort was allegedly targeted by an airstrike on 2015/09/20, and by another on 2017/09/27. Some of the user generated content posted on 2016/01/12 was recycled from those incidents.

Later strikes

A strike on 2016/01/17 allegedly hit the same location as the 12th January attack. This strike is reported to have killed Almigdad Mojalli, a reporter who had gone to that location to cover the strike a few days before.

Multiple social media users reported that there was a strike on this area on 2016/01/17, such as this Tweet warning that “they are bombing the bath again” and this Facebook post stating “Violent raids of alliance aviation on the bath of ar-Rus”.

No images or videos could be identified from the attack on 2016/01/17


Open source information indicates that on 2016/01/12 a large building at 14.998248, 44.155014 was hit by at least one, possibly two, munitions dropped from a plane. Many reports claimed that this building was a mosque, which was supported by other details, however this could not be proved beyond doubt. Several of the casualties appear to have been women and children and no evidence could be found in open sources that any of the casualties were military personnel, although at least one Facebook post claimed that Houthi forces were in that location.


IBB Press report‍
Yemen 24 report
Yemen Press report
Dhamar News report
Al Shahed Press report
Civil Conglomerate for Development and Freedom Image gallery
Tihamah News: Report from earlier airstrike on the facility on 20 September 2015

Social Media posts and images:
Facebook post with statement
Facebook post with images of casualties
Facebook post with images of casualties and location
Facebook post with images of casualties
Facebook post with landscape image
Facebook post with image of graves
Facebook post with image of books
Facebook post claim that first responders were targeted
Facebook post with high quality images of location
Facebook post with high quality images
Facebook post claiming Houthi’s were present
Earliest Tweet found
Tweet with images of strike
Tweet with images showing casualties & area from the south

Video report from 27 September 2015 airstrike
Al Masirah, a pro-Houthi channel, footage of people at the scene of the airstrike
Further Al Masirah report of people at the scene of the airstrike
News Yemen report on airstrike
Yemen Today report on airstrike
from user Naji Almaznaei
from TribuneofYemen
from user Yemen Now

Statement by the Yemeni Ministry of Tourism from 2015/09/20 airstrike
Livejournal post depicting this location in 2012

Sources for attack on 2016/01/17:
The Intercept
: Report from a later airstrike on the facility on 2016/01/17, citing an airstrike at the same location “several days” earlier.
Facebook post statement on 2016/01/17 attack (account appears to be a Houthi soldier)
Facebook post statement on 2016/01/17 attack (account is a verified journalist and activist)
Facebook post statement on 2016/01/17 attack
Facebook post reporting raids on this location on 2016/01/17
stating “they are bombing the bath again”