The Art of Facts

A LEGAL BLOG about Fact-finding and Armed ConflicTS

The Use of Architectural Techniques in Fact-Finding during Armed Conflict: A Time Machine?

It seems these days one can find fact-finding references pretty much everywhere in the news. I admit this statement may just be due to me suffering from job conditioning. But take for example the report published this week by Airwars, a project led by independent journalists “tracking and archiving the international air war against Islamic State, in both Iraq and Syria”, including data for the numbers of strikes and “non-combatant” deaths. While they acknowledged the methodological limits and challenges of this initiative, their report points to a significant disparity between the figures gathered by Airwars and the number of publicly available official investigations into civilian casualties by the US Central Command (Centcom, that leads the coalition’s airstrikes): hundreds of deaths were reported while apparently only four investigations were conducted by Centcom.

A core issue here may be the difference in the standard of proof used to decide on publishing investigation findings and that to report on allegations and incidents of civilian casualties. Airwars’ journalists insist on the wide range of sources they use for data gathering purposes and on the classification used, based on “a grading system to reflect the current quality of reporting for each alleged incident” between “Confirmed”, “Fair”, “Poor”, “Contested Events”, and “Disproven” (Report, p. 4). Centcom on the other hand is said to only publish investigations with a “preponderance of evidence” of civilian deaths. However, these evidentiary thresholds refer to two radically different stage in the process: the first one refers to assessing whether an allegation of air strikes causing civilian casualties is credible, the other relates to whether or not the findings of an investigation would be made public. The real contention would then be more about the type of standard Centcom relies on to initiate its investigations, considering that any credible allegation should require a proper investigation and warrant the publication of findings.

But today I want to address a slightly different, but related topic. I may have used this introduction to catch your attention because some may find the following lines really repetitive… Yes I will be talking about challenges in fact-finding on violations of the law on conduct of hostilities again! Well, let’s face it, this is a key issue and there is no way it’s going to be solved any time soon. But as fact-finding methodology evolves and gets more professionalized, fascinating tools emerge that may help addressing some of those challenges.

As I discussed in a previous post (here) fact-finding about IHL norms on the conduct of hostilities requires considering a fundamental time paradox: while the legality of an attack depends in part on an ex-ante evaluation by the attacker, the facts are established after the attack (ex-post). The use of remote sensing technologies, such as satellite imagery may help address this challenge.

Now architectural techniques seem to offer even further potential. Amnesty International and Forensic Architecture, a research team based at Goldsmiths, University of London, published a report last week (‘Black Friday’: Carnage in Rafah during 21014 Israel/Gaza Conflict), that provides a detailed account of one of the deadliest and most destructive episodes of the IDF Operation “Protective Edge” in the Gaza Strip. Interestingly, unlike other reports on IHL and human rights violations that mostly get picked up for their findings, this study also got a lot of news coverage for the methodology used (see here for example). This focus is even more relevant as Amnesty International was not granted access to the Gaza Strip and had to conduct research remotely. This report obviously relies on many sources of information, from traditional ones such as eyewitness testimonies and satellite imagery, to rather new techniques such as the analysis of photographic and audiovisual material found on social media, reinforcing the citizens’ role in fact-finding. But the report in addition to the growing use of metadata analysis also combines those sources with architectural techniques, such as 3D modeling and two-point perspective, a methodology that brings the researchers closer to using a time machine to travel back in time to get a clearer picture of the actual attacks, as well as their features and impact.

Those techniques as explained and used in the report for various materials deserve the following extensive quote with excerpts from the report’s methodology section:

Video to Space

… Photographs and videos taken by both journalists and citizens captured large amounts of spatial information about environments in which events were unfolding, including the architectural layout of sites, the location and the time of day. Video stills were collaged together to create a panoramic view of space. Working back from distinctive architectural features – for example, a water tower, a high building, a crossroads or a football pitch – Forensic Architecture located still images within the satellite images and three-dimensional models.


Geosyncing refers to the establishment of space-time coordinates of an event. To reconstruct the events of 1-4 August, Forensic Architecture employed digital maps and models to locate evidence such as oral description, photography, video and satellite imagery in space and time. As such the media were used to reconstruct events, and to verify findings by cross-referencing various sources. When the metadata in an image or a clip file was intact, Forensic Architecture identified it by using Adobe Lightroom and Adobe Bridge. The geosyncing is in this case a straightforward process, undertaken on software platforms such as QGIS. However, material from social media often came without metadata, or with the metadata damaged or inaccurate. In the absence of digital time markers Forensic Architecture used analogue or “physical clocks” – time indicators in the image – such as shadow and smoke plumes analysis to locate sources in space and time.

Shadow Analysis

Forensic Architecture built a digital model of Rafah to locate witnesses and photographs in space, as well as to use the model as a digital sundial; all standard architectural modelling software currently come with shadow simulators. To establish the time of a photograph Forensic Architecture matched these digital shadows with shadows captured in a photograph or a video. A match would provide information about the location, orientation and time of the image representation. Another use for shadow is that the length, seen on satellite imagery, provides information about the height of built volumes. …

Plume and Smoke Cloud Architecture

After a bomb blasts its smoke plume goes through several distinct phases. The plume forms into a mushroom cloud that slowly dissipates. Studying photographic representation of plumes, Forensic Architecture estimated how long after a strike the photograph was taken. Each explosion from air-dropped munitions results in a smoke plume whose form is unique to the moment and the strike. In this way, Forensic Architecture undertook detailed morphological analysis to identify the same strike in different pieces of footage and to synchronize the footage based on the phase of plume growth being observed. This analysis therefore offered a way of linking evidence together in space and time. The process of synchronization and triangulation helped reconstruct the space and time sequence of unfolding events. Forensic Architecture also measured and compared the size of plumes as captured in different media sources, to compare the plume caused by unknown strikes with known ones.

 One may right away point out that the highly technical modeling and methods used don’t make those techniques reliable or even relevant as such. It’s true. Here are a few brief thoughts though on this emerging methodology and its potential implications for fact-finding in armed conflict and related alleged violations of IHL norms on the conduct of hostilities.

First of all those techniques are not meant to be used in isolation, they are only useful if they are cross-referenced with other sources or if they contribute to answer the questions of “when” and “where” with respect to other types of information. Their purpose is therefore limited and is not to give a full account of a given incident. Secondly their value depends on the context. While they may be helpful to investigate a specific case, when parties to the conflict deny it ever happens, they also prove to be very valuable to describe the extent and pattern of a series of attacks, such as in the context of the activation by IDF of the “Hannibal Directive” in Rafah. This very doctrine relies on the use of massive firepower to respond to the capture of a soldier.

Architectural techniques bring a different perspective to reconstruct those events. Most importantly they can provide more precise information on the characteristics of those attacks, such as the type of weapons used, a mapping of the number, geographical scope, the frequency and impact of such attacks.

Those features are in themselves not enough to establish whether violations of IHL norms on the conduct of hostilities occurred. But the potential architectural techniques offer to travel back in time is particularly valuable during armed conflicts. Of course they won’t provide information on whether a given civilian building was used for military purposes at the time it was hit. But, in a specific context such as the Hannibal Doctrine they can contribute to highlight the gap between the military purpose of that tactic and the characteristics of the attacks and its impact on civilians and civilian objects. In itself this doctrine already raises serious concerns as to its legality under IHL if each attack carried out to implement this tactic is not based on the identification of a particular military objective. Architectural techniques helped to build a detailed factual account to support the claim that some of these attacks were at the very least disproportionate or otherwise indiscriminate on other grounds.

Leave a Reply