Solar Ephemeris Models and Shadow Geometry Enable Fact-Checkers to Timestamp Video Footage Without File Metadata

A video stripped of its metadata no longer hides when it was filmed. For decades, digital investigators relied on the hidden text files embedded in digital media—EXIF data—to determine the exact time and date a recording was made.[6]

But metadata is fragile. It can be stripped by social media platforms upon upload, or intentionally altered by malicious actors using basic software tools. When the digital breadcrumbs disappear, fact-checkers turn to the physical world. The technique is known as chronolocation, and it relies on a clock that cannot be hacked or spoofed: the solar system.[6]

By analyzing the geometry of shadows cast in a video frame, investigators can calculate the exact position of the sun. This process transforms any vertical object—a streetlamp, a building, or a person—into a sundial. Because the sun's trajectory across the sky is entirely deterministic, its position at any given moment is a unique mathematical signature.[1][7]

If you know where a video was filmed, the shadows will tell you exactly when. The discipline has moved from a niche intelligence gathering method to a foundational pillar of open-source journalism. Organizations like Bellingcat use shadow analysis to verify footage from conflict zones, exposing manipulated timelines and fake news.[1][8]

It is a physics-based invariant that forces digital forgeries to obey the laws of celestial mechanics. Generative artificial intelligence models can hallucinate pixels with astonishing fidelity, but they do not calculate the sun's position. If a video claims to show a midday protest in December, but the shadows are short and tight beneath the protesters, the geometry exposes the lie.[8]

The Geometry of Chronolocation

Chronolocation relies on two specific angles: solar azimuth and solar elevation. The azimuth is the compass direction of the sun, which dictates the direction a shadow falls on the ground. If a shadow points due north, the sun is due south, which immediately anchors the time of day to solar noon in the Northern Hemisphere.[2][7]

The second angle, solar elevation, measures how high the sun is above the horizon. This dictates the length of the shadow relative to the object casting it. A low sun casts a long shadow, while a high sun casts a short one. By measuring the ratio between an object's height and its shadow, investigators can calculate the elevation angle.[2]

These two angles form a coordinate pair that maps to a specific time and date. The mathematical relationship is governed by solar ephemeris models, which calculate the exact position of celestial bodies. The function runs in both directions: if you know the time and location, you can predict the shadow; if you know the location and the shadow, you can recover the time.[2][3]

To extract these measurements from a two-dimensional video frame, investigators use trigonometry. The tangent of the sun's elevation angle is equal to the object's height divided by the shadow's length. This requires finding a clear vertical reference point and a flat surface where the shadow falls without distortion.[2]

Tools of the Trade

The default shadow-analysis tool in the investigative journalism community is SunCalc. Originally developed as a simple mapping application, it models the relationship between date, time, geographic location, and the sun's position. Users can drop a pin anywhere on a digital map interface and drag a time slider to see the corresponding shadow angles.[1][3]

SunCalc provides the theoretical sun position with negligible error, relying on established astronomical algorithms. It outputs the exact azimuth and altitude for any minute of the year. When investigators input the height of an object, the software instantly calculates the expected length of its shadow, providing a baseline to compare against the video evidence.[2][3]

The tool is so foundational that Bellingcat integrated the SunCalc library into its own proprietary Shadow Finder tool in 2024. While SunCalc requires the user to know the location to find the time, Shadow Finder automates the reverse process. It searches the globe for locations that match a specific shadow measurement.[3]

For highly sensitive investigations, researchers often move offline. Because web-based tools rely on server requests that can expose an investigator's coordinates, security-conscious fact-checkers use offline solar calculators. Tools like the Python pvlib library or offline planetarium software like Stellarium provide the same ephemeris data without leaving a digital footprint.[3]

Exposing Digital Forgeries

The power of chronolocation is best demonstrated in conflict zones, where the information war is as fiercely contested as the physical battlefield. In 2020, Bellingcat used shadow analysis to investigate the execution of two Armenian men in Nagorno-Karabakh. By measuring the shadows in the execution video, they corroborated the timeline of the atrocity.[1]

The technique is equally effective at debunking false claims. When a video circulates claiming to show a recent military strike, fact-checkers first geolocate the terrain. Once the location is fixed, they measure the shadows. If the shadow geometry aligns with a summer sun trajectory, but the video is published in winter, the footage is exposed as recycled.[4][7]

Consider a case study from Madrid's Puerta del Sol. If a video surfaces claiming to show an event at noon on November 22, the solar elevation should be approximately 29.6 degrees. At that angle, a 1.8-meter person must cast a 3.16-meter shadow. If the video shows a shadow equal to the person's height, the timestamp is a fabrication.[1][8]

This level of precision turns chronolocation into a definitive forensic standard. It is not about guessing the time of day based on the lighting; it is about proving it mathematically. The shadow vector serves as a celestial anchor, tying the digital stream to the physical rotation of the Earth.[8]

When metadata is present, chronolocation serves as a vital cross-check. If the EXIF data says a photo was taken at 8:00 AM, but the shadows point east, indicating a setting sun, the metadata has been tampered with. The physical evidence always overrides the digital label.[6]

The Limits of Shadow Analysis

Despite its mathematical certainty, chronolocation has strict environmental limitations. The most obvious is weather. Overcast conditions produce diffuse light with no distinct shadows, rendering the technique useless. The sun must be shining brightly enough to cast a hard, measurable edge on the ground.[3]

The geometry of the scene also introduces potential errors. The mathematical models assume a perfectly vertical object casting a shadow onto perfectly flat ground. If the ground slopes upward, the shadow will appear artificially short; if it slopes downward, the shadow stretches.[3]

Investigators must account for topography using three-dimensional terrain models like ShadeMap. Camera distortion further complicates the measurements. Wide-angle lenses, common in smartphones and security cameras, warp the perspective of the image. A shadow that appears to point in one direction at the edge of the frame might actually be pointing in another.[3][5]

This requires investigators to digitally correct the lens distortion before measuring the angles. There is also an inherent ambiguity in solar elevation. Because the Earth tilts on its axis, the sun follows the exact same trajectory twice a year. A shadow measured in early April will look identical to a shadow measured in early September.[5][7]

To break the tie, investigators must look for seasonal clues like foliage, snow, or clothing. Finally, chronolocation cannot verify indoor footage unless sunlight is streaming through a window to cast a measurable shadow on the floor. For the vast majority of indoor content, fact-checkers must rely on other forensic methods.[3][7]

The Future of Verification

As generative artificial intelligence becomes more sophisticated, the role of physics-based verification will only grow. AI models are currently trained on two-dimensional pixel patterns, not spatial geometry or celestial mechanics. They struggle to maintain consistent shadow angles across multiple objects in a single frame, let alone align them with accurate ephemeris data.[8]

This creates a window of opportunity for fact-checkers. By codifying shadow geometry into automated forensic tools, researchers can scan thousands of videos for temporal-celestial mismatches. If the clock in a video moves but the shadow vector remains frozen, the footage can be flagged as synthetic before it goes viral.[7][8]

The discipline is evolving from a manual, artisanal process into an operational science. The integration of ephemeris models with machine learning algorithms will eventually allow for real-time chronolocation. A video uploaded to a news platform could be instantly checked against the sun's historical position.[8]

Until then, the verification of digital media remains a human endeavor. It requires patience, a grasp of trigonometry, and a willingness to look past the pixels. In an era where digital evidence is increasingly suspect, the most reliable witness is often the oldest one we have: the shadow on the ground.[8]