An object appears to race above the ocean.
Another seems to split in two and enter the water.
A third appears to leave a strange wake behind it.
These are powerful images.
They are also examples of why appearance and physical behavior are not always the same thing.
Case One: Go Fast
The Go Fast video was recorded by a U.S. Navy F/A-18F sensor in 2015.
To a viewer, the object appears small, low and extremely fast.
AARO's later reconstruction produced a different interpretation.
The sensor display contained range, sensor azimuth, sensor elevation, aircraft altitude, aircraft speed, and bank angle.
The aircraft's exact heading and position were not preserved in the available file.
Instead of pretending those missing values were known, AARO modeled the possible range.
Its analysis placed the object near 13,000 feet.
After wind was considered, AARO assessed an intrinsic speed range from roughly 5 to 92 miles per hour.
Motion parallax
The aircraft carrying the camera was moving rapidly.
When a fast observer passes a slower object, the background can sweep across the frame and make the object appear much faster than it is.
This is motion parallax.
In Go Fast, the geometry changed the interpretation from extraordinary speed to non-anomalous performance.
What remained unknown?
AARO did not definitively identify the object.
A case can become less mysterious without becoming fully identified.
Case Two: Puerto Rico
A 2013 infrared recording near Aguadilla, Puerto Rico, became famous because the object appeared to move rapidly, split into two, cross land and water, enter the ocean, and emerge again.
That interpretation suggests an extraordinary transmedium object.
AARO and an intelligence partner reconstructed the sensor line of sight and flight path.
Their assessment was different.
Two objects, not one splitting
The reconstruction indicated that two objects were traveling near one another.
Their overlap in the two-dimensional infrared image created the appearance of a single object dividing.
Apparent water entry
The reconstructed path did not require the objects to enter the ocean.
The sensor's moving viewpoint and background geometry changed where the objects appeared relative to the coastline and water.
AARO assessed that the objects traveled in a straight line near wind speed.
Again, the video itself was real.
The first interpretation of the geometry changed.
Case Three: Atmospheric Wakes
Infrared footage from military platforms showed an object with what appeared to be a strange trailing wake.
The visual could be interpreted as an unusual propulsion signature.
AARO examined longer footage and commercial flight data.
The object was assessed as likely being a commercial aircraft.
The apparent trailing feature was attributed to video compression.
Compression creates structure
Digital video does not preserve the scene perfectly.
Compression algorithms estimate and reuse information between frames.
Edges can smear.
Contrast can trail.
Small moving objects can produce shapes that look connected to them.
A visual artifact can therefore appear to be a physical wake.
Infrared does not mean direct reality
Infrared sensors are powerful.
They are still instruments.
The image is a representation of detected radiation processed through sensor response, gain, contrast settings, color palette, digital processing, and display resolution.
A bright shape is not automatically a hot metal craft.
A dark shape is not automatically a cold object.
The image must be interpreted in context.
Distance ambiguity
Without range, apparent size gives weak information.
A tiny object close to the sensor and a much larger object far away can occupy the same visual angle.
Distance uncertainty spreads into estimates of size, speed, altitude, and trajectory.
Missing metadata
A video may look technologically advanced while lacking the variables needed for analysis.
AARO's Go Fast methodology notes that the original file and metadata were no longer available.
The public compressed video still contained enough display information for a range of calculations, but not enough for one exact solution.
More pixels do not replace missing geometry.
Witness perception
This does not reduce a witness to being unreliable.
A person can sincerely report exactly what an event looked like.
The mistake happens when appearance is silently converted into measured physical performance.
AARO's historical review distinguishes credibility from reliability.
Someone can be honest while the observation remains affected by perspective, memory, expectation, or limited sensory information.
Misidentification is common
AARO's public case pages include reports resolved as balloons, birds, commercial aircraft, satellites, and other ordinary objects.
Its current reporting trends show that balloons and satellites make up large shares of closed cases.
That does not mean witnesses were foolish.
Many ordinary objects become difficult to recognize under unfamiliar viewing conditions.
Sensor artifact is not the answer to everything
The opposite mistake is equally weak.
Not every unresolved object should be dismissed as a sensor error.
AARO currently publishes cases where analysts assess with high confidence that a physical object is present but cannot attribute it more specifically.
It also publishes cases where the data are insufficient to determine whether an apparent infrared signature is physical or an artifact.
Different cases have different evidentiary states.
Genuine uncertainty
Genuine uncertainty exists when several explanations remain compatible with the data.
That state should not be filled with a preferred story.
A case can be unresolved, physically real but unidentified, visually ambiguous, insufficiently documented, unremarkable in performance, or worthy of additional data.
Those categories are more informative than one dramatic label.
Why case resolution matters
Resolved cases teach investigators what can imitate anomaly.
That improves future analysis.
A known balloon can become a comparison set.
A known aircraft can reveal how infrared looks at distance.
A compression artifact can help identify similar trails.
Ordinary explanations are not wasted investigations.
They improve the model used to recognize genuinely unusual data.
KEY TAKEAWAYS
What to Carry Forward
- Apparent motion is not always physical motion.
- Go Fast demonstrates how parallax and geometry can alter speed interpretation.
- The Puerto Rico case demonstrates how line of sight can alter apparent splitting and water entry.
- Atmospheric Wakes demonstrates how compression can imitate a physical trail.
- Infrared imagery requires interpretation and context.
- Missing range or metadata limits size and speed estimates.
- A sincere witness can still misjudge physical parameters.
- Sensor error is one possible explanation, not a universal dismissal.
- Some observations remain genuinely unresolved because the data cannot distinguish among alternatives.