A U.S. Customs and Border Protection aircraft tracked a small infrared target near Rafael Hernández Airport. The video appears to show high-speed travel, a water entry and a split into two objects. Years later, AARO published a very different reconstruction.
Aguadilla is one of the best examples of why range is everything in UAP analysis.
The footage itself is not the controversy. The recording is authentic government thermal imagery.
The argument is over where the target was.
If the object was close to the camera and low over the terrain, its apparent motion implies higher speed and can make the later frames look like a genuine water entry and re-emergence.
If it was much farther away, the same line-of-sight movement can be reproduced by a slow windborne object while the aircraft carrying the camera creates most of the apparent motion through parallax.
Authentic footage does not settle geometry. Every speed, size and “transmedium” claim depends on where the object is placed in three-dimensional space.
A Customs and Border Protection aircraft operating near Aguadilla tracked a compact infrared target moving through the sensor field.
The target appears as a small thermal blob rather than a detailed craft. No wings, rotors or fuselage are resolved clearly in the public imagery.
Because the platform itself is moving and circling, the background slides rapidly behind the tracked target.
This creates the central analytical trap: image motion is not target speed.
Target acquired near the airport environment.
Object appears to cross roads, structures and vegetation.
Line of sight approaches the shoreline and ocean.
Image appears to bifurcate into two thermal returns.
This is the feature that made Aguadilla famous.
In the extraordinary reading, the object crosses the shoreline, enters the water without obvious deceleration or splash, continues beneath the surface and later emerges again.
If true, that would make Aguadilla a rare filmed transmedium event.
But infrared perspective can be deceptive.
A distant object moving behind the horizon line or across a thermal background transition can appear to intersect the water even if it remains airborne.
AARO's current reconstruction explicitly rejects the water-entry interpretation and says the objects never entered the ocean.
Late in the recording, the thermal target appears to divide into two similar returns.
This has been interpreted as physical division: one object splitting into two while in flight.
AARO's current explanation is different. Its Intelligence Community partner reconstruction says two objects were travelling close together and the sensor only resolved them separately later.
Other conventional possibilities include infrared bloom, changing resolution, focus, contrast processing or overlapping objects becoming distinguishable as viewing geometry changes.
Without higher-resolution raw imagery, the frames cannot independently prove that one physical object literally divided.
The Scientific Coalition for UAP Studies published a detailed technical analysis arguing that the target followed an anomalous path, moved faster than a typical windborne object and appeared to enter the water.
SCU used telemetry, scene geometry, radar/context material and frame analysis to construct a trajectory.
It rejected simple lantern/balloon explanations, arguing they did not reproduce the line of sight, motion and apparent water behaviour adequately.
That report is important because it moves the case beyond impressionistic viewing and into explicit modelling.
Its conclusions, however, depend on assumptions about distance and object location.
AARO now lists the Puerto Rico/Aguadilla object as a resolved case.
It assesses with high confidence that the objects did not demonstrate anomalous speeds or flight behaviours.
According to AARO, an Intelligence Community partner reconstructed the observing aircraft's flight path and sensor look angle.
That reconstruction places the targets on a straight wind-driven path rather than a high-speed manoeuvring trajectory.
It also says two objects were travelling near each other from the start, rather than one object splitting.
Most importantly, AARO says the targets did not enter the water.
This materially changes the modern status of Aguadilla. The case is no longer accurately described as officially unresolved.
When a camera aircraft circles a distant target, the line of sight changes rapidly.
The tracked target can therefore appear to race over the ground even while drifting slowly.
This is parallax.
If the range is underestimated, reconstructed speed is exaggerated. If the range is increased, a windborne trajectory can fit the same angular motion.
That makes balloons, lanterns or similar airborne objects serious candidates.
The strongest conventional models also use the fact that the target remains small and unresolved throughout the video.
The weakness is identification: the public record still does not place a specific recovered balloon or lantern into the case with a complete chain of proof.
Object is close to the aircraft and low over the terrain. Apparent image motion becomes real high speed; shoreline crossing becomes physical water entry.
Object is farther away. Ownship motion generates much of the apparent speed; the target stays airborne and follows wind.
Both models can use the same frames.
The dispute is therefore not about whether the pixels exist. It is about converting two-dimensional sight lines into a three-dimensional trajectory.
Reports connected to the airport and surrounding airspace have been used to argue that radar detected anomalous traffic during the same period.
But a radar return only strengthens the video if it can be tied to the exact thermal target with matching time, azimuth, range and altitude.
The public case does not provide a universally accepted one-to-one radar track for the object shown in every frame.
This is why “radar corroborated Aguadilla” is too strong without qualification.
The proper question is which radar contact, at which second, corresponds to which line of sight from the sensor aircraft.
Thermal sensors display heat contrast, not ordinary visible-light outlines.
A small warm object can create a larger apparent blob through point-spread, bloom and processing.
That means dimensions should not be read directly from the visible black or white patch.
The apparent split is also sensitive to resolution. Two close thermal sources can merge into one unresolved patch and later separate as angle, focus or contrast changes.
This is another reason raw sensor data and processing history matter.
Speed requires range; range is disputed.
AARO says the targets never entered the water.
AARO says two objects travelled together.
No physical target was recovered for identification.
The best available public media chain remains imperfect.
Sensor-video/radar correlation remains contested.
AARO's resolution changes the evidential landscape because it introduces an official geometric reconstruction.
But a strong Case File should not simply replace one authority with another.
The correct test is reproducibility.
Can independent analysts recreate AARO's trajectory from released data? Are all assumptions about platform position, look angle, wind and target distance explicit? Do those assumptions fit the visible landmarks throughout the clip?
Likewise, SCU's anomalous reconstruction should be tested under the same standard.
The value of Aguadilla is that both sides can be asked to show their geometry.
Aguadilla looks extraordinary before the geometry is solved.
That visual impact is real and should not be dismissed.
But if a distant windborne path can reproduce the footage without water entry, high speed or physical splitting, then the extraordinary interpretation loses much of its force.
The remaining mystery becomes narrower: what exactly were the two objects?
That is less dramatic than a transmedium vehicle — but still a legitimate identification question.
The Customs and Border Protection aircraft was not a fixed camera on a tripod. It was flying, turning and changing its line of sight while the sensor operator kept the target near the centre of the frame.
That creates a classic moving-observer problem. When the camera platform moves laterally relative to a distant object, nearby terrain can sweep across the image quickly even if the target itself is drifting slowly.
This is why apparent speed in the video cannot be read by simply watching how fast the target seems to cross roads or coastline features.
To calculate true target speed, analysts must know the aircraft's position and velocity, the sensor azimuth and elevation, and the target's range. Errors in any of those inputs propagate directly into the reconstructed path.
Aguadilla is therefore a rare case where geometry is not a side issue. Geometry is the case.
The recording took place in an aviation environment where unidentified traffic matters for ordinary safety reasons before any UFO interpretation begins.
Rafael Hernández Airport serves civil and government operations, and the surrounding airspace includes aircraft, airport lighting, weather phenomena and potential windborne objects.
The airport setting strengthens the value of contemporaneous records because flight movements, weather observations and operational logs can potentially be cross-checked against the thermal recording.
It also makes overstatement easier. A temporary operational response or reported radar contact near an airport can become “the airport shut down because of the UFO” in later retellings unless the exact causal chain is documented.
The Case File should therefore keep airport operations as contextual evidence unless a primary record directly connects a specific action to the tracked target.
Different analysts can derive dramatically different speeds because they are solving for different distances.
If the target is placed close to the observing aircraft, the angular movement across the frame converts into a relatively high linear velocity.
If it is placed much farther away, the same angular motion can be produced by a slower object while the camera aircraft generates most of the background sweep.
This is why speed claims should always be paired with the assumed range.
A number without its geometry is not a complete result.
The strongest analysis therefore does not ask “How fast does it look?” It asks: “What three-dimensional path best fits all lines of sight over the full video?”
One of the research flags on Aguadilla is the absence of a universally available original lossless sensor file with complete provenance and transformation history.
That matters because stabilisation, compression, contrast adjustment, frame extraction and re-encoding can alter edge structure and apparent morphology.
For ordinary storytelling, those differences may be minor.
For a case where the central claims involve whether one blob becomes two, whether the target briefly disappears, and whether a boundary represents water entry, those differences matter a great deal.
Any enhanced or stabilised derivative should therefore be labelled as such, and measurements should be made from the least-processed source available.
When two warm objects are close together and below the sensor's resolving limit, their thermal signatures can merge into one apparent source.
As viewing angle or separation changes, the combined blob can resolve into two lobes.
To the eye, that transition looks like one object dividing.
This is not a hypothetical quirk invented for Aguadilla. It is a standard imaging problem whenever point sources approach the resolution limit of an optical system.
The competing extraordinary interpretation remains possible only if analysts can show that one physical source existed before the split and then physically became two after the split.
The public footage does not provide direct depth information capable of proving that sequence by itself.
A craft moving from air into water without slowing, splashing or changing behaviour would be extraordinary.
That is why Aguadilla became so prominent in UAP discussion.
But extraordinary behaviour must be demonstrated from the data rather than inferred from visual coincidence.
If the target merely passes behind the apparent shoreline from the camera's perspective, then no transmedium event occurred at all.
The distinction depends on target range, camera line of sight and the three-dimensional relationship between the object and coastline.
AARO's present resolution directly attacks this claim by stating that the reconstructed targets remained airborne throughout.
Any future extraordinary interpretation therefore has to defeat that geometry, not merely replay the video.
It is tempting to turn the dispute into a battle of authority: government office versus independent researchers.
That would miss the scientific opportunity.
SCU produced a detailed anomaly-focused reconstruction from the public evidence and argued that slow windborne explanations fail key sight lines.
AARO now says an Intelligence Community reconstruction produces a straight, wind-speed path with two objects and no water entry.
The strongest way to compare them is not by reputation but by inputs.
What coordinates did each model assign to the target at each time? What wind data did each use? What camera calibration? What assumptions about altitude? What error bars?
If both sides publish enough detail, Aguadilla can become a genuine reproducibility exercise.
AARO now considers the Puerto Rico object resolved in the sense that it does not assess anomalous speed or flight behaviour.
That is a major evidential development and should be represented clearly.
But for a technically literate reader, the useful question is how that conclusion was reached.
Resolution is strongest when the underlying reconstruction can be independently reproduced and when the remaining object identity is narrowed to a plausible class.
If AARO's geometry is correct but the exact objects remain unidentified, then the case changes category: from alleged transmedium craft to unidentified windborne objects.
That is a much smaller mystery, but still an identification problem.
Aguadilla still has qualities that make it worth close study.
The footage is continuous rather than a single blurry photograph. The camera platform is known. Telemetry exists. Geographic landmarks can be identified. Competing models can be tested numerically.
That makes it unusually suitable for transparent analysis.
Even if AARO's reconstruction is ultimately correct, the case remains an excellent example of how an authentic military or law-enforcement sensor recording can look extraordinary when range is unknown.
If AARO's reconstruction is incomplete or wrong, the same dataset gives researchers a path to demonstrate exactly where it fails.
Either way, Aguadilla is more useful than a mystery that cannot be tested.
Aguadilla remains one of the most useful UAP videos ever released because it can be analysed quantitatively.
The footage is authentic. The target is real in the sensor data. The dispute is about interpretation.
SCU built an anomalous trajectory involving high speed and water interaction. AARO now says a reconstructed geometry shows two windborne objects travelling straight, never entering the water and never demonstrating anomalous behaviour.
Aguadilla is no longer just a mystery video. It is a test of which three-dimensional reconstruction survives the data.