Thursday, August 20, 2015

Homeland Security UFO Video Analyzed

UPDATE: I made an error or two and have created a graph showing proper triangulation at the bottom of this page, with corrections, and with notes on conclusions!

A Puerto Rico "UFO" video has been making the rounds. I analyzed the video after it became a hot topic on Reddit, and posted about it there, but I felt I owed my readers analysis. You'll have to pardon my lazy writing here, because I'm already pretty talked-out about it as I write. Here is the video:


It looks weird, doesn't it? It's IR video, of course, so even a video of a human looks otherworldly in that context. Honestly, I was more than a little surprised the skeptics (some of them, at least) seemed flummoxed by this video. I'm not a skeptic, just an open minded investigator, and UFOs are not my thing, but still...

The video is from 2013 but this month the Science Coalition for Ufology released its lengthy 161 page report detailing their analysis of the video. Their conclusions were a bit surprising to me, since I immediately saw a problem with how they derived their numbers and concluded that this was an anomalous event - an Unidentified Aerial Phenomenon. Now, as I said, I'm a bit talked out on this subject. I didn't want to get into it because it's just not my "thing," but since I saw some errors that nobody else seemed to notice, I felt compelled to point them out.  So pardon the fact that I will be copy-pasting some of my Reddit comments as part of this article, but I can't handle rehashing it. Sorry.

Claims are made about the bird traveling 86 miles per hour, and later flying through the water at a similar speed. I think there were too many assumptions involved. Just looking at the display, there was an attempted target lock at around 43 seconds into the video (the box that briefly pops-up around the crosshairs) - this fails. This is a clue in itself: The target had plenty of heat signature, and once in the crosshairs, the operator pulled the trigger, but the lock failed. This was an immediate clue that we were looking at a very small target; one below the threshold of the software for target tracking. We understand that. Mostly these IR systems are used to lock onto aircraft, or cars, or sometimes people. But you don't want them accidentally  locking on a very small target such as a bird. Now the lack of a lock is important, because it means that all the target readings on the lower right display were not about this target. Instead, they are a mean average for the ground. This means that the coordinates cannot be used to establish any speed estimates whatsoever. And We will address that further, because it led to a fundamental error on the part of the SCU investigators.

The aircraft was turning and climbing during this, starting out heading WNW (302 degrees) and ending up going SE when they lost the target. The target was North of the aircraft for this entire event. They increased their distance to the target throughout this. I think it was 2.7 nautical miles downrange on lock attempt, and they were 3.4 nautical miles down range as the bird crossed the shoreline, then 4.0 nautical miles downrange when the bird lands in the water.

IMPORTANT POINT: As alluded earlier, the coordinates shown on the video display are those of the far distant ground, not of the target - and I will conclusively prove that at the end of this commentary. Because the investigators used the coordinates of the land, they got a total arc of 2.2 miles, and that is where they came up with their ground speed. Furthermore, they used the moving ground coordinates to try to establish a movement speed "underwater" while viewing a stationary target. The coordinates continue to move as the airplane moves, however, they have NOTHING to do with any target you are looking at, unless the target is locked! But back to the video for the moment....

You really catch the shoreline at 1:55 in the video; aircraft is still turning left, now at 155 degrees. Target is estimated 16 ft above ground, 3.4 nautical miles downrange (but of course, that's the LAND not the bird). Right after this is when we zoom in, at 1:58. At 2:00 to 2:04, you can really make-out the shape of a pelican: big curved neck, long bill, and wings in that classic glide arc. Now we're 3.7 nautical miles down range. Some people have objected that you can't see wings, but you can - only at certain points. You have to remember that this is infrared video; it's reading heat signatures. Since wings are thin and have low vascularization, they have a very low heat signature. To infrared, they are almost invisible.

At 2::05 you see the bird land in the water. 2:06 they lose the target when the bird lands in the water (heat signature disappears, of course). 4 nautical miles down range: Target is North/northeast of the aircraft, but the aircraft is flying SE. Now we have some controversy as the operator supposedly tracks the target underwater, moving as fast as they thought it was moving in the air!

I don't see a bird flying underwater (or a UFO). I see a heat signature artifact that appears briefly on screen after the camera operator loses the bird (because it landed in the water and the heat signature disappeared). You'll notice it's no longer a heat signature (dark) at that point, but a cold signature (white). He does seem to spot a floating pelican momentarily but perhaps didn't realize that was what he was following to start with, so he zooms back out again, to hunt.. The changes in coordinates you see on screen are describing the visible sea area as the plane moves away and turns, that's all. I will restate yet again, that the ground coordinates, when there is no locked target, describe a mean average of the ground (or sea) being views, and they are constantly in motion because the plane is moving. These coordinates cannot be used to make a speed estimate. We proved that there is no target lock and the coordinates describe only the visible landscape/seascape. About the waves and speed: They seem to be moving really fast (all of the waves, you'll notice) but that's because of the aircraft speed and the foreshortening of the long telephoto lens, not because they are turbocharged ocean waves.

Finally, zoomed out, the camera operator (while panning) spots two birds taking off from the water at about 2:32 and he zooms in again. There is another important point here: who says one of these two birds is the same one as the camera guy followed to the water? Pelicans are ubiquitous in the area and there could be thousands of them out there; all invisible to the IR because their feathers are at water temperature. That assumption was another mistake, and it is one of the assumptions behind the alleged underwater flying, along with the the error displayed throughout the analysis: Thinking that the ever-changing coordinates described a target, which they do not..

..Bob, the other half of udebunked despite the fact that he never writes anything, commented to me the thing I was thinking: This seems like  a very inexperienced IR operator. Perhaps his first-ever field work? Birds are seen on IR all the time. Everybody knows that the wings are hard to see when looking down from above with a warm (high heat signature) land background. Everyone knows their movement profile. What was this person's problem? We'll probably never know. This is part of the real mystery: If you believe the SCU report, nobody in the Homeland Security service or Air Force could figure it out... so they just gave it to UFO people. Yeah, that's believable... (wiping smirk off face) as in not believable at all. With the Air Force's IR profile database, the match could have been made by computer in like a tenth of a second. So then the real mystery is why some bird footage was proffered to UFO groups by the government..what were they up to? Some kind of test? If so, the UFO people failed that test. Maybe that is what the government wanted. But I digress....

---

Personally, I don't see any indicators of high speed at all. The heads-up display is misleading people into thinking the bird is covering a huge distance, but it's not a locked target so we're getting coordinates of distant ground. Those coordinates are always in constant motion, even when viewing a stationary but unlocked target. Don't believe me? Look at the first part of the video, before the operator starts chasing a bird. The ONLY way you can arrive at a speed estimation is to find a stationary ground target with a known width, and use that to triangulate vertically and horizontally, to establish a speed.. When you do that, it doesn't look fast at all.

Then we get to the landing in the water, which we actually see at 2:05, followed immediately by the loss of heat signature (and the white artifact) at 2:06. Again, no speed really. The waves look fast, because of foreshortening by the enormous zoom (remember the aircraft is now over 4 nautical miles away, so that's like equiv. 2,000 millimeter lens) and the effect of the aircraft making a left turn, which means the waves (like the bird) are being videoed in a moving arc, with the aircraft's speed being added to the apparent ground and wave speed. You could calculate how much apparent speed the aircraft is adding to the bird by tracking the turn of the aircraft throughout the video (goes from WNW around to the SE and finally to the NE in the same short time slice) - doing a sight line and tangent analysis - but I'm too lazy to do that, and it would be highly problematic anyway because altitude of the target would be a critical factor (and we don't have that).

The ground arc is a total of 2.2 miles, but much of that is accounted for the the turning aircraft and the fact that those coordinates are of the far distant ground (and constantly in motion because of the aircraft), NOT the bird.

 * ... To give an example of this: Bird crosses a two lane highway (airport runway) at 1:14 to 1:16 -  (coordinates 18 29 51N 67 07 04W ) to 18 29 54N 67 07 08W Google map, (THIS IS HIGHLY SIGNIFICANT EVIDENCE, as you will see) which shows that is a distance of 294ft. 294 feet covered in 2 seconds - IF the bird were flying at 0 feet above the ground and we were viewing the road from directly above - that's the only way their measurements would work!

But we know that isn't true: the bird is in the air and being viewed from 2.1 nautical miles away from an altitude of 2100 FT (corrected in graphic), and the road (airport runway) is seen at a very oblique angle, which tells you that the bird is closer to the aircraft than the coordinate estimates would imply. Applying a little trigonometry, it appears that the bird is covering about 65 air feet (corrected in graphic) in that same two seconds, yielding a lazy speed of 22.36 mph (corrected in graphic). So yeah, the speed and turn of the aircraft, the long-distance view and foreshortening of the zoom lens is giving a very misleading impression. Also note that while the road can be checked to confirm it is 294 feet, the displayed coordinates travel almost 500 feet as the bird crosses that road. Again, the coordinate motion, when there is no locked target, includes the relative speed of the aircraft.

Note on speed: You have multiple points which can be triangulated. You must use stationary ground targets of known width. Alternatively, as I said, sight line and tangents give you another way to get at a number, but they are again enslaved to the inaccurate coordinates and therefore problematic. The runway above has a known width. The aircraft had a known position when the bird was observed crossing in front of that road (airport runway). The aircraft was  2.1 nautical miles from the road (12760 feet) and 2100 feet in altitude. With that, application of Pythagorean theorem gives us the unknown side of the vertical triangle: 12932 feet. We will need that late for the speed calculations. Now we determine the horizontal triangulation: the space through which the bird actually flies, as you can see in the video. So we take the 12932 feet (Corrected in graphic) from aircraft to target (the ROAD, remember) and the width of the road (easily obtained from Google maps) which is 294 feet, and get our other side. Now we have a horizontal triangle, which is 294 feet wide at ground level, and makes a point at the aircraft. In other words, the width of the triangle gets smaller as we approach the aircraft.  Finally, we have to estimate the height of the bird. That's  a little fuzzy, but since elsewhere you see clearly the tops of buildings with the bird far above, you know it isn't skimming the land. I chose 100 feet. It was probably higher. If you go with 200 feet, the bird is traveling a little even slower, and if you go with 800 feet - then we're down to a walking speed. It isn't going 86 miles per hour, that's for sure.  So then we find our vertical height lines on the vertical triangle, and that gives us a new base for the horizontal triangle. I arrived at a 65 foot window that the bird crossed, as I said. There's fuzz here, but the fuzz factors are altitude and angle passing through the viewing area. At their ends, a perpendicular passage will give you 15 MPH at their suggested height, diagonal passage gives you 44 MPH and diagonal and diving from altitude gives you 72 MPH (however, in that case we would have to see our bird moving parallel to the lines of sight, which is not what is happening. I only mention this to illustrate that there is an outside speed that can be derived, which is still far lower than the SCU report estimates).. Nothing higher than that. Sorry, guys.

 ABOUT THE PROOF of the coordinates error: As you see from the Google map, plugging those coordinates into Google maps lands you directly on the road which is off in the distance and only seen obliquely, as you can see. Plug in the coordinates shown at any point, and you'll get the objects seen in the far distant background. THIS conclusively proves my original point that the coordinates shown are of the far distant background land, not the bird (otherwise those coordinates would put you on land in between the aircraft and the road, and not the road). Furthermore, you  can observe in this case that the moving ground coordinates always show a greater distance than we know is involved. The road, for example, being a known 294 feet but the ground coordinates reporting a much greater distance. The erroneous speed estimate was based on those land coordinates, yielding 2.2 miles covered/86 mph-ish, but the bird is between that far-distant land and the aircraft camera, and and probably flies under 3/4 mile in that same time frame. SCU used ground coordinates to establish speed, but we've already proven that these are useless for this process. I'd say that this is pretty much conclusive. Sorry, though, I'd rather it be something more interesting!

Mistake 1: Misunderstanding and failing to acquire the public information available about how the IR display rangefinding works. If they had done this, they would have understood that the lower right coordinates do not refer to a target, unless the target is locked. THE TARGET WAS NEVER LOCKED. Therefore, the numbers are constantly in motion even when looking at things like houses on the ground (which obviously aren't moving) because the aircraft is moving. Those number always show motion. They cannot be used to make speed calculations. All of this could have been avoided by simply sitting down with any experienced military IR operator: It would have taken 30 seconds for him/her to say, "It's a bird," and none of us would have had to go through this tedious and unfortunate process.

Mistake 2: Having misconstrued the readings, they then used those readings to estimate speed of the bird. And they used their wrong estimate to rule-out the fact that this is a bird. Speed estimations can only be made using stationary ground targets with known widths, and triangulating on those, using the altitude and distance of the aircraft. Once again, an experienced air-to-ground IR operator would have stopped this train wreck before it ever happened.

Mistake 3: When the bird lands, because they didn't understand the meaning of the display, they misinterpreted the target as still moving. Again, that data is always in motion unless there is a locked target which is stationary ...which never happened in this case.

Mistake 4: The investigators of SCU conflated a later sighting of two birds with the original bird sighting. They did this probably due to confirmation bias, since they had already set the idea in their heads that stationary targets were actually moving, due to their erroneous interpretation of the display information, then thinking that the same bird later emerged from the water further down range made sense to them and confirmed their wrong assumption about the motion and speed of the target..

Verdict: Debunked. The SCU report concluded that this was an Unknown Aerial Phenemon and an "Unknown Submerged Object" based on misunderstanding and misinterpreting the on-screen data. Once a proper speed analysis is done, there's nothing left to hang that UAP hat on, and once one realizes that they conflated two unrelated sightings to get the "submerged" moniker, there's nothing to left of that, either. It's probably a bird, and in that case, almost certainly a pelican. The investigators who analyzed this video assumed that the on screen coordinates were of the bird or related to the speed of the bird in some way, but I've proven that they are of the land in the far distance and show a constant motion factor even on stationary targets, because there is no target lock. The error of using target coordinates where there was never a locked target simply led to a cascade of errors, and all the mistakes which followed. Does anyone know how many pelicans there are in that area at that time of year??? Nobody checked, I bet. With the coordinates, the analysts could have just plugged the coordinates into Google and found out that they were not of the target but of the far distant land, but they didn't think of it.

In other words, despite issuing a lengthy scientific-sounding report, the investigators didn't really do their homework and didn't employ proper investigation procedures, nor did they apply accepted, appropriate analysis techniques.

The SCU investigators didn't know  how the coordinates displays work when there is no locked target. They didn't understand infrared heat signatures and their limitations. They probably didn't look up information about pelicans in the area and they certainly did not do the triangulation of stationary ground targets  necessary to determine the actual air speed - and we know that for a fact: The speed numbers they stated can ONLY be derived from the ever-changing coordinates (run the math and see for yourself) and those coordinates were of the distant ground not the bird, and were always in motion even when viewing stationary targets because of the actions of the aircraft. Proof that they did not use a triangulated ground target distance measurement. Even when they note in their report that they thought the coordinates were of the distant ground, they still used those ground coordinates to estimate speed instead of triangulating known, identifiable stationary ground targets.

You could still call it a UAP, of course. It looks like a pelican but that's subjective! I think the innocent mistakes in interpreting on screen data just messed you up and caused you to rule-out the likelihood that this was a bird from the get-go. I'm really sorry, guys! Stuff happens....

We all make mistakes. But next time, check out the details instead of assuming so much. I'm just an ordinary person, not an important scientist, but even I could immediately see where you went wrong. Mistakes were made. Someone (me) noticed. And that's the end of the story.

---

Here is the 2:02-2:04 part of the video where you can see that it is a pelican:


And here is a zoom-in of the same time sequence. Use the square on the lower right to go full-screen, and the circle on the lower left to repeat... and repeat, until you get it. :)


General video timeline (of the YouTube video):
  • 0.22 - bird flies into frame at upper right, with the brief appearance of apparent flapping.
  • 0.23-25 First opportunity to see wings flapping. Yes, wings are hard to see in IR, particularly when viewed from far above and at great distance, but try harder!
  • 1:14-1:16 - Bird crosses identifiable road. This was the critical data point: The coordinates entered into Google map prove that the tracking is of the far distant ground, not the bird. We can then get the width of the road and create X and Y triangulations to estimate the bird's distance from the plane and actual speed, which is far lower than the estimates created from taking the coordinates, erroneously assumed to be those of the bird.
  • 1:27-1:28 Bird crosses another road in the far distance. Yet another set of triangulation points and proof that the coordinates are of the far distant land, not the bird.
  • 2:05 - bird lands in water. IR artifact (white) remains on screen for a moment after this.
  • 2:11-2:13 - Bird is briefly spotted, floating. Note again that coordinates continue to (with the clearly unmoving bird) only because they are readings of the sea level, being read in an arc as the plane continues to turn. We've already proven that these are not measurements of the target. No point in trying to claim that they are otherwise.
  • 2:21 - bird head barely visible here and there, just before operator zooms out at 2:22. I don't know how many times one has to say: You can see the target is not moving (other than bobbing up and down in the waves), the ground coordinates continue to change because the aircraft is now 4 nautical miles down range and the visible seascape is changing with the arc of the airplane's turn. There's nothing more to it than that.
  • 2:22 - Operator zooms out and begins panning.
  • 2:26 - You can see a bird in the air, near the sea level, to the right of the crosshairs.This bird continues to fly and the operator follows it. The bird becomes more clear as it flaps its wings and develops more heat signature at 2:33.
  • 2:36 - Operator zooms in and you can now see that there are actually two birds. Probably neither is the same bird as we see in the first part of the video. There is an illusion of fast motion caused by 1) the foreshortening due to zoom lens usage, which makes the waves seem to be moving fast, even though you can see the target is not, and 2) the ever changing coordinates on the lower right, changes which match the changes of the aircraft coordinates as it turns and moves away, represent reciprocal values of each other. Once again, those numbers are not about the target, because there is NO locked target; they describe the area of sea being looked at, and that is all.
UPDATE: Graphic of proper triangulation!  I did make an error in the blog, which is corrected in the graphic. Sadly, it does not help SCU any. Bummer. I even went with their 40' height estimation for the bird, which seems too low but whatever...



---
DISCLAIMER: Nothing in this post or blog is meant to suggest that the supernatural or paranormal does not exist. The fact is, I don't know. All I can do is look at the evidence to see if it stands up to the test of very simple scrutiny. If it passes, yay. If it fails, oh well... there's always next time. Investigators are human. I'm human. We all make mistakes. My pointing out the mistakes that others make should not be taken to imply that they are wrong in general, or that the things they investigate are only the product of imagination.