REPORT ON PHOENIX LIGHT ARRAYS by
Bruce Maccabee, originally posted Sept. 30, 1998; updated to March, 2002
Bill Hamilton has presented his (and Tom King's) response to my analysis of the March 13, 1997, 10 PM video sightings at http://home.earthlink.net/~skywatcher22/plvideo.htm. In their report, "The Phoenix Lights Analysis," herein called TPLA, they have argued that my general conclusion, that the lights were south of the Estrella Mountains and much farther from Phoenix than the witnesses thought (e.g., 70 miles rather than 15-20 miles) is wrong. Instead, they claim that the lights were on the north side of the Estrella Range, about 15 miles south of Phoenix.
My estimate of the location and altitude of the lights as well as data on the motion of the lights (downward and to the left) led me to conclude that it was at least reasonable to accept the claim of the Air Force that these lights were probably very bright flares over the Barry Goldwater Air Force Test Range west of Tucson, AZ. Official statements indicated that flares were dropped as late as 10 PM.
On the other hand, H&K claim the lights couldn't be flares because they were the wrong color for (magnesium) flares of the type which were dropped by the A-10 aircraft (type LUU-2B) and because the lights were north of the Estrellas.
This addendum to my report addresses some of the claims and criticisms made in TPLA. I assume that anyone reading this report has read the preceding "Report on the Phoenix Light Arrays" which is referred to herein as RPLA. For the convenience of the reader some information from these reports will be repeated here, but to get the full "flavor" of the analysis one should study the other reports before reading this.
H&K end their paper with the statement that they did not see the January 14, 1998 light arrays and therefore they would have no comments on it. I can well understand the reluctance of a ufologist to discuss a sighting which he is not directly familiar with because there are too many chances for errors in the information. However, in this situation it is too bad he hasn't at least made a comment as to whether or not he thinks my analysis of the January 14, 1998 sightings is correct because I used those as a "learning exercise" before tackling the March 13, 1997 array sighting. (Note: from now on J14 refers to 1998 sighting and M13 to the 1997 sighting.)
Since my main interest was to determine the location of the M13 array, one might ask why start with the arrays videotaped on J14? I initially studied the J14 arrays arrays for two specific reasons: (1) several people videotaped geometric arrangements of lights (light arrays) that were so characteristic or unique in "shape" that there is no doubt that they videotaped the same arrays at the same times from widely separated locations thus making the triangulations definitive and (2) the witnesses all agreed that the J14 lights were very similar to, if not identical to the lights seen M13. This agreement meant that analysis of the J14 lights provided a conventient "training tool" for me to get my triangulation "chops up" before I tackled the more difficult M13 lights.
By way of contrast with the M13 videos, when I began my investigation there was some question as to whether or not all the witnesses filmed the same array on M13 and, even if one assumed they all filmed the same array, there was a question as to whether or not they filmed at the same time. Since my initial goal was to produce an accurate triangulation to determine how far the lights were from Phoenix I needed to have a high level of confidence, if not independent evidence, that (a) all the witnesses of the 10 PM event videotaped the same array and that (b) either they all vidotaped at approximately the same time (within a minute or so) or at least that the array was moving so slowly (if at all) that differences of several minutes in videotape times wouldn't matter.
When I began my investigation there was a question as to whether or not the M13 curved arc array videotaped by Mike Krzyston (K) was actually the same array of only 3 lights as videotaped by L (Dr. Lynn, which is a semi-pseudonym), nine lights videotaped by Mike Rairdon (R) and other numbers videotaped by Tom King (TK). This question existed quite independently of the questions surrounding the possible identification of the lights. These other questions included whether or not the lights could be flares dropped close to Phoenix or at high altitude far from Phoenix, whether or not the Air Force was lying to the public, whether or not it was possible to rule out flares based on a "sort of" spectral analysis (done by Jim Dilettoso at Village Labs), whether or not the Cognitech analysis (presented on the Discover channel program in late October, 1997) could be trusted (it suggested that the lights videotaped by K dropped behind the Estrella range as they moved downward and hence could be far from Phoenix), whether or not flares could be ruled out because the lights were silhouetted against the north side of the Estrellas (as presented by a local Fox TV station rebuttal to the Cognitech/Discover analysis), and so on. The local and national media had combined these video sightings with the 8:00 - 9:00 PM moving triangle sightings (PROBABLY UFO EVENT NOT DISCUSSED HERE!) into a circus sideshow with claims and counterclaims flying hither and yon and personality conflicts abounding, strong opinions having been voiced on both sides and, of course, the undercurrent of "UFOs? Junk, but worthwhile as entertainment." Media in, garble (or garbage) out. Anyway, the situation was so confusing that I really didn't know what I would find.
Fortunately for my sanity, I was able to "quickly" (after a couple of months of effort) come to some definite conclusions about the J14 light arrays. There was no doubt as to the timing of the videos because in many cases particular lights within the arrays could be seen to appear and disappear, with time durations between appearances and disappearances being the same on the different videos. For example, the "linear array" of four lights was very distinctive and the individual lights could be positively correlated between videos. I triangulated their locations using daytime videos for comparison, geographical features as reference points for sighting directions, camera calibrations, etc. (See RPLA.) The same was true for the triangle array and for lights "at the far right" (see RPLA). One added advantage to triangulating the J14 lights was that there was a video taken from a location 55 miles west of Mike Rairdon (the easternmost videographer), namely the P video (witness requests anonymity). This 55 mile baseline established conclusively that the J14 lights were some 70 miles from Phoenix and at altitudes between 15,000 and 20,000 ft.
All the J14 lights that I triangulated turned out to be over the northeast corner of the restricted Air Force range as indicated on a map of the area that I was given. Any particular light was seen for between 4 and 5 minutes which is consistent with the lifetime of the extremely bright LUU-2 type illumination flares (2 million candlepower). Hence it makes some sense to assume they were lights related to Air Force activities and, in fact, the Air Force stated that there had been flares dropped at a rather high altitude during the annual Air National Guard "Operation Snowbird."
As you read the next sections keep in mind the claim of the witnesses that the J14 lights looked just as bright and "astounding" as the M13 lights.
MARCH 13, 1997
I analyzed three M13 videos: Krzyston (K), L and Rairdon(R). (I requested a copy of King's video but never got it.) The MK video shows the lights "turning on" because he had seen an earlier light and had gotten his videocamera to tape it (see Figures 14 and 15 of RPLA). That light had gone out by the time the left hand light designated #1 appeared. It was some seconds later that the lights in the arc array, designated in order of appearance as #2 - #9, appeared (see RPLA). (I should point out that the "turn on" of a light was not immediate. The video shows that they brightened up rapidly but not instantaneously, i.e., not within a single frame time of 1/30 sec.) Lights 2 through 9 appeared one after another, each one appearing to the left of the previous one thereby giving the sensation of motion from right to left, in a sequence making an arc pattern. Light #1 was to the left of light 9 (and lower) and was not part of the arc. The K video also shows the lights going off...which is more of a fade out than an abrupt disappearance. However, they did not go out in the order in which they appeared.
L claims to have seen six lights in the sky at about 10 PM. She had seen such lights previously and for other reasons I won't go into here had decided that it was important to record these light events. She ran to get her videocamera and managed to record only three of the lights, the others having gone out (see Figure 16 or RPLA). These three lights form a straight slanted line, as do the lights 7,8, and 9 of the K video (and the same three for the R video discussed below). Hence it is reasonable to assume she videotaped the last three lights to disappear on the K video. However, the fact is that the order of disappearance on the L video does not match the order of disappearance of the last three lights videotaped by K. This fact initially raised the question as to whether or not she videotaped the same lights.
Finally, I studied the R videotape which begins with all nine lights on (see Figure 17; note that lights 7,8,9 are nearly in a straight line). A simple comparison of the R and K videos shows an immediate problem in correlating the lights. Most of the K video shows only lights #1 (to the left of the array),5,6,7,8, and 9. Only at the very beginning of the K array video can one see lights 2 (at the far right), 3 and 4. These are visible before 5,6,7,8 and 9 appear and they go out almost immediately. However, a "synthetic" array created by combining frames from the K video in order to show all the lights 1 through 9 is almost identical to the array in the R video. Perfect "identicality" is not to be expected since the perspective views were different (R is about 32 miles east of K). The effect of the perspective difference is most obvious when the spacings between lights 1 and 9 are compared on the two videos (after taking magnication factors into account): the spacing is smaller in the R video than in the K video (see Figure 18 of RPLA).
It appears to be impossible to get an exact temporal correlation between the lights in the K, L and R videos. One has only to look at the order of disappearance to see the problem. For convenience I number the last three lights to appear in the L video as 9,8,7 , numbering from left to right, to make these three appear to be the same as 9,8,7 in the other videos.
The lights are listed in **ORDER OF DISAPPEARANCE** with DURATIONS in parentheses:
2(first to disappear, duration at least 16 sec), 3(second to disappear, duration at least 72 sec), 4(third to disappear, duration at least 74 sec), 1(at least 108 seconds; TPLA gives 130 sec), 9(79), 6(115), 5(126), 7(112) and 8(121; this was erroneously given as 130 sec duration in RPLA) Note: a camera stop of unknown duration prevents perfect accuracy in determining the durations of the first four lights.
7(19), 8(26), 9(60)
1(39), 2(85), 3(88), 5(95), 4(98), 7(105), 6(117), 8(130), 9(151)
(A comparison of these numbers those in TPLA shows agreement within a couple of seconds for the K video and for the R video these numbers range from 4 to 6 seconds larger than in TPLA. Note that the lights tend to fade rather than extinguish abruptly so it is sometimes difficult to decide when a light is finally "out.")
Looking just at the last three lights to disappear in each video we see the orders of disappearance to be different. In the K video the order is 5, 7, 8; in the L video, 7, 8, 9; in the R video, 6, 8, 9. The K an L videos agree in the disappearance order of 7 and 8 and the L and R videos agree in the order of disappearance of 8 and 9 (assuming L videotaped lights 7,8,and 9 as defined for the other videos!) (Keep in mind that numeric designation of the the lights is based on their order of appearance on the K video.)
This disparity in the order of extinguishing is what prevented me from beginning my analysis with the M13 lights. There was such controversy over what had happened that I didn't feel "safe" tackling the M13 array until I had successfully triangulated other lights. Fortunately the J14 arrays had none of this ambiguity.
Nevertheless, I believed the probability to be 100%, or nearly so, that K, L and R videotaped the same array. Furthermore, I guessed that they all videotaped within a few minutes of one another. (This is a rigorous conclusion rather than just a "guess" if the lights were flares because the each flare burns for less than 5 minutes.) I also studied the motions of a few of the lights by very carefully comparing the positions at the beginning and end of the K and R vidoes (see RPLA, Figures 20 - 25). Although the lights did move downward and to the left, there were were no large changes in position of the lights. (Note: Jim Deardorff, an expert in atmospheric physics, has learned that the general upper altitude wind condition in the late afternoon, the closest measurement time to the time of the sighting, was 305 degrees azimuth at 18 m/sec. This sort of wind from the northwest would have caused flares to drift to the left in the video.) Since the light positions didn't change rapidly it was reasonable to carry out the triangulation to see what I would find. I used some of the calibration data generated during the analysis of the J14 arrays and obtained the chart (map) shown in Figure 18 of RPLA. As this figure shows, under the assumptions set forth (same array, very little change in position over the times between videos) the lights were about 70 miles from Phoenix and over the northeast corner of the restricted area.
Assuming that they all did videotape the same lights and that the lights were some 70 miles away, the next question is, why is there little or no correlation between the light durations or the order of extinguishing? The answer to this is based on the following facts: (a) the lights were dropping downward slowly (see Figures 21-21 of RPLA), (b) the lights were low on the horizon,(c) the witnesses were looking over mountains which provide a jagged horizon that is different for each observer. Hence it is reasonable to propose that the lights dropped down behind mountain peaks, just as MUFON investigator Richard Motzer suggested in his July 1997 MUFON Journal article. This explantion was suggested again by the analysis done by the image processing company, Cognitech, as shown during the Discover Channel documentary on October 26, 1997. Cognitech combined daytime video of the scene in front of Mike Krzyston's house with his nighttime video of the arc array. By properly compensating for hand vibrations and differences in magnification they were able to overlay the nighttime video on the daytime video, frame by frame, and show the lights first above the ridgeline and then descending and disappearing at the jagged mountain ridgeline. Thus, according to the Cognitech analysis, the differences in disappearance times resulted from the differences in altitude of the various locations along the ridgeline. Figures A1 and A2 of this addendum show my relatively crude attempt to do the same thing for a single frame.
It is very difficult to be accurate because the nighttime video can only be scaled to the daytime video by using the nearby sloping hill as a reference and the outline of the hill is indistinct, as illustrated in Figure A1. Figure A2 shows the calculated positions of the lights of the arc superimposed on the daylight video after compensating for magnification differences.
The calculated positions in Figure A2 are actually higher than they should be, but even moving them down a little bit would not place them below the ridgeline. The main point is, however, that they were initially above the ridgeline and that as they dropped down they disappeared behind it. Because of the jagged edge of the ridgline they disappeared ("went out") at different times. (Note: I carried out this sort of analysis before RPLA was published. However, I did not include it because I was able to refer to the much better Cognitech analysis.) According to an article by Tony Ortega in the Phoenix New Times, March 5, 1998, Dr. Paul Scowen repeated the Cognitech type of analysis for a single frame and obtained a similar result with the flares just above the ridgeline.
Station KSAZ, Channel 10, a Fox TV station, did a similar analysis by attempting an overlay of a daytime video frame with a nighttime frame. Their analysis shows the lights below the ridgeline of the mountains. I suspect that this was because in overlaying the nighttme video on the daytime video they did not properly match the magnification factors of the daytime and nighttime video frames. (Perhaps the nighttime video was taken at a lower magnification factor - shorter zoom focal length - than the daytime comparison.)
TPLA begins with a very brief summary of the results in RPLA and points out the problem with temporal correlation. It then discusses the problems with the Air Force statements. According to TPLA, the Maryland National Guard claimed that the training session with flare drops was held over the North Tac Range, not over the East Tac Range where the triangulation placed them. Initial Air Force statements were that there were no Air Force planes flying after 8:30 PM. However, about 4 months after the sighting the Maryland National Guard stated that it had use of the range between 9:30 and 10 PM. There is no doubt that there was confusion on the part of the "authorities" in this regard.
In late July, 1997, Captain Eileen Benz, spokeswoman for the Arizona National Guard said she had learned from National Guard helicopter pilots that they had seen a group of A-10's, the aircraft which drop the LUU-2 flares, heading for Davis-Monthan AFB at about 10 PM on March 13, 1997. She then learned that the Maryland Air National Guard had used the Barry Goldwater range. According to Beinz, the A-10's dropped flares at an altitude of 15,000 ft at 10 PM over the "North Tac Range" which she placed at 30 miles southwest of Phoenix. (I don't know where that would be. I suspect she had the distance wrong, or else the flares were not ejected over the North Tac Range.) Capt. Drew Sullins, spokesman for the Maryland Air National Guard, also in late July 1997, stated that a squad of A-10 jets had been using the Barry Goldwater range for training missions some 60 miles southwest of Phoenix and that the planes had "dumped several flares" at high altitude. Thus the statements by the National Guard officials indicate that there were flares ejected in the general area of the arc of lights.
During operational exercises flares are dropped at altitudes below 8,000 ft. At these lower altitudes they would be invisible to people in Phoenix because of the Estrella Range (and South Mountain). Probably most of the flares were dropped under normal operating conditions (low altitude) over the North Tac Range far west of Tucson, as stated by the MNG. However, according to a newspaper story about the National Guard training with flares, the aircraft are not allowed land with unused flares but rather must throw them out. Unburned flares falling from high altitude could be dangerous (!!!) but burned flares much less dangerous since they are designed to burn up during the fall (even the aluminum casing is burned). Therefore what I suspect happened is this: two planes on the way back to Tucson (Davis-Monthan AFB) were flying generally eastward at altitudes around 15,000 ft when they ejected unused flares. I suspect that one plane ejected a single flare that became light #1. I suspect a second plane flying to the AFB (right to left from the point of view of witnesses in Phoenix) then ejected 8 unused flares, the maximum number carried by an A-10. I conjecture that plane was making a gradual turn to the right while climbing somewhat while ejecting the flares, thus making the arc of 8 lights. (I saw something similar to this while in Gulf Breeze in 1992. To the naked eye it appeared as a series of lights, one after another, appearing in a row with each one going out shortly after it appeared. A high power telescope proved there was a large airplane ejecting flares. With the telescope I could see the airplane, only about 20 miles away, lit by the light of the flares because the flares ignited *close* to the airplane!)
H&K point out that the colors of the lights seem to be too orange to be magnesium flares of the LUU-2 type, since magnesium burns at high temperature with a white light. I suggested that light traveling through the atmosphere over 50 miles could be reddened (like the moon or sun) by dust and moisture droplets in the atmosphere. They, however, argue that any reddening wouldn't be sufficient to produce the orange color they and the others saw. This can only be resolved by experiment. (Note: although the magnesium burns white, the consumption of the aluminum cylinder that contains the flare "candle" may add some orange to the light.) The problem of color may be resolved with controlled observations during similar training exercises expected to take place during early 1999.
Flares burn at uneven rates and therefore fluctuate in brightness. The lights on the video also fluctuate in brightness. The lights in the videos also show flare-like characteristics in the way they appeared and disappeared. The appearances were reasonably fast but the disappearances were more gradual, sometimes taking seconds to complete disappear (burn out).
H&K point out that I did not discuss any of "orb" sightings from other parts of Arizona (or from other parts of the world). Of course, I did not think it was necessary since I was attempting to learn about the particular lights in the videos. There have been orange orb sightings throughout the world. Many or most of these remain unexplained. It is my impression from having read the literature that most of these are relatively close to the observer, not tens of miles away.
H&K discuss the their own sighting from a location in Awahtukee. They assume that they were looking toward Montezuma's peak, which they estimated at 2,500 ft above them and 8 miles away. They calculated that if the lights were above the peak and at a distance of 67 miles from them the minimum altitude of the lights would have to be 2,500 x 67/8 = 21,000 ft, which is much higher than I had calculated. However, their calculation has to be revised for several reasons. Bill Hamilton supplied me with daytime panarama photos showing the *southern end* of the Montezuma range (Figures A3, a composite of several photos, and A4).
Notice in particular the entry way lights and ground lights.
He also provided a couple of video frames showing that there were ground lights visible to use a reference points and one the location of one of the "orbs." (Figure A5).
By combining the daytime and nighttime images one could show that they were looking over Montezuma's Head (southern end of the Estrella Range, azimuth 221-223 degrees) rather than Montezuma's Peak (azimuth 235 degrees). The distance from them to the Head was about 14 miles, their altitude was about 1,200 ft and that of the Head is about 3,400 ft (according to the geological survey map) so the altitude difference was about 2,200 ft. The distance from them to the array, as I had located it by triangulation using the K and R videos, was about 60 miles. Hence the minimum altitude above sea level of the lights (if 60 miles away) was about 1200+2,200 x 60/14 = 10,600 ft, which is within the altitude range expected for the flares. (The minimum angular elevation looking over Montezuma's Head was about arctan[2200/(5280 x 14)] = 1.7 degrees. That is, from their position no light lower than 1.7 deg elevation - approximately - could be seen if on the far side of the mountain.)
This altitude (10,600 ft) can be compared with an estimate of the altitude of light 9 as determined by the K video. The calculated height is based on calibration of the camera and the known altitudes of Krzyston's house and the altitude and distance of the 4512 ft high Montezuma's Peak in the center of a daytime video taken from Krzyston's house (see Figure A2; it is the highest peak in the distant mountain chain). Calculations show that the flat top of the nearby hill that appears in clearly in his comparison daytime video and in silhouette against the glow from the city lights in his nighttime video, is about 0.4 degrees in angular elevation above Krzyston's altitude, which is about 1600 ft above sea level. The angle is difficult to measure accurately but it appears that light 9 is about 0.9 degrees higher than the flat top of the hill, i.e., at about 1.3 degrees elevation above K. Hence, projecting this angle to the distance of 77 miles (see Figure 18 of RPLA) gives an altitude of light 9 of 1600 + 77 x 5280 x tan 1.3 = 10,800 ft, which is very close to the previously calculated 10,600 ft. Hence, according to these calculations K's light 9 would have been high enough for H&K to see/videotape it.
[NOTE: The close agreement between 10,800 ft and 10,600 ft, the minimum altitude for visibility from the H&K location, is surprising (and gratifying) but must be considered fortuitous. These calculations are derived from estimates of altitudes that are based on the interpolation between contour lines on a survey map and can easily be off by several tens of feet to more than a hundred feet. For example, if K's altitude were 1700 ft instead of 1600 the altitude calculated for light 9 would be closer to 11,000 ft. Also, at a distance of 77 miles an error of only 0.1 of a degree in the elevation angle corresponds to a height variation of about 600 feet. The angle estimate could easily be low by 0.1 degrees, so the height might be as much as 11,200 ft or more. The main point is of this calculation is that it establishes an altitude consistency between what H&K could have seen if the lights were 60 miles away from them and the lights videotaped by K and R.)
H & K have claimed that the lights videotaped by Krzyston were below the mountain skyline. They base this conclusion on the video analysis presented by the Phoenix station KSAZ in which the technician superimposed a daytime with the nighttime video. As I have already pointed out, this rather quick video analysis is contradicted the much more careful analysis done by the Cognitech company for the Discover TV show, by an independent analysis of Dr. Paul Scowen and by my own analysis as illustrated in Figure A2.
H & K claimed that their sighting direction was toward Montezuma's Peak at about 235 degrees azimuth. However, any lights that were in that direction and also in an area that could appear in Krzyston's video would be on the north side of the Estrellas and at an azimuth of around 265 degrees, not 235 degrees, from the location of H & K. But of more importance is the fact, mentioned above, that the daytime comparison photos show they were looking toward Montezuma's Head, a nearly flat topped portion of the Estrella range at an azimuth of about 221-223 degrees. Projection of a line along the 221 degree azimuth starting at the location of H&K passes through the location of the arc of lights as determined by the K, L and R videos (see Figure 18 of RPLA). Hence it would appear that, because they were looking in the direction of Montezuma's Head, they were also looking in the direction of the arc array.
Hamilton and King have analyzed the turn on and turn off times and durations of the lights in the various videos. They conclude, referring to the relative times of the lights, "This is the most telling calculation. Between the K and R videos the 'power down' of the lights is in a different sequence than the 'power up' and they differ from each other in sequence and duration implying that the videos did not shoot the array in the same time frame. This would nullify the triangulation results."
The last sentence is wrong unless (a) there was considerable movement of the lights between the times of the various videos or (b) the identifications of the various lights in the several videos are wrong, which is to say light 1 of the K video was not light 1 of the R video and light 8 of the K video was not light 8 of the R video, etc. Possibility (a) is ruled out by the measurements of the locations of several lights when they appear and a couple of minutes later before they disappear: there was slight motion to the left and downward (downward motion is not important for the triangulation) so the triangulation cannot be ruled out based on motion. Possibility (b) would certainly obviate the triangulation, but then it would also obviate Hamilton and King's comparative time analysis of the lights of the array in one video as compared to another. Of course there is another possibility, namely that there were several arrays that night and the various videographers shot different ones. However, all these videos were taken at roughly the same time, about 10 PM, (presumably within minutes of one another) and all in generally the same direction (southwest). Therefore it seems to me unlikely that each videographer saw a different array, as opposed to all seeing the same array but at different perspectives and under different circumstances.
I conclude from the time analysis and the evidence that the lights dropped downward that the simplest explanation for the different durations is that offered initially by Richard Motzer, namely that the lights dropped down behind the jagged ridgline of the mountains, thereby creating a "random turn off" effect that was uncorrelated between the various videos.
The preceding analysis shows that (1) H&K were looking toward the array location determined by the K, L and R videos, (2) the angular elevation of the highest lights of the array was higher than the angular elevation of the portion of the mountain range they were looking over (Montezuma's Head), (3) the seemingly random "turning off:" of the lights can be accounted for by assuming they disappeared behind jagged ridgelines. Hence I conclude that the lights actually were at a great distance and very probably flares ejected at the end of a training exercise by the Maryland National Guard. This conclusion DOES NOT PRECLUDE an experiment in which identical types of flares are viewed and videotaped from a great distance as they fall.