ANALYSIS OF THE PHOTOS OF AN UNIDENTIFIED OBJECT OBSERVED BY THE ASTRONAUTS OF SKYLAB 3
by
Bruce Maccabee and Brad Sparks
SKYLAB 3
FIGURE 1 THE FOURTH PHOTO
The sighting by Alan Bean, Owen Garriott and Jack Lousma, of a red "satellite" occurred on
September 20, 1973 (day 263, revolution 1863 of their spaceflight) at approximately 1635 to 1645 GMT.
(Brad Sparks was alerted to this sighting by a brief mention of it in an Associated Press report.
It was also mentioned in Robert Emenneger's 1974 book "UFOs, Past Present and Future.")
OWEN GARRIOTT JACK LOUSMA
SKYLAB 3 ASTRONAUTS AT A PRESS CONFERENCE
.........................................................................
The sighting is reported in the following document compiled by James Oberg in 1977:
The second manned Skylab visit was within five days of returning to earth after a record-breaking
59 day expedition. The crew had awakened at 0700 GMT and would go to sleep at 2300 GMT. They had
eaten lunch in the wardroom and were doing a 'procedures review' in the wardroom prior to beginning a
new series of experiments in the afternoon. During the sighting the crew was out of contact with the
ground control. The on-board tape recorder was not turned on and the first recorded mention of the
incident was on a ground link some 4.5 hours after the event. Garriott had taken four photographs.
During the taped discussion the conversation went as follows:
LOUSMA: "Did you tell him about that satellite we saw?
BEAN: Yes, we saw a great satellite. We didn't know if we told you about it.
LOUSMA: The closest and brightest one we've seen.
BEAN: Huge one.
LOUSMA: We've seen several. It was a red one.
CAPCOM: No, you may have told somebody, but it wasn't this team. I don't remember hearing about it.
LOUSMA: I guess we didn't report it. It was reflecting in red light and oscillating at, oh, counting
it's period of brightest to dimmest, about ten seconds. It led us into sunset. That was about three
revs ago, I think. Something like that, wasn't it Owen?
(no answer by Owen; topics changed.)
Note: "Three revolutions ago," at about 1.5 hours per revolution is 4.5 hours before. Garriott said later
that the object did not lead the Skylab into sunset, but rather followed the Skylab into sunset. (see
below) Below is the original Oberg document.
Note that the above document indicates that the location was over the southwestern Indian
Ocean during revolution # 1863 (see map below).
...........................................................................
Further information is available from debriefing that took place on or about October 4.
Transcriptions are given below the documents.
SKYLAB 3 DEBRIEFING Oct 4, 1973
"GARRIOTT: Do you want to talk about that satellite?
LOUSMA: I saw a couple of satellites that appeared like a satellite would on earth. I saw one that
was not like one you would see on earth, so why don't you mention it?
GARRIOTT: OK. About a week or 10 days before recovery and we were still waiting for information to be
supplied to us about the identification. Jack first notices this rather large red star out the
wardroom window. Upon close examination, it was much brighter than Jupiter or any of the other
planets. It had a reddish hue to it, even though it was well above the horizon. The light from the
Sun was not passing close to the Earth's limb at the time. We observed it for about 10 minutes prior
to sunset. It was slowly rotating because it had a variation in brightness with a 10-seconds period.
As I was saying, we observed it for about 10 minutes, until we went into darkness, and it also
followed us into darkness about 5-seconds later. From the 5 to 10 second delay in it's disappearance we
surmised that it was not more than 30 to 50 nautical miles [35 to 58 statute miles or 56 to 93 km]
from our location. From its original position in the wardroom window, it did not move more than
10 or 20 degrees over the 10 minutes or so that we watched it. Its orbit was very close to that of
our own. We never saw it on any earlier or succeeding orbits and we'd be quite interested in having
its identification established. It's all debriefed in terms of time on channel A, so the precise
timing and location can be picked up from there."
NOTE: according to Oberg the Channel A tape recorder was not on, so the exact time cannot be determined.
NOTE: Soon after this debriefing Garriott told Sparks his best
estimate of the time interval between the Skylab going into sunset and the red
light's disppearance was about 5-6 seconds. Garriott explained exactly how he
counted out the seconds "one thousand one, one thousand two, etc."
One of us (Sparks) asked an expert in satellite observation to take the NORAD tracking data and use
his satellite observation program SkyMapPro, to determine Skylab 3's latitude-longitude-time
coordinates and shadow entry, which turned out to be between 1645 and 1646 GMT (Greenwich Mean Time)
or UTC (Universal Time Coordinated)(it was sunlit at 1645 and in the shadow at 1646). This fits the
above reported 1645 GMT time and places 10-minute duration sighting at about 1635-1645. At 1645-1646
GMT Skylab was over Madagascar Channel at the extreme western edge of the Indian Ocean.
Ground track of Skylab Orbital Workshop
Time Time Latitude Longitude Altitude Speed Lighting
hr:min km(nm) km/sec(nm/sec)
---- ------ --------- -------- ----- ----- -------
20 Sep 1973 16:40 04°46'15"S 026°06'28"E 434.6 (234.7) 7.651(4.13) SUN
20 Sep 1973 16:41 07°44'11"S 028°21'55"E 435.7 7.650 "
20 Sep 1973 16:42 10°41'08"S 030°39'27"E 436.8 7.649 "
20 Sep 1973 16:43 13°36'44"S 032°59'52"E 438.1 7.648 "
20 Sep 1973 16:44 16°30'35"S 035°24'06"E 439.3 7.647 "
20 Sep 1973 16:45 19°22'15"S 037°53'03"E 440.6 7.646 "
20 Sep 1973 16:46 22°11'17"S 040°27'43"E 442.0 7.645 SHADOW
20 Sep 1973 16:47 24°57'09"S 043°09'10"E 443.4 7.643 "
20 Sep 1973 16:48 27°39'17"S 045°58'31"E 444.7 7.642 "
20 Sep 1973 16:49 30°17'01"S 048°56'57"E 446.1 7.641 "
20 Sep 1973 16:50 32°49'38"S 052°05'44"E 447.4 (241.6) 7.640(4.12) "
(270-278 statute (4.75 statute
miles/sec) miles/second)
Latitute-Longitude values are to nearest 1 arcminute and altitudes to nearest kilometer.
A map has been constructed from the above data, showing the shadow boundary on the earth and the track
of the Skylab (white line):
THE PICTURES
The first and last photos presented here are pictures of the earth taken before
and after the sighting. These photos are mentioned in the NASA publication, "Photographic Index and Scene
Identification," NASA-TM-X-69780, that was tabulated by Richard Underwood and co-workers in November, 1973.
(See click here). In that document the photos
are listed twice. Page 41 lists the photos that were contained in film magazine #CX-35. These photos were
taken with a 35 mm Nikon camera with a lens of either 55 or 300 mm focal length using a film called
SO-368 which is comparable to a type of Ektachrome (MS2448). According to the tabulation the photos
labelled SL3-118-2132 through 2135 (frames 1- 4) were associated with the "S-230 experiment." Then photos
SL3-118-2136 and 2137 (frames 5,6)were taken during the Goddard laser beacon experiment. These photos are
described as "underexposed" and "blurred." According to the report on the laser beacon experiment
("Evaluation of Skylab Earth Laser Beacon Imagery," March, 1975;
see Click Here), these beacon photos were taken on September
4 with a hand-held Nikon camera set at f/4.5 to f8 and with a shutter time of 1/500 to 1/250 of a second.
It is important to know this because these may have been the settings and focal length when the four
photos of interest here were taken, photos SL3 - 118 - 2138, 2139,2140 and 2141. The "Photographic Index...."
lists these four as "Satellite, unmanned" on page 41 and as "blank" on page 245. The two different
descriptions may have resulted as follows (speculation): when the photo index team first went through
the photos they did not know that Garriott had photographed a "red satellite," so, looking quickly
at the photos they saw nothing obvious and listed them as blank; later they learned that he photographed
(what he thought was) a satellite and they changed the description, placing the new description in the
front section of the report but leaving the original description at the "far end" (appendix) of the report.
The next photo, SL3 - 118 - 2142 is identified as "Lake Erie, Ohio, Ontario, clouds."
Photos 2137,38,39,40,41 and 42 are shown below.
HERE IS THE SECOND PHOTO OF THE LASER BEACON EXPERIMENT. Although Garriott, who took the picture,
said he could see the beacon it is not evident in this photo (nor is it evident in the preceding photo).
HERE IS THE FIRST PHOTO OF THE OBJECT.
The first photo of the object shows a featureless red "dot".
HERE IS THE SECOND PHOTO OF THE OBJECT.
The second photo shows a slightly larger image, slightly off round and yellowish in the center. The
yellowish color at the center is typical of an overexposed color photo of a distant, bright
red object or light.
HERE IS A BLOWUP OF THE ABOVE IMAGE
THE RED "DOT" IS BARELY VISIBLE IN THE FOLLOWING PHOTO
SO I HAVE DRAWN A CIRCLE AROUND IT.
PHOTO 2141 CONTAINS THE MOST INTERESTING IMAGE BECAUSE IT ALLOWS FOR AN ESTIMATE OF SIZE
AS REPRESENTED BY THE SPACINGS BETWEEN THE RED "BLOB" IMAGES
The photo shown at the beginning of this article is a blowup of the fourth photo to show the odd shape
of the red glowing object (or objects?).
HERE IS ONE MORE OF THE EARTH
AND HERE IS ANOTHER OF THE SKYLAB
..................................................................................
ANALYSIS
The available information does not allow for a specific identification of the red object but
it does allow us to decide whether or not this could have been a man-made earth satellite by making
an estimate of the size. The size estimate is based on the distance estimate and the maximum
spacing between red "dot" or "blob" images in the fourth photo. The estimate is not based on the
size of the image of a single "blob" or "dot" because the "dot" or "blob" image made by a distant
bright light is generally larger than the image would be if the optical system could provide a
perfect "geometrically accurate" image. The extra image size is a result of "diffraction" and "lens
aberration" which combine to make light spread sideways on the film plane.
(NOTE: The image sizes of extremely distant lights that make only "dot" images, like images of stars,
depend only upon the brightness, with the brighter lights making bigger images.
Astronomers have made use of this phenomenon to determine relative star brightnesses from
photographs of stars.) Thus, in the case of a "dot" or "blob" image the most one can say is that
the reflective object or light source which made the image was no larger than the geometric size,
as described below, that corresponds to the size of the dot or blob image.
For objects which are close enough or large enough to be "resolved" by the optical system
(i.e., large enough to have obvious structure such as distinctly separated image dots or shapes)
there is a way to estimate the object size from the image size. For a resolved image the size is
geometrically related to the actual object size by a ratio equation that can be stated as follows:
the image size(width, length, spacing of "blobs" as measured in the focal plane), herein called I,
divided by the focal length, F, is equal to the size (width or length, as measured in a plane that is
perpendicular to the line of sight) of the object, O, divided by the distance to the object, D.
In equation form this is: I/F = O/D, which is based on the law of similar triangles with a common
vertex at the center of the lens. (Usually the resolution capability of the camera is determined
by whether or not this relation holds for a particular image of a particular object at a particular
distance.) What can and what can't be resolved depends upon the optical quality of the camera lens
and other factors related to the camera and film.
If an object or light makes a "dot" image on the film, then the calculated size of the object, O = D(I/F),
is (probably) much larger than the actual size of the object, as described in the previous paragraph.
However, when there are distinguishable features of the image that are clearly separated by distance on
the film it is possible to estimate the spacing between the features which made the images.
In case of the first three Skylab 3 photos it would be of little value to try to estimate the size
of the red object by measuring the size of individual "dot" images and multiplying by D/F. But in the
fourth photo there is a considerable spacing between images of "dots". This means that the distance
between dots has been resolved by the camera even though the dots themselves have not been resolved.
[There "dots" are somewhat elongated so, strictly speaking, they are not exactly "dots."]
On a direct copy (1:1) of the original film the maximum spacing between "dots," specifically the
spacings betwee the centers of the upper and lower dots is about 0.87 mm. The focal length
of the camera was (probably) 300 mm (it might have been 55 mm; see below). Thus the ratio I/F
equals about 0.87/300 = 0.0029 (in radian measure) (or, if the focal length was 55 mm, the
ratio is .87/55 = 0.0158). As pointed out above, this ratio is proportional to the object size, as
measured perpendicular the line of sight (i.e., as "projected" onto a plane surface that is perpendicular
to the line of sight), divided by the distance.
Now it is necessary to know the estimated distance D in order to calculate
the object size, O, which is the "projected" spacing between whatever made the upper and lower dots.
Garriott, who was the first scientist at the Skylab, had the presence of mind to count the seconds
between when the Skylab went into the earth's shadow and when the object, which appeared to be following
the Skylab, disappeared. He assumed the object was visible only by reflected light, an assumption
which would be reasonable if the object had been another orbiting satellite. (If the object had been a
bright source of red light [there were no satellites with bright red lights attached] then his
assumption would not necessarily apply.) He also assumed that the red object was behind the Skylab and
traveling at the same speed which, according to the table above, was about 7.64 km/sec.
Based on these assumptions, and his estimated time until it disappeared, 5 - 10 seconds, the calculated
distance between the Skylab and the object was (5 to 10 seconds) times 7.64 km/sec = 38 to 76 km.
Assuming that the fourth photo was taken just before Skylab went into the shadow (an "unfounded"
assumption since we don't know at what times, during the sighting, any of the photos were taken) and
using the shorter distance estimate along with the ratio I/F = 0.0029 we find O = D(I/F) = D(0.0029) =
38 km x 0.0029 = 0.11 km = 110 m or about 336 ft as projected onto a plane perpendicular to the line of sight.
If D were 76 km the calculated size would be twice as large. (Sparks spoke to Garriott in October,
1973, soon after the mission had been completed. Sparks asked how Garriott estimated the time
until the "satellite" went into the shadow. Garriott said he counted off the seconds as "one thousand one,
one thousand two", etc. and estimated 5 to 6 seconds this way. Hence these authors believe that
the lower time estimate is probably more likely to be correct.) The width of the image, measured from the
middle "dot" to the end of the "extension" at the right side of the middle dot, is about 1/3 of the spacing
between the upper and lower dots. Thus the width would be about 37 m or 112 ft if 38 km away and 74 m
or 224 ft if 76 km away.
Now let's assume that the various dots that form the fourth image are reflections from points or small
reflective areas of a "typical" earth satellite (manmade!). Then this calculation indicates that the
spacing between two reflective portions of the satellite was 336 or more feet. However, in 1973
there was no satellite remotely approaching that size (and only the space station is comparable now,
over 30 years later).
(NOTE: The Skylab-3 itself with docked Apollo module was the largest satellite
object in orbit at the time, about 150 feet long.)
The calculation just done is a lower bound on the spacing between the outermost lights for the assumed
distance. The calculation assumes that an imaginary straight line connecting the upper and lower red lights
(or red reflective areas) on the object was perpendicular to the line of sight from the camera to the object.
If a line drawn between the red lights was not perpendicular to the line of sight, then the spacing
between them was greater than estimated above (by the factor 1/sinA where A is the acute angle between
the sighting line and the line between points on the object; there is no way of knowing what A was so 90
degrees is assumed here, resulting an an estimate of the minimum spacing between "lights").
(NOTE: when this discussion was first published there was a question as to whether the focal length
was 300 or 55 mm. One of us (Maccabee) had made a notation (in 1977 when copies of these photos were first
obtained) that the focal length was 300 mm, but we recently learned that there were 5 Nikon cameras and each
had either a 55 or 300 mm lens. This raised the question of whether or not the focal length could have
been 55 mm. The question has now been resolved by the combination of the photo index and the report on the
beacon experiment, as described above. Although it was 16 days between the laser beacon experiment and the
date of these four photos, we believe it highly unlikely that they would have changed the lens on this camera
if they wanted to take wide angle (55 mm) photos since they had four other Nikon cameras available,
several of which had 55 mm lenses. We assume that Garriott simply grabbed a Nikon with the 300 mm lens,
namely, the one used many days before during the beacon experiment.)
WHAT COULD IT HAVE BEEN?
The possible identifications divide into two general classes: (a) conventional explanations that assume a
manmade/earth-launched satellite and (b) unconventional explanations that assume it was not an earth-
launched satellite. Explanations of type (b) allow for the object to do "anything" because the object is
assumed not to be limited by earth's gravity. In particular, (b)-type explanations allow for the object to
make non-orbital motions, such as travel at velocities greater or less than the orbital velocity that is
appropriate for its altitude. (For an object that is "trapped" by earth's gravity the orbital velocity depends
upon the radial distance from the center of the earth). For example, one might imagine that the reason the
object appeared as a dot in three photos and then as a structured object in the fourth is that the object
was at a large distance from the Skylab during the (unknown) time duration of the first three photos and then,
between the third and fourth photos, it suddenly moved much closer to the Skylab, at which time
its structure could be resolved by the camera. If this happened, then the object made motions that are not
consistent with being a satellite subject to orbital motion around the earth. There are many other non-orbital
motions which one could imagine if the object was not a satellite and one could spend a long time trying to
study all of these. However, the main point is that if the object was not confined to an orbit, since there
were no man-made objects capable of such motions in the vicinity of the Skylab, the object would have to be
considered truly anomalous, an Unidentified Space Object (USO, not to be confused with "USO: Unidentified
Submerged Object").
Explanations of type (a) are the only ones that can be tested against known physics of satellites so the
question arises, does the conventional hypothesis, that it was a satellite, fit the data?
This hypothesis runs into trouble immediately when compared with the photographic and visual data. Both visually
and photographically the object appeared bright red in color. Noting this color but assuming it was a
satellite or some object that could have been reflecting sunlight, Garriott pointed out that when it was first
seen and for about 10 minutes afterward the object and Skylab were not in a region of the orbit
where the sunlight would have passed through the earth's atmosphere and therefore would
have been reddened by scattering from particulate matter in the atmosphere (as in a red
sunset). Hence the red color could not have been a result of the reflection of reddened sunlight from a
typical satellite. To explain the red color Garriott may have assumed (although he never mentioned this
assumption) that the satellite was painted red. He awaited an identification by NASA, but there was none
because there were no satellites with red surfaces. The surfaces are typically painted black or white
or are unpainted bare metal.
An alternative hypothesis is that it was a satellite emitting intense red light. Garriott probably didn't consider
this to be a possible explanation since there were no satellites with bright red lights.
The assumption that the astronauts were watching the red lights on an object and not reddened solar reflections
leads to the conclusion that Garriott's method of estimating the distance could have been wrong. If the object
had red lights then the fact that these lights disappeared shortly after Skylab went into the earth's shadow would
not necessarily mean that it had also traveled into the earth's shadow. It could have been at any distance
and simply turned off its lights 5 - 10 sec after the Skylab "disappeared" into darkness. But if this were true
then Garriott's estimate of distance as described above could be a happy coincidence or it could have been a
mistake.
One interesting consequence of the red light hypothesis, which includes the assumption that the lights were turned
off 5 - 10 seconds after Skylab went into darkness, is that the object could have been a lot closer to Skylab
than Garriott estimated. Assume, for example, that the spacing between the lights on the assumed "satellite" was
10 m (33 ft), the maximum size of man-made satellites (including booster rockets) in 1973. Then the formula
used above can be rewritten to give the distance that an object would have to be away from the camera in order
to produce an image of the measured size. With O = 10 m, D = O(F/I) = O(1/0.0029) = 10/0.0029 = 3.4 km (2.1 mi).
If the spacing of lights were less than 10 m then the distance would also be proportionally smaller.
Another hypothesis has been considered in order to explain the difference in image size between the first
three photos and the fourth. This hypothesis is based on the assumption that the failure of the camera and
film to resolve the structure during the first three photos is a result of a great distance to the supposed
satellite. It is further assumed that the supposed satellite was in an elliptical orbit and was approaching the
Skylab from the rear. The astronauts estimated that the object was in sight for about 10 minutes. The large
difference in image sizes can be explained if one assumes that the first three photos were taken during the first
minute while the satellite was still very far away and that the fourth was taken roughly 9 - 10
minutes later when the assumed satellite was much closer.
Further to this hypothesis, one can assume that during the first minute, when the object was varying in luminosity
with about a 10-second cycle period, the brightness (or size) of the image might have varied depending on when the
photo was taken during the 10-second cycle (during a low brightness phase or high or in between), and this
variation could have happened without much actual distance change. For example, if the object had been about 10 m
in size and approaching the Skylab from behind at a speed 0.1 km/sec (360 kph) greater than that of the Skylab and
if the first 3 photos had been taken within a ten second period near the end of the first minute of observation,
then, during this initial ten second period, the distance to the Skylab would have decreased by only 1 km. Suppose
the separation had been 63 km when the object was first seen. Then, about 10 minutes = 600 seconds after it was
first seen, i.e., at the assumed time of the fourth photo, the separation would have been only about
63 km - (600 sec)(.1 km/sec) = 3 km and the angular size of the object, (O/D), would have been twenty times
greater than at 60 km and more of the structure would be resolved by the camera.
(Note: this hypothesis differs from the similar one
described previously in the following way: whereas previously it was assumed that the object maintained a
roughly fixed, large distance during the time of the first three photos and then it suddenly made an
"un-satellite-like" rapid motion toward the Skylab, in this hypothesis the object/satellite is assumed to
have approached in a steady manner consistent with an elliptical orbit as described below.)
The preceding hypothesis can be made somewhat more specific. Comparing the size of the image in the
fourth photo with the sizes of the "dot" images in the preceding photos
it would appear that the object would have had to be 20 or more times farther from the Skylab during the first
three photos than at the time of the last photo in order to be far enough away to prevent the camera from
resolving the image. Calculations presented above show that a man-made satellite (ca. 10 m in length) would
have to be about 3 km away in order to make an image the size of the one in the fourth photo,
assuming the 300 mm focal length lens was used. Twenty or more times farther away would place the initial
distance at 60 km or more. The above hypothesis is that the object crossed this 60 km distance in about 10
minutes or 600 seconds meaning that it would have been traveling faster than the Skylab, specifically,
60 km /600 sec = 0.100 km/s (about 200 mph) faster The Skylab was traveling at about 7,640 m/s, so the object,
under these assumptions, would have traveled at about 7,740 m/s. The Skylab was in a nearly perfectly circular
orbit (eccentricity esentially zero). The object is assumed to have been traveling in the same orbital plane
and at the altitude of the Skylab when it disappeared, about 442 km. It was traveling faster, under these
assumptions, and so its orbital eccentricity was larger than zero. The equation for eccentricity,
e, based on the radial distance from the center of the earth at perigee, Rp, and the velocity at
perigee, Vp, is e = (RpVp^2/GM) - 1 where GM = 4 x 10^14 m^3/sec^2. Using Rp = A + Re = (altitude +
radius of earth) = 442 km + 6378 km = 6820 km and Vp = 7,740 m/s yields e = .0214.
The semimajor axis of this ellipse is a = Rp/(1-e) = 6,969 km and the apogee height that corresponds to
these values of Rp and e is Ra - Re = 2a - Rp - Re = 7,118 km - 6,378 km = 740 km. The period of the object in
the elliptical orbit would have been P = 2 (pi)(a^1.5)/(GM)^.5 = 5,779 s or 96.3 minutes which is only about 2
minutes longer than the Skylab orbit time (93.4 min). The result of this calculation is that the object
would have been very close to the Skylab for many minutes. The results would be similar for an object
in a slightly different orbital plane, a plane close to the 50 degree inclination of the Skylab orbit
in order to be within viewing range for up to 10 minutes. Few satellites (if any) other than
the Skylab have been launched into 50 degree orbits.
This set of assumptions does have a problem with the timing. Assume, as above, that the fourth photo was
taken just before the Skylab went into the shadow. Then the object/satellite was about 3 km away from the
Skylab at that time. It was approaching the shadow at a speed of about 7.7 km/sec. If it had been directly
behind the Skylab it would have crossed the shadow boundary about (3 km/7.7 km/sec =) 0.39 sec after the Skylab,
clearly much less time than estimated by Garriott. Suppose, instead, that the satellite had been 3 km higher
and "above" (rather than directly behind) the Skylab when the Skylab went into the shadow. Then the distance
(along its orbit) of the object from the shadow boundary would have been about 3 km/tan 21 = 7.8 km at the time
the Skylab went into the shadow. (In this equation 21 degrees is the vertical angle-the angle measured in
the vertical plane-that the orbit makes with the shadow boundary at the point where the orbit crosses the
boundary.) The satellite would travel this distance in about 1 second, again much less than the time difference
estimated by Garriott.
The preceding calculations illustrate the problem with assuming that the object was the size of a large
man-made satellite and therefore about 3 km away when the fourth photo which was taken. The problem is that,
regardless of the direction assumed from the Skylab to the supposed satellite, the satellite would pass
into the shadow no longer than about 1 second after the Skylab. If the assumed satellite were smaller than
10 m in size, say a meter or so, such as something ejected from the Skylab then it would have been even
closer than 3 km and the time lag would have been even less.
Assume, again, that the satellite or object was directly behind the Skylab. For it to pass into the shadow
5 seconds after the Skylab, the distance of the satellite from the shadow boundary, as measured along its
orbit, would have had to have been at least (5 sec x 7.7 km/sec =) 38.5 km at the time when the Skylab
crossed the shadow boundary. Thus at the time of the fourth photo (assumed to have been taken just before
the Skylab went into the shadow) the distance would have been 38 km or greater.
If the supposed satellite had been at a higher altitude and "above" the Skylab when Skylab crossed the shadow
boundary it could have been closer to the Skylab than 38 km and still take 5 seconds to reach the shadow boundary.
For this to occur the distance above the Skylab would have been (38.5 tan 21 =) 14.8 km which means that
at the time of the fourth photo the distance would have been around 15 km. Using this distance to calculate
size yields (0.0029 x 15,000 m =) 43 m, which is still four times the size of the largest satellite
(other than the Skylab itself). Thus we see that it is "difficult" (impossible?) for the satellite (or
ejected debris) hypothesis to satisfy two sighting requirements at the same time:
1) the size of the object, as determined by the distance and angular size (from the fourth photo)
must be comparable to or smaller than that of a large satellite and
2) the time difference between when the Skylab crossed the shadow boundary and when the object/satellite
crossed the boundary must be 5 seconds or more.
Unfortunately there is no information on the actual direction from the Skylab to the object. Garriott
said the direction to the object moved no more than 10 - 20 degrees, but he didn't provide the actual
sighting direction. His assumption that the object followed Skylab into darkness suggests that
the object was "behind" the Skylab, but in what direction behind we do not know. We can, therefore,
drop the tacit assumption used above that the red object was directly behind the Skylab and allow it to approach
from some other direction. The simplest alternative assumption is that the red object was in an orbit
at the same altitude as the Skylab and traveling at the same orbital speed but that it was in a different
orbital plane. Assume that the object and Skylab were approaching the location where the orbital planes
cross. Then the distance between them was decreasing, finally reaching "zero" where the planes cross.
(In 3 - D space the intersection of two different orbital planes forms a line that passes through the center of
the earth. Imagine that the orbital planes are viewed from "above" - a point far from the earth - and seen edge on.
Then the crossing of the planes (the intersection line) will appear to be a point. The tracks of two satellites
in different orbital planes, in the vicinity of the orbital plane intersection, appear to be (nearly) straight
lines that cross at the intersection point. This crossing point is above the earth at the altitude of the
satellites. This corresponds to a "sideswipe" scenario where the approach velocity is completely due to a
difference in orbital inclination angle from that of the Skylab's 50-degree inclination.
This scenario is illustrated in the track map, below, where the red object is shown approaching the Skylab
at this 45-degree angle from the west rather than on the other (north) side. In this illustration the angle
between the two orbits is greatly exaggerated.
It is possible to describe a specific notional scenario that will allow for an estimate of the
difference in orbital plane inclinations. Built into this scenario are the following assumptions: (a) the
object must have been about 20 times farther from the Skylab when the first photos were taken than it was
when the last photo was taken, (b) the first photos were taken during the first minute of observation
(c) they were separated by 3 km when the fourth photo was taken, and (d) the last photo was taken about
9.5 minutes after the initial photos. The latter assumption means that the Skylab traveled
about 9.5 x 60 sec x 7.645 km/sec = 4588 km during the assumed time between photos. Twenty times 3 km is 60 km,
so assume that the object was 60 km away from Skylab when it was first observed. Now imagine drawing
two converging straight lines, each almost horizontal, on a paper. The point where they cross
represents the orbital plane crossover line as seen "end-on." Place the Skylab at some distance to the left
of the crossover point on the upper of the two lines and place the object at the same distance on the lower of
the two lines. Draw a line between the points. This line is the separation, S1 and the triangle just
created is isosceles: it has two equal sides. The Skylab and the object are imagined to be moving at
the same speed along these lines toward the right, toward the intersection point. Now move to the
right toward the intersection point and at some position, still to the left of the intersection point,
draw another line that is parallel to S1. This line, called S2, forms a much shorter isosceles triangle.
The distance along each orbit line from S1 to S2 is called A. By the above assumptions
A = 4588 km, S1 = 60 km and S2 = 3 km. Now it is necessary to calculate the vertex angle, a,
from these data. The calculation is straightforward because of the simplified geometry. Using the law
of similar triangles and trigonometric relations one can show that a = 2 arcsin{[(S1/2)-(S2/2)]/A} =
2 arcsin{[30 - 1.5]/4588} = 0.71 degrees. Other reasonable assumptions about the initial and final
separations will lead to similarly small angles.
(By these assumptions, when the Skylab and object were 3 km apart and Skylab was going into darkness,
they were still at some considerable distance from the orbit plane crossover point. That distance,
from the 3 km separation to the crossover, was B = (3/2)/sin(.71) = 121 km.)
divided by 5.5 which makes it even closer to the Skylab orbit.)
Although the scenario just discussed must be immediately discarded since the unknown would pass into
the shadow at within a second of the Skylab, it does illustrate
the following point: if the red object is assumed to have been a man-made satellite, then, because
of the long time it was visible, its orbital inclination must have been very close
to that of Skylab, probably within a degree or so, i.e., say, 49 - 51 degrees.
[Another scenario that has been considered is based on three assumptions that are the "opposite" of what
has been assumed above: 1) all the photos were taken at the beginning of the sighting, say, within
the first minute, 2) the object was closest at the beginning of the sighting and 3) the object was
traveling more slowly than the Skylab. It is also assumed that the object was a man-made satellite 10 m
(or less) in size so that the initial distance at the time of the fourth photo was about 3 km.
(It is further assumed that the three "dot" images that precede the fourth photo
were a result of some rotation of the object such that only single red reflections or single red lights were
visible to the astronauts, an assumption that already rules out this explanation.) Garriott's estimate of
about 5 sec time lag indicates that the object would have been about 38 km behind the Skylab when
it passed into the shadow. This scenario raises the following questions: 1) why didn't the astronauts
continue to take photos since they clearly watched the object for, maybe, 9 minutes after taking 4
photos during the first minute (as assumed), 2) why didn't they indicate in their report that it was
getting dimmer and smaller and clearly moving away from them as it dropped back? Of course, as shown
above, its orbit would have been very close to that of the Skylab.]
A 51-52-degree inclination is in fact one of the most common inclinations in space history for
satellite objects, as it is the prime launch heading for many Russian launches from its original Tyuratam
(Baikonur) launch site. But the higher inclination alternative, 51 degs instead of 49 degs, must have
the red object coming from the north towards the Skylab, whereas, from the testimony of the astronauts,
it appears that the red object was west of the Skylab.
But Russian space objects are carefully tracked by both US and Russian observers as well as others.
NASA signed an agreement with NORAD for NORAD to provide NASA with continually updated computer
projections warning of potentially dangerous close approaches of any and all space objects during a manned space
mission. These projections are generated with a NORAD/USAF computer program called COMBO (Computation for Miss
Between Orbits). The COMBO program does not require that the satellites be directly in view of NORAD
radars and other sensors at the moment of close approach -- the projections are made to any point
in the satellite orbits anywhere in the world using the principle that manmade
satellites follow predictable flight paths.
As of 1973, according to NORAD/ADC, "NASA has not had to maneuver a manned spacecraft to avoid a collision
with another satellite" based on the COMBO warnings. See article by Major Samuel C. Beamer, Commander,
Det 8, 14th Missile Warning Squadron, ADC, Laredo, Texas, "Nerve Center for Space Defense," Air
University Review, September-October 1973 (online at
http://www.airpower.maxwell.af.mil/airchronicles/aureview/1973/sep-oct/beamer.html)
The point of the above calculations and discussion is this: had it been another satellite
in an elliptical orbit that was only a few kilometers or so from Skylab at its closest point,
then it should have been known to the organizations that keep track of all satellites. It
probably would have been within a few tens of kilometers of Skylab when Skylab went through
the areas of the world where there were NASA and NORAD radars and sensors operating. There
was no report that NASA or NORAD detected any object near Skylab, and a near-miss of only
a few kilometer would have been an extremely dangerous situation as tracking radars often have
errors of that magnitude (meaning it could have been a 0-distance, i.e., a direct and
fatal collision).
DISCUSSION
Whatever the object was, it traveled a long distance with the Skylab. They saw it for about 10 minutes.
During that time it and the Skylab traveled about 4,600km (2,800 miles). The big question, then, is this:
is it possible to prove that there was, or at least could have been, a (man-made) earth
satellite which either reflected sunlight as red light (in the absence of atmospheric reddening of sunlight)
or which had several bright red lights and which could have been within a few miles of Skylab for many
minutes without NORAD/NASA detecting it? If an earth satellite, then its orbit was "uncomfortably " close
to the Skylab orbit and it certainly should have been picked up by NORAD at some time during its orbit.
Based on the available information, these authors conclude that there was no man-made satellite that could
explain this sighting and hence the object was truly anomalous.
Further data are being sought.
We thank James Smith for helpful comments and the above-listed web sites for NASA data on Skylab 3.
Note that the above document indicates that the location was over the southwestern Indian
Ocean during revolution # 1863 (see map below).
...........................................................................
Further information is available from debriefing that took place on or about October 4.
Transcriptions are given below the documents.
SKYLAB 3 DEBRIEFING Oct 4, 1973
"GARRIOTT: Do you want to talk about that satellite?
LOUSMA: I saw a couple of satellites that appeared like a satellite would on earth. I saw one that
was not like one you would see on earth, so why don't you mention it?
GARRIOTT: OK. About a week or 10 days before recovery and we were still waiting for information to be
supplied to us about the identification. Jack first notices this rather large red star out the
wardroom window. Upon close examination, it was much brighter than Jupiter or any of the other
planets. It had a reddish hue to it, even though it was well above the horizon. The light from the
Sun was not passing close to the Earth's limb at the time. We observed it for about 10 minutes prior
to sunset. It was slowly rotating because it had a variation in brightness with a 10-seconds period.
As I was saying, we observed it for about 10 minutes, until we went into darkness, and it also
followed us into darkness about 5-seconds later. From the 5 to 10 second delay in it's disappearance we
surmised that it was not more than 30 to 50 nautical miles [35 to 58 statute miles or 56 to 93 km]
from our location. From its original position in the wardroom window, it did not move more than
10 or 20 degrees over the 10 minutes or so that we watched it. Its orbit was very close to that of
our own. We never saw it on any earlier or succeeding orbits and we'd be quite interested in having
its identification established. It's all debriefed in terms of time on channel A, so the precise
timing and location can be picked up from there."
NOTE: according to Oberg the Channel A tape recorder was not on, so the exact time cannot be determined.
NOTE: Soon after this debriefing Garriott told Sparks his best
estimate of the time interval between the Skylab going into sunset and the red
light's disppearance was about 5-6 seconds. Garriott explained exactly how he
counted out the seconds "one thousand one, one thousand two, etc."
One of us (Sparks) asked an expert in satellite observation to take the NORAD tracking data and use
his satellite observation program SkyMapPro, to determine Skylab 3's latitude-longitude-time
coordinates and shadow entry, which turned out to be between 1645 and 1646 GMT (Greenwich Mean Time)
or UTC (Universal Time Coordinated)(it was sunlit at 1645 and in the shadow at 1646). This fits the
above reported 1645 GMT time and places 10-minute duration sighting at about 1635-1645. At 1645-1646
GMT Skylab was over Madagascar Channel at the extreme western edge of the Indian Ocean.
Ground track of Skylab Orbital Workshop
Time Time Latitude Longitude Altitude Speed Lighting
hr:min km(nm) km/sec(nm/sec)
---- ------ --------- -------- ----- ----- -------
20 Sep 1973 16:40 04°46'15"S 026°06'28"E 434.6 (234.7) 7.651(4.13) SUN
20 Sep 1973 16:41 07°44'11"S 028°21'55"E 435.7 7.650 "
20 Sep 1973 16:42 10°41'08"S 030°39'27"E 436.8 7.649 "
20 Sep 1973 16:43 13°36'44"S 032°59'52"E 438.1 7.648 "
20 Sep 1973 16:44 16°30'35"S 035°24'06"E 439.3 7.647 "
20 Sep 1973 16:45 19°22'15"S 037°53'03"E 440.6 7.646 "
20 Sep 1973 16:46 22°11'17"S 040°27'43"E 442.0 7.645 SHADOW
20 Sep 1973 16:47 24°57'09"S 043°09'10"E 443.4 7.643 "
20 Sep 1973 16:48 27°39'17"S 045°58'31"E 444.7 7.642 "
20 Sep 1973 16:49 30°17'01"S 048°56'57"E 446.1 7.641 "
20 Sep 1973 16:50 32°49'38"S 052°05'44"E 447.4 (241.6) 7.640(4.12) "
(270-278 statute (4.75 statute
miles/sec) miles/second)
Latitute-Longitude values are to nearest 1 arcminute and altitudes to nearest kilometer.
A map has been constructed from the above data, showing the shadow boundary on the earth and the track
of the Skylab (white line):
THE PICTURES
The first and last photos presented here are pictures of the earth taken before
and after the sighting. These photos are mentioned in the NASA publication, "Photographic Index and Scene
Identification," NASA-TM-X-69780, that was tabulated by Richard Underwood and co-workers in November, 1973.
(See click here). In that document the photos
are listed twice. Page 41 lists the photos that were contained in film magazine #CX-35. These photos were
taken with a 35 mm Nikon camera with a lens of either 55 or 300 mm focal length using a film called
SO-368 which is comparable to a type of Ektachrome (MS2448). According to the tabulation the photos
labelled SL3-118-2132 through 2135 (frames 1- 4) were associated with the "S-230 experiment." Then photos
SL3-118-2136 and 2137 (frames 5,6)were taken during the Goddard laser beacon experiment. These photos are
described as "underexposed" and "blurred." According to the report on the laser beacon experiment
("Evaluation of Skylab Earth Laser Beacon Imagery," March, 1975;
see Click Here), these beacon photos were taken on September
4 with a hand-held Nikon camera set at f/4.5 to f8 and with a shutter time of 1/500 to 1/250 of a second.
It is important to know this because these may have been the settings and focal length when the four
photos of interest here were taken, photos SL3 - 118 - 2138, 2139,2140 and 2141. The "Photographic Index...."
lists these four as "Satellite, unmanned" on page 41 and as "blank" on page 245. The two different
descriptions may have resulted as follows (speculation): when the photo index team first went through
the photos they did not know that Garriott had photographed a "red satellite," so, looking quickly
at the photos they saw nothing obvious and listed them as blank; later they learned that he photographed
(what he thought was) a satellite and they changed the description, placing the new description in the
front section of the report but leaving the original description at the "far end" (appendix) of the report.
The next photo, SL3 - 118 - 2142 is identified as "Lake Erie, Ohio, Ontario, clouds."
Photos 2137,38,39,40,41 and 42 are shown below.
HERE IS THE SECOND PHOTO OF THE LASER BEACON EXPERIMENT. Although Garriott, who took the picture,
said he could see the beacon it is not evident in this photo (nor is it evident in the preceding photo).
HERE IS THE FIRST PHOTO OF THE OBJECT.
The first photo of the object shows a featureless red "dot".
HERE IS THE SECOND PHOTO OF THE OBJECT.
The second photo shows a slightly larger image, slightly off round and yellowish in the center. The
yellowish color at the center is typical of an overexposed color photo of a distant, bright
red object or light.
HERE IS A BLOWUP OF THE ABOVE IMAGE
THE RED "DOT" IS BARELY VISIBLE IN THE FOLLOWING PHOTO
SO I HAVE DRAWN A CIRCLE AROUND IT.
PHOTO 2141 CONTAINS THE MOST INTERESTING IMAGE BECAUSE IT ALLOWS FOR AN ESTIMATE OF SIZE
AS REPRESENTED BY THE SPACINGS BETWEEN THE RED "BLOB" IMAGES
The photo shown at the beginning of this article is a blowup of the fourth photo to show the odd shape
of the red glowing object (or objects?).
HERE IS ONE MORE OF THE EARTH
AND HERE IS ANOTHER OF THE SKYLAB
..................................................................................
ANALYSIS
The available information does not allow for a specific identification of the red object but
it does allow us to decide whether or not this could have been a man-made earth satellite by making
an estimate of the size. The size estimate is based on the distance estimate and the maximum
spacing between red "dot" or "blob" images in the fourth photo. The estimate is not based on the
size of the image of a single "blob" or "dot" because the "dot" or "blob" image made by a distant
bright light is generally larger than the image would be if the optical system could provide a
perfect "geometrically accurate" image. The extra image size is a result of "diffraction" and "lens
aberration" which combine to make light spread sideways on the film plane.
(NOTE: The image sizes of extremely distant lights that make only "dot" images, like images of stars,
depend only upon the brightness, with the brighter lights making bigger images.
Astronomers have made use of this phenomenon to determine relative star brightnesses from
photographs of stars.) Thus, in the case of a "dot" or "blob" image the most one can say is that
the reflective object or light source which made the image was no larger than the geometric size,
as described below, that corresponds to the size of the dot or blob image.
For objects which are close enough or large enough to be "resolved" by the optical system
(i.e., large enough to have obvious structure such as distinctly separated image dots or shapes)
there is a way to estimate the object size from the image size. For a resolved image the size is
geometrically related to the actual object size by a ratio equation that can be stated as follows:
the image size(width, length, spacing of "blobs" as measured in the focal plane), herein called I,
divided by the focal length, F, is equal to the size (width or length, as measured in a plane that is
perpendicular to the line of sight) of the object, O, divided by the distance to the object, D.
In equation form this is: I/F = O/D, which is based on the law of similar triangles with a common
vertex at the center of the lens. (Usually the resolution capability of the camera is determined
by whether or not this relation holds for a particular image of a particular object at a particular
distance.) What can and what can't be resolved depends upon the optical quality of the camera lens
and other factors related to the camera and film.
If an object or light makes a "dot" image on the film, then the calculated size of the object, O = D(I/F),
is (probably) much larger than the actual size of the object, as described in the previous paragraph.
However, when there are distinguishable features of the image that are clearly separated by distance on
the film it is possible to estimate the spacing between the features which made the images.
In case of the first three Skylab 3 photos it would be of little value to try to estimate the size
of the red object by measuring the size of individual "dot" images and multiplying by D/F. But in the
fourth photo there is a considerable spacing between images of "dots". This means that the distance
between dots has been resolved by the camera even though the dots themselves have not been resolved.
[There "dots" are somewhat elongated so, strictly speaking, they are not exactly "dots."]
On a direct copy (1:1) of the original film the maximum spacing between "dots," specifically the
spacings betwee the centers of the upper and lower dots is about 0.87 mm. The focal length
of the camera was (probably) 300 mm (it might have been 55 mm; see below). Thus the ratio I/F
equals about 0.87/300 = 0.0029 (in radian measure) (or, if the focal length was 55 mm, the
ratio is .87/55 = 0.0158). As pointed out above, this ratio is proportional to the object size, as
measured perpendicular the line of sight (i.e., as "projected" onto a plane surface that is perpendicular
to the line of sight), divided by the distance.
Now it is necessary to know the estimated distance D in order to calculate
the object size, O, which is the "projected" spacing between whatever made the upper and lower dots.
Garriott, who was the first scientist at the Skylab, had the presence of mind to count the seconds
between when the Skylab went into the earth's shadow and when the object, which appeared to be following
the Skylab, disappeared. He assumed the object was visible only by reflected light, an assumption
which would be reasonable if the object had been another orbiting satellite. (If the object had been a
bright source of red light [there were no satellites with bright red lights attached] then his
assumption would not necessarily apply.) He also assumed that the red object was behind the Skylab and
traveling at the same speed which, according to the table above, was about 7.64 km/sec.
Based on these assumptions, and his estimated time until it disappeared, 5 - 10 seconds, the calculated
distance between the Skylab and the object was (5 to 10 seconds) times 7.64 km/sec = 38 to 76 km.
Assuming that the fourth photo was taken just before Skylab went into the shadow (an "unfounded"
assumption since we don't know at what times, during the sighting, any of the photos were taken) and
using the shorter distance estimate along with the ratio I/F = 0.0029 we find O = D(I/F) = D(0.0029) =
38 km x 0.0029 = 0.11 km = 110 m or about 336 ft as projected onto a plane perpendicular to the line of sight.
If D were 76 km the calculated size would be twice as large. (Sparks spoke to Garriott in October,
1973, soon after the mission had been completed. Sparks asked how Garriott estimated the time
until the "satellite" went into the shadow. Garriott said he counted off the seconds as "one thousand one,
one thousand two", etc. and estimated 5 to 6 seconds this way. Hence these authors believe that
the lower time estimate is probably more likely to be correct.) The width of the image, measured from the
middle "dot" to the end of the "extension" at the right side of the middle dot, is about 1/3 of the spacing
between the upper and lower dots. Thus the width would be about 37 m or 112 ft if 38 km away and 74 m
or 224 ft if 76 km away.
Now let's assume that the various dots that form the fourth image are reflections from points or small
reflective areas of a "typical" earth satellite (manmade!). Then this calculation indicates that the
spacing between two reflective portions of the satellite was 336 or more feet. However, in 1973
there was no satellite remotely approaching that size (and only the space station is comparable now,
over 30 years later).
(NOTE: The Skylab-3 itself with docked Apollo module was the largest satellite
object in orbit at the time, about 150 feet long.)
The calculation just done is a lower bound on the spacing between the outermost lights for the assumed
distance. The calculation assumes that an imaginary straight line connecting the upper and lower red lights
(or red reflective areas) on the object was perpendicular to the line of sight from the camera to the object.
If a line drawn between the red lights was not perpendicular to the line of sight, then the spacing
between them was greater than estimated above (by the factor 1/sinA where A is the acute angle between
the sighting line and the line between points on the object; there is no way of knowing what A was so 90
degrees is assumed here, resulting an an estimate of the minimum spacing between "lights").
(NOTE: when this discussion was first published there was a question as to whether the focal length
was 300 or 55 mm. One of us (Maccabee) had made a notation (in 1977 when copies of these photos were first
obtained) that the focal length was 300 mm, but we recently learned that there were 5 Nikon cameras and each
had either a 55 or 300 mm lens. This raised the question of whether or not the focal length could have
been 55 mm. The question has now been resolved by the combination of the photo index and the report on the
beacon experiment, as described above. Although it was 16 days between the laser beacon experiment and the
date of these four photos, we believe it highly unlikely that they would have changed the lens on this camera
if they wanted to take wide angle (55 mm) photos since they had four other Nikon cameras available,
several of which had 55 mm lenses. We assume that Garriott simply grabbed a Nikon with the 300 mm lens,
namely, the one used many days before during the beacon experiment.)
WHAT COULD IT HAVE BEEN?
The possible identifications divide into two general classes: (a) conventional explanations that assume a
manmade/earth-launched satellite and (b) unconventional explanations that assume it was not an earth-
launched satellite. Explanations of type (b) allow for the object to do "anything" because the object is
assumed not to be limited by earth's gravity. In particular, (b)-type explanations allow for the object to
make non-orbital motions, such as travel at velocities greater or less than the orbital velocity that is
appropriate for its altitude. (For an object that is "trapped" by earth's gravity the orbital velocity depends
upon the radial distance from the center of the earth). For example, one might imagine that the reason the
object appeared as a dot in three photos and then as a structured object in the fourth is that the object
was at a large distance from the Skylab during the (unknown) time duration of the first three photos and then,
between the third and fourth photos, it suddenly moved much closer to the Skylab, at which time
its structure could be resolved by the camera. If this happened, then the object made motions that are not
consistent with being a satellite subject to orbital motion around the earth. There are many other non-orbital
motions which one could imagine if the object was not a satellite and one could spend a long time trying to
study all of these. However, the main point is that if the object was not confined to an orbit, since there
were no man-made objects capable of such motions in the vicinity of the Skylab, the object would have to be
considered truly anomalous, an Unidentified Space Object (USO, not to be confused with "USO: Unidentified
Submerged Object").
Explanations of type (a) are the only ones that can be tested against known physics of satellites so the
question arises, does the conventional hypothesis, that it was a satellite, fit the data?
This hypothesis runs into trouble immediately when compared with the photographic and visual data. Both visually
and photographically the object appeared bright red in color. Noting this color but assuming it was a
satellite or some object that could have been reflecting sunlight, Garriott pointed out that when it was first
seen and for about 10 minutes afterward the object and Skylab were not in a region of the orbit
where the sunlight would have passed through the earth's atmosphere and therefore would
have been reddened by scattering from particulate matter in the atmosphere (as in a red
sunset). Hence the red color could not have been a result of the reflection of reddened sunlight from a
typical satellite. To explain the red color Garriott may have assumed (although he never mentioned this
assumption) that the satellite was painted red. He awaited an identification by NASA, but there was none
because there were no satellites with red surfaces. The surfaces are typically painted black or white
or are unpainted bare metal.
An alternative hypothesis is that it was a satellite emitting intense red light. Garriott probably didn't consider
this to be a possible explanation since there were no satellites with bright red lights.
The assumption that the astronauts were watching the red lights on an object and not reddened solar reflections
leads to the conclusion that Garriott's method of estimating the distance could have been wrong. If the object
had red lights then the fact that these lights disappeared shortly after Skylab went into the earth's shadow would
not necessarily mean that it had also traveled into the earth's shadow. It could have been at any distance
and simply turned off its lights 5 - 10 sec after the Skylab "disappeared" into darkness. But if this were true
then Garriott's estimate of distance as described above could be a happy coincidence or it could have been a
mistake.
One interesting consequence of the red light hypothesis, which includes the assumption that the lights were turned
off 5 - 10 seconds after Skylab went into darkness, is that the object could have been a lot closer to Skylab
than Garriott estimated. Assume, for example, that the spacing between the lights on the assumed "satellite" was
10 m (33 ft), the maximum size of man-made satellites (including booster rockets) in 1973. Then the formula
used above can be rewritten to give the distance that an object would have to be away from the camera in order
to produce an image of the measured size. With O = 10 m, D = O(F/I) = O(1/0.0029) = 10/0.0029 = 3.4 km (2.1 mi).
If the spacing of lights were less than 10 m then the distance would also be proportionally smaller.
Another hypothesis has been considered in order to explain the difference in image size between the first
three photos and the fourth. This hypothesis is based on the assumption that the failure of the camera and
film to resolve the structure during the first three photos is a result of a great distance to the supposed
satellite. It is further assumed that the supposed satellite was in an elliptical orbit and was approaching the
Skylab from the rear. The astronauts estimated that the object was in sight for about 10 minutes. The large
difference in image sizes can be explained if one assumes that the first three photos were taken during the first
minute while the satellite was still very far away and that the fourth was taken roughly 9 - 10
minutes later when the assumed satellite was much closer.
Further to this hypothesis, one can assume that during the first minute, when the object was varying in luminosity
with about a 10-second cycle period, the brightness (or size) of the image might have varied depending on when the
photo was taken during the 10-second cycle (during a low brightness phase or high or in between), and this
variation could have happened without much actual distance change. For example, if the object had been about 10 m
in size and approaching the Skylab from behind at a speed 0.1 km/sec (360 kph) greater than that of the Skylab and
if the first 3 photos had been taken within a ten second period near the end of the first minute of observation,
then, during this initial ten second period, the distance to the Skylab would have decreased by only 1 km. Suppose
the separation had been 63 km when the object was first seen. Then, about 10 minutes = 600 seconds after it was
first seen, i.e., at the assumed time of the fourth photo, the separation would have been only about
63 km - (600 sec)(.1 km/sec) = 3 km and the angular size of the object, (O/D), would have been twenty times
greater than at 60 km and more of the structure would be resolved by the camera.
(Note: this hypothesis differs from the similar one
described previously in the following way: whereas previously it was assumed that the object maintained a
roughly fixed, large distance during the time of the first three photos and then it suddenly made an
"un-satellite-like" rapid motion toward the Skylab, in this hypothesis the object/satellite is assumed to
have approached in a steady manner consistent with an elliptical orbit as described below.)
The preceding hypothesis can be made somewhat more specific. Comparing the size of the image in the
fourth photo with the sizes of the "dot" images in the preceding photos
it would appear that the object would have had to be 20 or more times farther from the Skylab during the first
three photos than at the time of the last photo in order to be far enough away to prevent the camera from
resolving the image. Calculations presented above show that a man-made satellite (ca. 10 m in length) would
have to be about 3 km away in order to make an image the size of the one in the fourth photo,
assuming the 300 mm focal length lens was used. Twenty or more times farther away would place the initial
distance at 60 km or more. The above hypothesis is that the object crossed this 60 km distance in about 10
minutes or 600 seconds meaning that it would have been traveling faster than the Skylab, specifically,
60 km /600 sec = 0.100 km/s (about 200 mph) faster The Skylab was traveling at about 7,640 m/s, so the object,
under these assumptions, would have traveled at about 7,740 m/s. The Skylab was in a nearly perfectly circular
orbit (eccentricity esentially zero). The object is assumed to have been traveling in the same orbital plane
and at the altitude of the Skylab when it disappeared, about 442 km. It was traveling faster, under these
assumptions, and so its orbital eccentricity was larger than zero. The equation for eccentricity,
e, based on the radial distance from the center of the earth at perigee, Rp, and the velocity at
perigee, Vp, is e = (RpVp^2/GM) - 1 where GM = 4 x 10^14 m^3/sec^2. Using Rp = A + Re = (altitude +
radius of earth) = 442 km + 6378 km = 6820 km and Vp = 7,740 m/s yields e = .0214.
The semimajor axis of this ellipse is a = Rp/(1-e) = 6,969 km and the apogee height that corresponds to
these values of Rp and e is Ra - Re = 2a - Rp - Re = 7,118 km - 6,378 km = 740 km. The period of the object in
the elliptical orbit would have been P = 2 (pi)(a^1.5)/(GM)^.5 = 5,779 s or 96.3 minutes which is only about 2
minutes longer than the Skylab orbit time (93.4 min). The result of this calculation is that the object
would have been very close to the Skylab for many minutes. The results would be similar for an object
in a slightly different orbital plane, a plane close to the 50 degree inclination of the Skylab orbit
in order to be within viewing range for up to 10 minutes. Few satellites (if any) other than
the Skylab have been launched into 50 degree orbits.
This set of assumptions does have a problem with the timing. Assume, as above, that the fourth photo was
taken just before the Skylab went into the shadow. Then the object/satellite was about 3 km away from the
Skylab at that time. It was approaching the shadow at a speed of about 7.7 km/sec. If it had been directly
behind the Skylab it would have crossed the shadow boundary about (3 km/7.7 km/sec =) 0.39 sec after the Skylab,
clearly much less time than estimated by Garriott. Suppose, instead, that the satellite had been 3 km higher
and "above" (rather than directly behind) the Skylab when the Skylab went into the shadow. Then the distance
(along its orbit) of the object from the shadow boundary would have been about 3 km/tan 21 = 7.8 km at the time
the Skylab went into the shadow. (In this equation 21 degrees is the vertical angle-the angle measured in
the vertical plane-that the orbit makes with the shadow boundary at the point where the orbit crosses the
boundary.) The satellite would travel this distance in about 1 second, again much less than the time difference
estimated by Garriott.
The preceding calculations illustrate the problem with assuming that the object was the size of a large
man-made satellite and therefore about 3 km away when the fourth photo which was taken. The problem is that,
regardless of the direction assumed from the Skylab to the supposed satellite, the satellite would pass
into the shadow no longer than about 1 second after the Skylab. If the assumed satellite were smaller than
10 m in size, say a meter or so, such as something ejected from the Skylab then it would have been even
closer than 3 km and the time lag would have been even less.
Assume, again, that the satellite or object was directly behind the Skylab. For it to pass into the shadow
5 seconds after the Skylab, the distance of the satellite from the shadow boundary, as measured along its
orbit, would have had to have been at least (5 sec x 7.7 km/sec =) 38.5 km at the time when the Skylab
crossed the shadow boundary. Thus at the time of the fourth photo (assumed to have been taken just before
the Skylab went into the shadow) the distance would have been 38 km or greater.
If the supposed satellite had been at a higher altitude and "above" the Skylab when Skylab crossed the shadow
boundary it could have been closer to the Skylab than 38 km and still take 5 seconds to reach the shadow boundary.
For this to occur the distance above the Skylab would have been (38.5 tan 21 =) 14.8 km which means that
at the time of the fourth photo the distance would have been around 15 km. Using this distance to calculate
size yields (0.0029 x 15,000 m =) 43 m, which is still four times the size of the largest satellite
(other than the Skylab itself). Thus we see that it is "difficult" (impossible?) for the satellite (or
ejected debris) hypothesis to satisfy two sighting requirements at the same time:
1) the size of the object, as determined by the distance and angular size (from the fourth photo)
must be comparable to or smaller than that of a large satellite and
2) the time difference between when the Skylab crossed the shadow boundary and when the object/satellite
crossed the boundary must be 5 seconds or more.
Unfortunately there is no information on the actual direction from the Skylab to the object. Garriott
said the direction to the object moved no more than 10 - 20 degrees, but he didn't provide the actual
sighting direction. His assumption that the object followed Skylab into darkness suggests that
the object was "behind" the Skylab, but in what direction behind we do not know. We can, therefore,
drop the tacit assumption used above that the red object was directly behind the Skylab and allow it to approach
from some other direction. The simplest alternative assumption is that the red object was in an orbit
at the same altitude as the Skylab and traveling at the same orbital speed but that it was in a different
orbital plane. Assume that the object and Skylab were approaching the location where the orbital planes
cross. Then the distance between them was decreasing, finally reaching "zero" where the planes cross.
(In 3 - D space the intersection of two different orbital planes forms a line that passes through the center of
the earth. Imagine that the orbital planes are viewed from "above" - a point far from the earth - and seen edge on.
Then the crossing of the planes (the intersection line) will appear to be a point. The tracks of two satellites
in different orbital planes, in the vicinity of the orbital plane intersection, appear to be (nearly) straight
lines that cross at the intersection point. This crossing point is above the earth at the altitude of the
satellites. This corresponds to a "sideswipe" scenario where the approach velocity is completely due to a
difference in orbital inclination angle from that of the Skylab's 50-degree inclination.
This scenario is illustrated in the track map, below, where the red object is shown approaching the Skylab
at this 45-degree angle from the west rather than on the other (north) side. In this illustration the angle
between the two orbits is greatly exaggerated.
It is possible to describe a specific notional scenario that will allow for an estimate of the
difference in orbital plane inclinations. Built into this scenario are the following assumptions: (a) the
object must have been about 20 times farther from the Skylab when the first photos were taken than it was
when the last photo was taken, (b) the first photos were taken during the first minute of observation
(c) they were separated by 3 km when the fourth photo was taken, and (d) the last photo was taken about
9.5 minutes after the initial photos. The latter assumption means that the Skylab traveled
about 9.5 x 60 sec x 7.645 km/sec = 4588 km during the assumed time between photos. Twenty times 3 km is 60 km,
so assume that the object was 60 km away from Skylab when it was first observed. Now imagine drawing
two converging straight lines, each almost horizontal, on a paper. The point where they cross
represents the orbital plane crossover line as seen "end-on." Place the Skylab at some distance to the left
of the crossover point on the upper of the two lines and place the object at the same distance on the lower of
the two lines. Draw a line between the points. This line is the separation, S1 and the triangle just
created is isosceles: it has two equal sides. The Skylab and the object are imagined to be moving at
the same speed along these lines toward the right, toward the intersection point. Now move to the
right toward the intersection point and at some position, still to the left of the intersection point,
draw another line that is parallel to S1. This line, called S2, forms a much shorter isosceles triangle.
The distance along each orbit line from S1 to S2 is called A. By the above assumptions
A = 4588 km, S1 = 60 km and S2 = 3 km. Now it is necessary to calculate the vertex angle, a,
from these data. The calculation is straightforward because of the simplified geometry. Using the law
of similar triangles and trigonometric relations one can show that a = 2 arcsin{[(S1/2)-(S2/2)]/A} =
2 arcsin{[30 - 1.5]/4588} = 0.71 degrees. Other reasonable assumptions about the initial and final
separations will lead to similarly small angles.
(By these assumptions, when the Skylab and object were 3 km apart and Skylab was going into darkness,
they were still at some considerable distance from the orbit plane crossover point. That distance,
from the 3 km separation to the crossover, was B = (3/2)/sin(.71) = 121 km.)
divided by 5.5 which makes it even closer to the Skylab orbit.)
Although the scenario just discussed must be immediately discarded since the unknown would pass into
the shadow at within a second of the Skylab, it does illustrate
the following point: if the red object is assumed to have been a man-made satellite, then, because
of the long time it was visible, its orbital inclination must have been very close
to that of Skylab, probably within a degree or so, i.e., say, 49 - 51 degrees.
[Another scenario that has been considered is based on three assumptions that are the "opposite" of what
has been assumed above: 1) all the photos were taken at the beginning of the sighting, say, within
the first minute, 2) the object was closest at the beginning of the sighting and 3) the object was
traveling more slowly than the Skylab. It is also assumed that the object was a man-made satellite 10 m
(or less) in size so that the initial distance at the time of the fourth photo was about 3 km.
(It is further assumed that the three "dot" images that precede the fourth photo
were a result of some rotation of the object such that only single red reflections or single red lights were
visible to the astronauts, an assumption that already rules out this explanation.) Garriott's estimate of
about 5 sec time lag indicates that the object would have been about 38 km behind the Skylab when
it passed into the shadow. This scenario raises the following questions: 1) why didn't the astronauts
continue to take photos since they clearly watched the object for, maybe, 9 minutes after taking 4
photos during the first minute (as assumed), 2) why didn't they indicate in their report that it was
getting dimmer and smaller and clearly moving away from them as it dropped back? Of course, as shown
above, its orbit would have been very close to that of the Skylab.]
A 51-52-degree inclination is in fact one of the most common inclinations in space history for
satellite objects, as it is the prime launch heading for many Russian launches from its original Tyuratam
(Baikonur) launch site. But the higher inclination alternative, 51 degs instead of 49 degs, must have
the red object coming from the north towards the Skylab, whereas, from the testimony of the astronauts,
it appears that the red object was west of the Skylab.
But Russian space objects are carefully tracked by both US and Russian observers as well as others.
NASA signed an agreement with NORAD for NORAD to provide NASA with continually updated computer
projections warning of potentially dangerous close approaches of any and all space objects during a manned space
mission. These projections are generated with a NORAD/USAF computer program called COMBO (Computation for Miss
Between Orbits). The COMBO program does not require that the satellites be directly in view of NORAD
radars and other sensors at the moment of close approach -- the projections are made to any point
in the satellite orbits anywhere in the world using the principle that manmade
satellites follow predictable flight paths.
As of 1973, according to NORAD/ADC, "NASA has not had to maneuver a manned spacecraft to avoid a collision
with another satellite" based on the COMBO warnings. See article by Major Samuel C. Beamer, Commander,
Det 8, 14th Missile Warning Squadron, ADC, Laredo, Texas, "Nerve Center for Space Defense," Air
University Review, September-October 1973 (online at
http://www.airpower.maxwell.af.mil/airchronicles/aureview/1973/sep-oct/beamer.html)
The point of the above calculations and discussion is this: had it been another satellite
in an elliptical orbit that was only a few kilometers or so from Skylab at its closest point,
then it should have been known to the organizations that keep track of all satellites. It
probably would have been within a few tens of kilometers of Skylab when Skylab went through
the areas of the world where there were NASA and NORAD radars and sensors operating. There
was no report that NASA or NORAD detected any object near Skylab, and a near-miss of only
a few kilometer would have been an extremely dangerous situation as tracking radars often have
errors of that magnitude (meaning it could have been a 0-distance, i.e., a direct and
fatal collision).
DISCUSSION
Whatever the object was, it traveled a long distance with the Skylab. They saw it for about 10 minutes.
During that time it and the Skylab traveled about 4,600km (2,800 miles). The big question, then, is this:
is it possible to prove that there was, or at least could have been, a (man-made) earth
satellite which either reflected sunlight as red light (in the absence of atmospheric reddening of sunlight)
or which had several bright red lights and which could have been within a few miles of Skylab for many
minutes without NORAD/NASA detecting it? If an earth satellite, then its orbit was "uncomfortably " close
to the Skylab orbit and it certainly should have been picked up by NORAD at some time during its orbit.
Based on the available information, these authors conclude that there was no man-made satellite that could
explain this sighting and hence the object was truly anomalous.
Further data are being sought.
We thank James Smith for helpful comments and the above-listed web sites for NASA data on Skylab 3.
It is possible to describe a specific notional scenario that will allow for an estimate of the
difference in orbital plane inclinations. Built into this scenario are the following assumptions: (a) the
object must have been about 20 times farther from the Skylab when the first photos were taken than it was
when the last photo was taken, (b) the first photos were taken during the first minute of observation
(c) they were separated by 3 km when the fourth photo was taken, and (d) the last photo was taken about
9.5 minutes after the initial photos. The latter assumption means that the Skylab traveled
about 9.5 x 60 sec x 7.645 km/sec = 4588 km during the assumed time between photos. Twenty times 3 km is 60 km,
so assume that the object was 60 km away from Skylab when it was first observed. Now imagine drawing
two converging straight lines, each almost horizontal, on a paper. The point where they cross
represents the orbital plane crossover line as seen "end-on." Place the Skylab at some distance to the left
of the crossover point on the upper of the two lines and place the object at the same distance on the lower of
the two lines. Draw a line between the points. This line is the separation, S1 and the triangle just
created is isosceles: it has two equal sides. The Skylab and the object are imagined to be moving at
the same speed along these lines toward the right, toward the intersection point. Now move to the
right toward the intersection point and at some position, still to the left of the intersection point,
draw another line that is parallel to S1. This line, called S2, forms a much shorter isosceles triangle.
The distance along each orbit line from S1 to S2 is called A. By the above assumptions
A = 4588 km, S1 = 60 km and S2 = 3 km. Now it is necessary to calculate the vertex angle, a,
from these data. The calculation is straightforward because of the simplified geometry. Using the law
of similar triangles and trigonometric relations one can show that a = 2 arcsin{[(S1/2)-(S2/2)]/A} =
2 arcsin{[30 - 1.5]/4588} = 0.71 degrees. Other reasonable assumptions about the initial and final
separations will lead to similarly small angles.
(By these assumptions, when the Skylab and object were 3 km apart and Skylab was going into darkness,
they were still at some considerable distance from the orbit plane crossover point. That distance,
from the 3 km separation to the crossover, was B = (3/2)/sin(.71) = 121 km.)
divided by 5.5 which makes it even closer to the Skylab orbit.)
Although the scenario just discussed must be immediately discarded since the unknown would pass into
the shadow at within a second of the Skylab, it does illustrate
the following point: if the red object is assumed to have been a man-made satellite, then, because
of the long time it was visible, its orbital inclination must have been very close
to that of Skylab, probably within a degree or so, i.e., say, 49 - 51 degrees.
[Another scenario that has been considered is based on three assumptions that are the "opposite" of what
has been assumed above: 1) all the photos were taken at the beginning of the sighting, say, within
the first minute, 2) the object was closest at the beginning of the sighting and 3) the object was
traveling more slowly than the Skylab. It is also assumed that the object was a man-made satellite 10 m
(or less) in size so that the initial distance at the time of the fourth photo was about 3 km.
(It is further assumed that the three "dot" images that precede the fourth photo
were a result of some rotation of the object such that only single red reflections or single red lights were
visible to the astronauts, an assumption that already rules out this explanation.) Garriott's estimate of
about 5 sec time lag indicates that the object would have been about 38 km behind the Skylab when
it passed into the shadow. This scenario raises the following questions: 1) why didn't the astronauts
continue to take photos since they clearly watched the object for, maybe, 9 minutes after taking 4
photos during the first minute (as assumed), 2) why didn't they indicate in their report that it was
getting dimmer and smaller and clearly moving away from them as it dropped back? Of course, as shown
above, its orbit would have been very close to that of the Skylab.]
A 51-52-degree inclination is in fact one of the most common inclinations in space history for
satellite objects, as it is the prime launch heading for many Russian launches from its original Tyuratam
(Baikonur) launch site. But the higher inclination alternative, 51 degs instead of 49 degs, must have
the red object coming from the north towards the Skylab, whereas, from the testimony of the astronauts,
it appears that the red object was west of the Skylab.
But Russian space objects are carefully tracked by both US and Russian observers as well as others.
NASA signed an agreement with NORAD for NORAD to provide NASA with continually updated computer
projections warning of potentially dangerous close approaches of any and all space objects during a manned space
mission. These projections are generated with a NORAD/USAF computer program called COMBO (Computation for Miss
Between Orbits). The COMBO program does not require that the satellites be directly in view of NORAD
radars and other sensors at the moment of close approach -- the projections are made to any point
in the satellite orbits anywhere in the world using the principle that manmade
satellites follow predictable flight paths.
As of 1973, according to NORAD/ADC, "NASA has not had to maneuver a manned spacecraft to avoid a collision
with another satellite" based on the COMBO warnings. See article by Major Samuel C. Beamer, Commander,
Det 8, 14th Missile Warning Squadron, ADC, Laredo, Texas, "Nerve Center for Space Defense," Air
University Review, September-October 1973 (online at
http://www.airpower.maxwell.af.mil/airchronicles/aureview/1973/sep-oct/beamer.html)
The point of the above calculations and discussion is this: had it been another satellite
in an elliptical orbit that was only a few kilometers or so from Skylab at its closest point,
then it should have been known to the organizations that keep track of all satellites. It
probably would have been within a few tens of kilometers of Skylab when Skylab went through
the areas of the world where there were NASA and NORAD radars and sensors operating. There
was no report that NASA or NORAD detected any object near Skylab, and a near-miss of only
a few kilometer would have been an extremely dangerous situation as tracking radars often have
errors of that magnitude (meaning it could have been a 0-distance, i.e., a direct and
fatal collision).