Could Columbia have been Imaged by a Keyhole?

by Ted Molczan - 2003 February 17

 

Intoduction

1. Factors Affecting Image Quality

2. ERS-1 Imaged by Spot 4

3. STS 107 Encounters with USA 129 and USA 161

4. STS 1 Encounters with KH 11-2

5. STS 2 Encounters with KH 11-3

6. Skylab Encounters with KH 8-38
 

Revisions

2003 Mar 11: added URL to Aviation Week and Space Technology report that STS 2 had been imaged by a KH-11 (Section 5).

2003 Mar 14: added URL to Washington Post article of same date (Introduction); added analysis of STS 107 encounters with USA 161 (Section 3).

2003 Mar 15: added URL to Washington Post article of same date to Introduction; rewrote Introduction.

2003 Mar 28: added Mar 25 letter from NASA to NIMA, revealing that NIMA will automatically image all future shuttle missions, and provide the images to NASA (Introduction).

2003 Apr 03: added Section 6: Skylab and KH 8-38.

Introduction

The tragic loss of Columbia and her crew renewed public interest in the possibility of using the Keyhole (aka KH-11) electro-optical imaging satellites to assist in detecting damage to shuttles.

One journalist sought confirmation of decades-old reports of Columbia having been imaged by a KH-11 on the inaugural shuttle mission, STS 1. Another asked the same question in regard to STS 107. Had there been any encounters sufficiently close to have enabled high-resolution imaging? Their curiosity aroused mine, so I agreed to do some research.

My orbital analysis software was not designed to efficiently analyse satellite encounters, but with some effort I coaxed from it sufficient information to conclude that Columbia could have been imaged by a KH-11 during STS 1 and STS 107, as reported by MSNBC and The Washington Post.

Subsequently, I wrote software to compute detailed information about satellite encounters, used to produce the results presented here.

Reports published by The Washington Post on 2003 Mar 14 and Mar 15 left little doubt that the KH-11 satellites are capable of imaging other satellites at high resolution.

Most revealing was the report on Mar 14 that, "a NIMA employee who had 'heard' NASA might need an image of the shuttle", contacted NASA, and twice offered, "during Columbia's flight to produce an image that would clarify whether the shuttle had damage that might put its crew at risk during its landing Feb. 1." Since NIMA disseminates KH-11 imagery to the intelligence community, it seems likely that its employee had a factual basis in making the offer.

The article also cited un-named sources as having said, "there was no question that NIMA's satellites could have photographed the shuttle with sufficient detail to make clear whether its crew faced any risk on landing because of the damage." (The reference to damage relates to the foam impact that occurred during ascent.)

As a result of a meeting with NASA on 2003 Mar 13, NIMA agreed to automatically image shuttles as targets of opportunity, and provide the images to NASA, as revealed in the letter of 2003 Mar 25 from the NASA Administrator to the NIMA Director.

As of this writing, the precise cause of the Columbia accident is unknown, so it is impossible to know whether or not high-resolution imagery would have been useful. There may not have been any outwardly visible signs. Even if there was something portentous to be seen, the outcome might have been the same. At this point in the investigation, having clear imagery of the left wing probably would be valuable, but that is easy to say in hindsight.

1. Factors Affecting Image Quality

The technical capabilities of the Keyholes are classified; however, analysis of public information reveals that their resolution is about 10 cm at a distance of 300 km.

Orbital encounters may occur at a very high relative angular velocity, which could impair resolution due to excessive blurring. A KH-11's maximum angular velocity relative terrestrial targets is 1.6 deg/s, so it should be able to image satellites moving at least as fast.

KH-11 satellites and their successors are believed to take advantage of their motion relative to target to build their images one line at a time. Evidence of this appears in this leaked KH 11 image from 1984 - note the parallel streaks in the dark portions of the scene. This type of sensor may be more tolerant of high rates of angular velocity than those that image an entire frame at once, so KH-11s may be able to cope with angular velocities exceeding 1.6 deg/s.

Illumination also affects image quality. If the target is between the observer and the sun, then the observer will see only its shadow. At the other extreme, the target is fully bathed in sunlight, washing out details. The ideal occurs somewhere between these extremes, creating some shadows on the target that may reveal surface irregularities.

2. ERS-1 Imaged by Spot 4

One example of a satellite imaging another satellite is this Spot 4 image of ERS-1, taken in 1998. I have computed the circumstances of the encounter, showing that ERS-1 was 41 km below Spot 4, moving with a relative angular velocity of 0.05 deg/s - about one tenth Spot 4's angular velocity relative terrestrial targets.

Spot 4's 50 cm resolution image of ERS-1 is impressive, but a Keyhole has nearly 40 times greater resolution for a given distance from target.

3. STS 107 Encounters with USA 129 and USA 161

Below are Columbia's seven sunlit encounters within 500 km of USA 129 and USA 161; details and supporting information are available here.

                           Range  Vang   ILL  Res
Keyhole   Date     UTC      km    deg/s   %   cm
-------  ------  --------  -----  -----  ---  ---
USA 161  Jan 18  02:25:42   468    1.36   64   16

USA 129  Jan 18  12:46:47   306    1.56   36   10

USA 129  Jan 20  05:18:22   190    1.61   22    6
USA 129  Jan 20  05:18:49   198    1.48   45    6

USA 129  Jan 25  05:18:12   221    1.60   76    7
USA 129  Jan 25  05:18:45   221    1.61   80    7

USA 129  Jan 29  08:13:45   317    1.58   25   11
USA 129  Jan 29  08:14:07   315    1.60   22   11

USA 129  Jan 30  03:42:00   227    1.55   86    8
USA 129  Jan 30  03:42:37   228    1.54   73    8

USA 129  Jan 30  23:10:43   429    1.14   97   14

Estimated resolution varied from 6 cm to 16 cm, HRSI tiles are 15 cm by 15 cm. The RCC panels on the leading edges of the wings are roughly 66 cm wide.

Most of the above occurred about 15 s before or after closest approach, selected on the basis that angular velocity did not exceed the maximum encountered by KH-11's imaging terrestrial targets.

By far the highest angular velocity, 9.9 deg/s, occurred at the closest approach of 2003 Jan 20 at 05:18:35 UTC; however, there was an even greater problem - Columbia would have been illuminated from behind, so USA 129 would have been looking into its shadow. As shown above, about 14 s before and after, the angular velocity was at or below 1.6 deg/s, and the illumination was much better. At those points, the range was just under 200 km, corresponding to resolution of about 6 cm.

USA 129 was tracked frequently throughout STS 107 by Anthony Beresford, of Australia, and Greg Roberts, of South Africa. Pierre Neirinck, of France, provided orbital analysis in support of the observations.

Originally, my analysis did not include USA 116 and USA 161, because during most of STS 107, they were not in a visibility window for any of my fellow hobbyists who track such objects. Their visibility window for the Southern Hemisphere observers ended about Jan 16, the day Columbia was launched. At that time, there was some reason to believe that re-boost manoeuvres by either one could have been imminent. Such manoeuvres would have been sufficiently large to have invalidated any encounter calculations based on the pre-manoeuvre orbits, so I was not sufficiently confident in their orbits to include USA 116 and USA 161 in my analysis.

USA 161 was recovered on 2003 Feb 20, by Bjorn Gimle, of Sweden, soon after a visibility window opened for Northern Hemisphere observers. Analysis of subsequent tracking by David Brierley, of the U.K., revealed that the object had not manoeuvred since it was last observed, on 2003 Jan 16. Moreover, its rate of decay had been fairly constant during the interim, which provided a confident basis to analyse the possibility of close encounters with STS 107. A single encounter met my criteria: range within 500 km, and Columbia illuminated by the sun, as listed in the above table.

As of this writing, the precise cause of the Columbia accident is unknown, so it is impossible to know whether or not high-resolution imagery would have been useful. There may not have been any outwardly visible signs. Even if there was something portentous to be seen, the outcome might have been the same. At this point in the investigation, having clear imagery of the left wing probably would be valuable, but that is easy to say in hindsight.

4. STS 1 Encounters with KH 11-2

STS 1 was launched on 1981 Apr 12 and landed on Apr 14. Six months later, Aviation Week and Space Technology reported that Columbia had been photographed by a KH-11 to check for damaged tiles. I have found several possible high resolution imaging opportunities during encounters with KH 11-2 on three consecutive revolutions on Apr 13:

                  Range  Vang   ILL  Res
 Date     UTC      km    deg/s   %   cm
------  --------  -----  -----  ---  ---
Apr 13  15:11:34   319    1.56   82   11
Apr 13  15:12:09   318    1.58   77   11

Apr 13  16:41:57   208    1.56   47    7
Apr 13  16:42:30   211    1.52   44    7

Apr 13  18:12:34   451    1.38   25   15

The late Anthony Kenden reported contrary findings in his 1983 paper, Was Columbia Photographed By A KH-11? - Journal of the British Interplanetary Society, Vol. 36 pp. 73-77.

Kenden found the same close encounters, but concluded, "there was just one short period [near 16:42:12 UTC] when it could have been photographed by a KH-11 ... the conditions for photography at this encounter were at best marginal, and it would appear unlikely that much in the way of damage assessment could have been performed. If some were indeed performed, it must be regarded as a considerable achievement."

Kenden's pessimistic finding derives mainly from his guess that, "the resolution attainable by a KH-11 in spacecraft-to-spacecraft imaging would be 1 m per 200 km of range." Over the 20 years since Kenden's report, it has emerged that KH-11's resolution is about 15 times better than that.

Major re-evaluations are commonplace in the study of highly classified programs - many of my own assumptions, guesses and analyses have not stood the test of time.

In evaluating the potential for useful imaging, Kenden's assumption of a relatively low resolution constrained him to consider only the brief period near closest approach - when angular velocity was several times KH-11's known highest rate of 1.6 deg/s. "Whether the KH-11 could cope with such high angular rates, so far beyond its design levels, must be regarded as doubtful."

My analysis revealed good prospects for high resolution imaging a bit further from closest approach, without exceeding KH-11's known maximum angular velocity.

5. STS 2 Encounters with KH 11-3

Aviation Week and Space Technology reported on 2003 Feb 28 that a KH-11 had been employed during the second shuttle flight in 1981. I have found one close encounter between STS 2 and KH 11-3, which could have supported imaging at about 9 cm resolution, with good illumination, and angular velocity not exceeding a KH-11's known maximum of 1.6 deg/s relative terrestrial targets.

                  Range  Vang   ILL  Res
 Date     UTC      km    deg/s   %   cm
------  --------  -----  -----  ---  ---
Nov 14  13:43:08   283    1.60   21    9
Nov 14  13:43:43   283    1.60   71    9
Also in orbit was KH 11-4, which I could not evaluate, because its orbital elements, though not classified at the time, have since been classified. If anyone has official pre-classification elements or hobbyist elements (or observations), and is willing to share them, then I will analyse them.

6. Skylab Encounters with KH 8-38

Dwayne Allen Day has reported that "the NRO used a KH-8 reconnaissance satellite to image the Skylab".

Skylab sustained serious damage during its ascent into orbit on 1973 May 14, after lift-off at 17:30 UTC. One of its two solar arrays was torn off; the other failed to deploy - jammed by a piece of debris. A solar shield required to maintain a habitable temperature within the station was lost.

The launch of the first crew was delayed, first from May 15 to May 20, then to May 25, to allow time for the development and construction of a replacement solar shield, and to develop procedures to deploy it, and to release the jammed solar array.

KH 8-38 was launched on May 16 at 16:40 UTC, into the standard KH-8 orbit, 140 km x 400 km, inclined at 110.5 deg. KH-8s are believed to have achieved a resolution of about 10 cm at perigee. Their maximum angular velocity relative terrestrial targets was 3.2 deg/s. They carried two film-return re-entry vehicles.

To have provided data in time to be useful in repairing Skylab, KH 8-38 would have to have taken its images and returned them to Earth as soon as possible before the launch of the first crew, on May 25.

I have found a number of close encounters between Skylab and KH 8-38, the best of which, on May 18 and 19, could have supported imaging at better than 20 cm resolution, with good illumination, and angular velocity not exceeding a KH-8's known maximum of 3.2 deg/s relative terrestrial targets:

                  Range  Vang   ILL  Res
 Date     UTC      km    deg/s   %   cm
------  --------  -----  -----  ---  ---
May 18  02:00:35   236    3.15   64   17
May 18  02:00:51   236    3.14   20   17

May 18  02:46:54   197    3.06   17   14
May 18  02:47:13   200    2.96   80   14

May 19  13:50:45   261    3.23   48   19
May 19  13:50:54   260    3.25   28   19

May 19  14:36:52   230    3.03   18   16
The KH-8 must have become operational less than two days after launch to have taken images at the above times. To deliver them in time to be useful to the first crew, one of its two film-return re-entry vehicles must have been returned almost immediately.

[Home Page]


Contact: webmaster@satobs.org