On Sept. 11, 1990, a Soviet Resource satellite swept across the southern border of occupied Kuwait headed north. Its camera clicked seven times between Kuwait's southern border with Saudi Arabia and the northern border with Iraq. Each foot-square photograph recorded scenes below with a sharpness never seen before in images from a commercial satellite, often picking up objects as small as 16 feet square. In isolated instances, resolution might have been as great as 10 feet _ an encroachment on the domain of the spy satellite.
In its brief passage over Kuwait last September, the Resource produced an enigma of the Persian Gulf war: No Iraqi deployments could be detected in the Soviet pictures of Kuwait.
The Pentagon was claiming at the time that more than 250,000 Iraqi soldiers, accompanied by more than 2,000 armored vehicles, were in Kuwait. If that were true, they should have been unmistakable at resolutions as good as 10 to 16 feet.
Readers of the St. Petersburg Times saw these pictures in their Jan. 6, 1991, Sunday paper. In a story accompanying the photographs, Times Washington correspondent Jean Heller posed the question: Where were the Iraqis?
A year has passed since the pictures were taken. It is time to raise the question again. And this time, there is at least a partial answer:
The Bush administration either was deceived successfully by the Iraqis about their troop and armor strength in Kuwait in September 1990; or misread the intelligence tea leaves; or exaggerated that strength by as much as a factor of four for reasons that have not become clear; or the explanation could be a combination of all three.
But at the same time, it seems clear now that the Iraqi force in Kuwait _ whatever the numbers _ probably did pose a threat to Saudi Arabia and American interests in that country.
This is what could be concluded from the Soviet satellite pictures:
The Iraqi army hadn't sent its tanks sweeping across the Kuwaiti desert; the signs of movement of such a force would have been easy to see in the Soviet photos. Even in satellite pictures with resolutions of 33 feet _ many times poorer than those the Soviets provided _ analysts could read the story of armored battles of the Iran-Iraq war in tracks left in the sand.
The Iraqis had not established large base camps in Kuwait. Camps for 250,000 to 300,000 troops would have meant large numbers of tents and support services that would have been spotted easily from space. And what would have passed for streets _ chewed up by the passage of military vehicles _ would have stood out clearly against virgin desert had they existed.
There were Iraqi forces in Kuwait in early September 1990, one month after the invasion; that is not subject to question. The problem is determining their number, their locations and their intentions, given the lack of information in the Soviet pictures.
Triangular fortresses that analysts expected the Iraqi army to erect were nowhere to be seen in the September 1990 photographs. The high earthen walls, hundreds of feet on a side, are formed by bulldozers scraping and packing sand. They were the trademark of the Iraqi defenses and were easy to locate in satellite pictures of the Iran-Iraq war.
Former military photoanalyst William Kennedy saw the Soviet pictures shortly after they came to the United States. He said he remains convinced that they show no evidence of Iraqi military activity.
If the Iraqi army had begun using large bases, Kennedy said, they would have traveled on the major highways, the easy routes for moving troops and supplies to potential defense lines in southern Kuwait.
But as the pictures that appeared in the St. Petersburg Times showed, many of those roads were blocked with blown sand, making the supply of a large force far more complicated than if the roads had been maintained.
More recent satellite photos show a great deal more.
Iraqi trench lines and strong points are clearly visible in U.S. Landsat pictures taken after the war ended in March 1991 (see graphic on this page).
One such trench stretches from a burning oil well for more than 5 miles across the desert. Crossing points, newly bulldozed strong points, artillery or tank emplacements and a lacy network of roads behind the trench all are easy to see.
Landsat has a resolution of just 100 feet, five to 10 times worse than the lenses of the Soviet cameras.
If installations anything like the ones in the Landsat picture had been in place in September 1990, they would have shown up in the images from the Resource satellite.
The Soviet photos were, in retrospect, the first clue that U.S. intelligence overestimated Iraq's numbers, a fact the Defense Department has confirmed, at least in part.
Estimating the enemy's strength is not an easy job. American analysts identify units in the theater by listening to their radio communications, noting who talks to whom _ and about what. Call signs, types of radio traffic and direction-finding all play roles in assessing the size of the opposition.
But signal intelligence can't do the whole job.
A radio intercept seems to tell U.S. intelligence that the Iraqi Medina Luminous Division is in the field. But the traffic alone doesn't establish whether the Medina Division is at full strength or represented by a corporal's guard of communications technicians sending and receiving their messages from a script written in Baghdad expressly to mislead the Americans.
It is considered prudent for an intelligence analyst to assume that all units are at full strength unless there is proof that they are not.
Military deception has a long and honored history, dating from the Old Testament to the present.
"In wartime," Winston Churchill remarked, "truth is so precious that she should always be attended by a bodyguard of lies."
Signal intelligence can be subverted when the enemy knows it is the cornerstone of our intelligence, and when he is shrewd enough to turn our skills against us.
After the war ended, the Pentagon conceded that its knowledge of true Iraqi troop strength was incomplete, and further, that it probably overestimated the number of Iraqis in Kuwait when the ground war began.
What does all this say about the September photographs?
Kennedy, the former military photoanalyst, said the absence of Iraqi fortifications in Kuwait in September indicates that "if the Iraqis were in country, they were probably in some kind of offensive posture."
Others have come to the same conclusion.
One brief passage of Bob Woodward's book The Commanders reports that CIA photos showed "Iraqi tanks dug into the desert with sand embankments on the front and two sides of each tank." This is a temporary deployment used for offensive purposes.
Individual tanks deployed at intervals inside small excavations would not have shown up in the pictures published by the Times.
There were, Woodward wrote, just 70,000 troops within striking distance of the Saudi border, a number well below administration estimates but sizeable enough to pose a threat to the Saudi frontier.
When an army places its tanks in large fortresses, as the Iraqis did later in the war, it is signaling that it intends to stay where it is and repel any assault.
The temporary nature of the Iraqi deployments in late summer of 1990 sent the other signal: Their troops were waiting for the order for the tanks to back out of their dugouts, scoot around the sides and roll forward in an assault on Saudi Arabia.
Peter Zimmerman is an expert in satellite imagery. He served as an adviser to the U.S. delegation to the SALT treaty talks and to the U.S. Arms Control and Disarmament Agency. He is a nuclear physicist, currently serving as a research professor of engineering at George Washington University.