I knew we would have aircraft losses throughout the war, but I wanted those losses to be the lowest possible until the ground war started. Then I expected the aircraft doing close air support would be taking greater risks, flying lower to identify the target or to get under the smoke and weather in order to support our engaged coalition ground forces.

An air commander manages loss rates. If you have time, and the target is not urgent (either in time or in priority), then you back off from great risk (on those occasions, for example, when the weather will make it difficult to visually acquire SAM missiles and outmaneuver them). On the other hand, if the situation is dire—if, say, a ground unit is being overrun—then you order the force into greater danger. At Khafji, we lost an AC-130 that stayed too long over the battlefield. Day came, and the sky grew light enough for an Iraqi shoulder-fired SAM gunner to visually acquire him and shoot his missile. Our pilot was not reckless. He elected to err on the unsafe side because there were lots of high-priority targets for him to shoot. It was a bad judgment, but that’s how airmen think and evaluate risk. If the target had been more ordinary and he had been facing the same risk, he’d have gone somewhere else and hit an easier target.

Because I absolutely had to knock out the Iraqi air defense command and control, I risked sending all the F-117s to Baghdad (I didn’t realize then how good Stealth was). However, if I had the same appreciation of Stealth that I had during the opening moments of the war, and the target was the Baath Party headquarters, I would not have sent the F-117s. That target wasn’t worth the gamble.

To me, the goals of the air superiority campaign were threefold:

(1) To render the Iraqi fighters inoperable. We would blind them by cutting off their command and control, terrorize them by shooting down anything that flew, and make life difficult by bombing airfields and radar sites.

(2) To render the radar-guided surface-to-air missiles impotent. We would attack command and control and use support jamming EA-6Bs and EF-111s, in order to force each piece of the system into autonomous operations (in that way they had to radiate with their own radars to find a target, making them more vulnerable to HARM attacks). We would terrorize the operators to induce them not to use their radars. We would kill them on a priority basis, using Wild Weasels and USN HARM-equipped aircraft. We would self-protect by using ECM pods, and by flying in VFR conditions, so aircraft could see the radar-guided SAMs and outmaneuver them.

(3) To render guns and shoulder-fired IR SAMs useless. We would fly at medium altitude; keep the time at low altitude (such as at the bottom of a dive bomb pass) at a minimum; and—the age-old lesson we always relearn in combat—we would not make multiple passes on the same target.