This letter responds to two articles by Stephen P. Aubin. The first, "How Newsweek Missed the Target," appeared May 19 on AFA's Web site. A fuller and somewhat different version, "Newsweek and the 14 Tanks," appeared in the July issue of Air Force Magazine.

Aubin has a problem. He hasn't seen the documents, so he doesn't know what he's talking about.

Responding to Newsweek's story of the suppressed Kosovo report, Aubin first parroted USAFE's initial line: no such study. The Air Force having abandoned that, Aubin, too, has to switch. The study not only exists, he now says, but was widely circulated. Both stories can't be true; and in fact neither is. Yes, the study exists. No, USAFE didn't circulate it. Aubin is wrong: The GAO wasn't given it.

Aubin's history is wrong, too. SACEUR sent the team into Kosovo precisely to check out Serb claims of minimal damage to the VJ [Yugoslav army]. USAFE wanted to survey fixed targets, but had zero interest in mobiles. After a SACEUR/USAFE tussle, the team split-some looking at fixed, some at mobiles. (Aubin quotes Lt. Col. David Duvall. Wrong man. Duvall ran the fixed targets group.)

The mobiles group anticipated briefing their report round NATO. It was killed. SACEUR's actual words to CINCUSAFE [Gen. John P.] Jumper were: "I can't go to Javier Solana and all those political leaders and people who have said we destroyed this [much equipment] and say 'Um, we made a mistake. You know, we just went around and took a hasty look on the ground and we didn't see a whole helluva lot.' "

Those are the facts. None of Aubin's huffing and puffing--"Not since CNN's Tailwind fiasco," etc.--changes them.

Defending the higher figures then confected by [Brig. Gen. John] Corley's team, Aubin again shifts ground. He claimed in his May 19 Web response that each pilot's mission report (misrep) of a kill "had to be corroborated by multiple sources." Corley certainly said that in his Sept. 16, 1999, SHAPE presentation. He even claimed: "Frankly, more than 85 percent of the time three or more sources were present." But, as Newsweek pointed out, that wasn't true. So Aubin now says the misrep itself counted as a source which only "had to be corroborated by at least one other source" to give the "multiple sources" Corley claimed.

That's a huge climb-down; but the new version isn't true either. Corley & Co. asserted flatly to Newsweek that "the misrep was a point of departure. We never used the mission report from the pilot as a source of validation" of a kill. And: "We call it an empty claim. ... To validate that claim we had to get something else ... multiple sources, two other validating sources." But in reality: "assessed hits based on multiple sources ... represent 45 percent of the total assessed hits." (No names, I'm afraid. The Air Force insisted the long session be on background. So much for Aubin's jibe about Newsweek's unnamed sources.)

Bottom line: Fewer than half the "validated" kills were backed by "multiple sources." Worse: Among the 55 percent backed by only a single datum point, just over four in 10 had as lone source a bomb flash picked up by IR sensors on the DSP satellite. Which confirms only that the pilot dropped a bomb; in most cases it says nothing about what, if anything, the bomb hit.

"Corley's team was conservative in its approach," says Aubin. Huh? Take artillery. NATO pilots claimed 857 hits on Serb artillery positions. The Joint Analysis Center (JAC)--NATO's scorekeeper--estimated that, at most, the pilots might have struck 341. But, their report says, the on-site team "did not consider 'artillery positions' because USAFE/IN could not confirm the position contained actual equipment." USAFE didn't know what had been down there. Yet Corley and his team "confirmed" that artillery pieces had actually been struck in 389 positions. How?

Or take Corley's claim of 93 confirmed tank kills. USAFE documents show that Corley's team actually managed to construct a case for 77 only. Then, in a final flurry, 16 strikes initially logged as multiple hits were reclassified as separate kills. That's "conservative"?

Aubin tries to rebut Newsweek's "accusation that the Air Force was flying too high" by pleading that laser-guided bombs work fine from 15,000 feet. But LGBs were only a tiny fraction of the munitions used against mobile targets. Take tanks: NATO pilots claimed 181 hits. USAFE considered 124 of these plausible; the JAC figured 110. Against these NATO had dispatched 956 munitions. Just 40 were PGMs: 27 laser-guided bombs and 13 Mavericks. The other 916 were inaccurate dumb bombs.

Altitude also bedeviled target ID, as the team found: "Many locations reported to have a tank or APC kill had numerous destroyed [military and civilian] vehicles. ... It is reasonable to assume a number of the military vehicles [we] counted, and even many civilian vehicles, were incorrectly identified ... as APCs or tanks."

Aubin's explanation for the missing equipment? The Serbs removed it. Please. That's been the Air Force line ever since Allied teams in World War II Normandy first catalogued massive discrepancies between pilot claims and kills found.

The Kosovo team visited virtually every site where NATO pilots had claimed a kill. Not merely did they find few bombed-out hulks; though they scoured the sites and craters, they found no debris either. That the Serbs might remove damaged vehicles is plausible. That they would vacuum every crater is not.

The VJ had no heavy lifting gear in Kosovo. Their only option would have been to drag damaged tanks to the nearest road. The team looked for drag marks, but reported "no evidence of equipment removal such as tracks, HET [heavy equipment transporter] marks, or the presence of V-bar equipped tanks used to tow a disabled vehicle." Corley misrepresented this at his Sept. 16 presentation: "The team further discovered that equipment had been towed out of bomb-damaged revetments to the main road and transported away. The ground earth scarring is clearly evident in multiple examples." A distortion that adds weight to the charge of suppression.

NATO intel analysts reckon they monitored perhaps 90 percent of the Serb withdrawal. Analysts have pored over the imagery. They've identified a few damaged vehicles, but nothing on the scale posited by USAFE.

Aubin's final assertion is that since the "combined effects" of military and other actions brought victory, "the number of tanks destroyed" is irrelevant as a metric. Rubbish. If airpower is poor at finding and destroying scattered mobile targets in difficult terrain covered by multiple air defenses-the challenge in Kosovo-let us acknowledge that and either improve Air Force capabilities or resolve not to fight such battles again. Aubin does nobody a service, least of all the Air Force, by trying to fudge the problem.

John Barry
Newsweek National Security Correspondent
Washington, D.C.

From Stephen P. Aubin:

Newsweek's John Barry and Evan Thomas claim that NATO aircraft, during the 78 days of Operation Allied Force, struck a mere 14 tanks, 18 APCs, and 20 artillery pieces. That is the crux of "The Kosovo Cover-Up" (Newsweek, May 15), but it is untrue. They were wrong when they reported it then, and they are wrong now. NATO aircraft struck 93 tanks, 153 APCs, and 389 artillery pieces. At least.

Barry and Thomas based their claims largely on what they termed a "suppressed" NATO report. The claims, in essence, were three: NATO airpower didn't hit much. NATO covered up that fact. And NATO invented higher numbers. Barry's letter repeats all three claims. They are false.

I will take each in turn, but I'll first deal with a somewhat minor Barry claim-that I initially denied the existence of a NATO report. I didn't, as is plain from the text. I denied the existence of a "suppressed" NATO report, and still do.

1. Newsweek asserts NATO airpower didn't hit much. The claim rests largely on the so-called "suppressed report" containing low figures. Barry doesn't tell you the true nature of the document. It was a working draft report prepared in July 1999 by SHAPE's Munitions Effectiveness Assessment Team (MEAT). It presents results of a postwar Kosovo ground survey--a snapshot of a cold battle area, nothing more. It makes no pretense to being the last word on the war. It will never yield the whole picture. For that, one must go to the final SHAPE report, NATO's Kosovo Strike Assessment, which Newsweek essentially ignored. More on that below.

What Barry has in his possession is a document that lists only the number of vehicle hulks found in Kosovo at least one and as many as three months after the strikes took place. Newsweek's claim that the ground survey represents the totality of NATO's successes is, on its face, ludicrous.

2. Whatever Barry claims, nobody "killed" any study. The "mobiles" part of the MEAT draft report is still very much alive. It and the final report are archived in Europe and Washington and at Air Force Historical Research Agency, Maxwell AFB, Ala. They are available to anyone who has a proper security clearance.

Moreover, the draft report has been widely circulated. It was provided to the US Army, Center for Naval Analyses, Office of the Secretary of Defense, and General Accounting Office.

In advancing his "suppression" claim, Barry suggests the existence of factions, one focused on fixed targets, the other on mobiles. In reality, there was one Munitions Effectiveness Assessment Team. The leader of the Kosovo Strike Assessment, Brig. Gen. John Corley, USAF, said the same team member names are found on both fixed-target and mobile-target working drafts.

Lt. Col. Michael (not "David," as Barry says) Duvall was deputy leader for the entire team, not just for fixed targets. He noted to me that team members were divided each day into "fixed" and "mobile" survey groups. Team members were interchangeable and came from all over, not just from USAFE, as Barry suggests.

3. Newsweek evidently believes NATO conspired to produce inflated strike data, arguing that NATO airpower did not kill 93 tanks, 153 APCs, and 389 artillery pieces.

What is the source of these figures? It is NATO's Kosovo Strike Assessment, the fruit of a nine-week-long, round-the-clock effort by 200 personnel. Its sources of information included not only the MEAT draft but also national satellite images, cockpit video, UAV video, and other intelligence. Data were correlated to establish what happened. For some reason, Barry simply refuses to accept use of such sources to confirm or disprove strike claims.

It is true that 55 percent of NATO's validated "successes" are based on a pilot's mission report and one additional source. Barry implies they are weak cases. However, these strikes make up what Corley calls the "definitive" category; the second source was strong enough to erase all doubt.

Each remaining NATO "success" (45 percent of total) also began with a pilot mission report. However, validation required at least two more sources. This caused confusion. Corley, in his September 1999 SHAPE briefing, did say three or more sources were available "85 percent of the time." I asked Corley about this discrepancy and, as it turns out, the 85 percent remark refers only to the 45 percent requiring two or more additional sources. Corley concedes that his statement was not very clear.

As Barry says, Corley's team did validate 77 tank strikes. However, these 77 were in addition to the 26 hulks of the MEAT draft report. It turned out, however, that 10 tanks were double counted. Basic arithmetic--add 26 and 77, subtract 10--yields the figure of 93 tanks. There was no "final flurry" to add 16 fraudulent tank kills, as Barry claims.

Barry also confuses readers about NATO's use of dumb bombs and precision munitions. When a Serb vehicle or vehicles (tanks, for instance) were in the open and risk of collateral damage was low, NATO might use a profusion of dumb bombs. NATO tended to use PGMs to hit single vehicles hidden near civilians. USAFE credits 81.7 percent of tank kills to PGMs, the rest to dumb bombs. The fact is, though, that either type can be "accurate," even from 15,000 feet. It depends on the nature and location of the target.

Barry scoffs at the idea that the Serbs "cleaned" the battlefield and greatly reduced the number of vehicle carcasses left in view. Yet Corley, in his SHAPE briefing, showed actual video of Serb transport vehicles hauling out APCs and other equipment covered by tarps. Barry's "NATO intelligence analysts," who are said to have "monitored perhaps 90 percent of the Serb withdrawal," were monitoring only Serb equipment still in Kosovo at the end of the 78-day campaign. By definition, they didn't see what was already gone. The Serbs had ample opportunity to move equipment during gaps in NATO surveillance.

There was no "Kosovo Cover-Up." Barry and Thomas were used by individuals whose desire to discredit airpower is obvious.