PROBLEMS WITH FLEXIBLE TARGETING

Yet another disappointment in the air war’s performance centered on what turned out to be NATO’s almost completely ineffective efforts to attack mobile VJ forces in the KEZ. By the end of the third week, despite determined attempts by allied aircrews over the preceding week, NATO analysts were unable to confirm the destruction of a single VJ tank or military vehicle, owing to the success of enemy ground units at dispersing and concealing their armor. That disappointment underscored the limits of conducting air operations against dispersed and hidden enemy troops in conditions in which weather, terrain, and tactics all favored the enemy and where no friendly ground combat presence was on hand to compel those forces to concentrate and expose themselves. Had Serb commanders any reason to fear a NATO ground invasion, they would have had little alternative but to position their tanks to cut off roads and other avenues of attack, thus making their forces more easily targetable by NATO air power. Instead, having dispersed and hidden their tanks and armored personnel carriers (APCs), Serb army and paramilitary units were free to go in with just 20 or more troops in a single vehicle to terrorize a village in connection with their ethnic cleansing campaign.

Indeed, the opportunity to get at fielded enemy ground units with air power alone had been essentially lost by NATO even before Operation Allied Force commenced. As General Jumper later recalled, during the Rambouillet talks in early March 1999, “we watched 40,000 Serbian troops mass north of Kosovo, we watched them infiltrate down into Kosovo, we watched heavy armor come down into there, all under the umbrella of the peace conference, and we weren’t able to react.”[268]Once those forces had completed their massing on March 15 and had begun a substantial incursion into Kosovo, any chance for allied air power to be significantly effective against them promptly disappeared. Once safely dispersed, VJ units simply turned off the engines of their tanks and other vehicles to save fuel, hid their vehicles in barns, churches, forests, and populated areas, hunkered down, and hoped to wait the air effort out. By the end of April, General Clark frankly conceded that after six weeks of bombing, there were more VJ, MUP, and Serb paramilitary forces in Kosovo than there had been when Allied Force began. That attested powerfully to the latter’s near‑total ineffectiveness, at least up to that point, in halting the Serbian ethnic cleansing rampage throughout Kosovo.

Once the targeting of enemy troops in Kosovo became a SACEUR priority at the start of the third week, Yugoslavia was divided into four large search sectors. The two USAF E‑8 Joint STARS aircraft that had been committed to supporting the air effort were tasked with searching for ground targets in the KEZ and with providing near‑real time intelligence and targeting information to the CAOC in Vicenza and to the EC‑130 ABCCC. Depending on the possibility of collateral damage, Joint STARS was sometimes cleared to communicate directly to airborne FACs and to direct NATO strikes against fleeting targets of opportunity, with the goal of getting target information and coordinates to orbiting strike aircraft within minutes.[269]

Before long, three broad approaches to what came to be called “flex” targeting emerged for prompt employment against mobile VJ and MUP forces operating in Kosovo and against pop‑up IADS assets deployed in Serbia. In the first, called “alert flex” targeting, combat aircraft were apportioned from the very outset as designated “flex” sorties in the ATO and reserved for launch on short notice against any pop‑up targets that might be detected and identified within the ATO cycle. Initially, such designated aircraft were kept on ground alert. Later in the operation, they were placed on airborne tanker alert, which reduced their response times by as much as two hours.

The second approach entailed redirecting aircraft already en route to preplanned fixed targets. Strikers would be diverted either to alternate high‑value fixed targets in Serbia or to recently detected mobile targets in Serbia or Kosovo. Because of the large number of NATO fighters already preapportioned and available on call for use as alert flex assets, however, such en route diversions occurred only rarely. All three heavy bombers (the B‑52s, B‑1s, and B‑2s) were also diverted to new targets on occasion, requiring real‑time changes in their preplanned ingress routes.

The third category of flexible targeting involved dedicated sorties launched into holding orbits for on‑call attacks against detected mobile VJ forces in Kosovo after the KEZ was declared on Day 20 of the air effort. This approach, which evolved progressively over time, entailed the use of F‑16s, A‑10s, or Tornados serving as airborne forward air controllers. Their FAC‑qualified pilots would search for ground targets in predesignated kill boxes, attempt a visual identification of any suspected target candidates, and assess the potential for collateral damage after determining that the target candidates were valid. Depending on the prevailing rules of engagement, the FAC pilots would first request ABCCC or CAOC approval to attack the target and then, upon being cleared to release weapons, would drop their munitions on the approved target while directing their wing‑men to drop on adjacent targets. In the event that multiple targets were detected and approved, additional strike aircraft would be called in if they were close at hand. Because NATO had no fielded ground forces in the combat zone, the FACs could not request ground assistance and were on their own in locating and identifying mobile targets.

As noted earlier, a major problem that inhibited the effectiveness of Joint STARS in support of these missions was Kosovo’s mountainous terrain, which required the aircraft to fly unusually close to enemy territory so its sensor operators could look into valleys and minimize the enemy’s opportunities to take advantage of terrain masking. Even then, the high ridgelines often made it impossible for Joint STARS crews, from their standoff orbits, to peer into some valleys where VJ forces were thought to have been concentrated. Joint STARS also had only a limited ability to detect and monitor ground targets in dense woods and built‑up areas. Because of these constraints, NATO had little by way of wide‑area airborne surveillance and cueing of the sort that had made coalition operations against enemy ground forces so effective in Desert Storm. That deficiency placed a doubly high premium on hitting enemy ground‑force targets as they moved into open areas and were visually detected by airborne FACs. It also, in effect, ceded the tactical initiative to VJ forces, since the latter could decide when and where to reposition themselves. The net result was a need for large numbers of combat aircraft continuously orbiting over the KEZ but producing little tactical return, compounded–indeed, largely caused–by the absence of a NATO ground threat to force enemy troops into more predictable patterns of behavior.[270]

The performance of Joint STARS against dispersed and hidden enemy forces was less than satisfactory not only because of the constraints described above, but also because of an unfortunate failure by air operations managers to make the most of the aircraft’s inherent capabilities for supporting counterland operations. That failure partly reflected a continuing slowness on the part of the U.S. Air Force to develop and institutionalize a detailed appreciation for how land forces operate and, in turn, to acquire the conceptual wherewithal that is essential for making air power more effective in defeating those forces. Surprisingly little progress was registered by the Air Force over the nine years since Desert Storm in developing a concept of operations for using Joint STARS in a surveillance and control team that also includes AWACS, Rivet Joint, airborne FACs, and UAVs, all working as a synergistic collective against elusive enemy ground forces.

As one telling testament to this failure, the inclusion of Joint STARS in the air war’s equipment roster had been requested by the Army, not by the Air Force.[271]Because of the predominant USAF focus on attacking fixed infrastructure targets, few in the Air Force fully appreciated the E‑8’s capability for providing wide‑area, all‑weather standoff coverage of the KEZ and its resultant ability to provide USEUCOM’s and NATO’s operational‑level commanders with real‑time situation awareness regarding the status and activity of VJ forces. It took days for Joint STARS even to be included in the ATO. Once there, the aircraft was typically thought of as a surveillance platform operating in the service of the intelligence community, rather than as a strike support asset working to provide direct and immediate assistance to NATO aircrews conducting flexible targeting missions. With the right teaming, connectivity, and practice, the use of Joint STARS to cue UAVs might have reduced, if not eliminated, the “searching‑through‑a‑soda straw” problem, lessened UAV exposure to hostile fire, and helped maintain tactical surprise for NATO aircrews engaged in the search for VJ targets of opportunity. No measures of that sort, however, were attempted until quite late in Allied Force.

Yet another complicating influence on the air effort’s attempts against dispersed and hidden enemy forces stemmed from the command and control arrangements that had been hastily cobbled together at the operational and tactical levels once it became clear that NATO was committed to an air war for the long haul. Although the CAOC eventually worked out a means of using real‑time imagery to detect fielded VJ forces in the KEZ and to “flex” allied air assets to attack those newly developed targets in an orderly fashion, those doing the “flex” decisionmaking during the first half of Allied Force did so with no apportionment or targeting guidance whatever. As one expert observer noted, “if the detected target was militarily significant, it was struck, regardless of [General Short’s] priorities or intentions. There was no link to an assessment mechanism, so that once a target was struck, there was no way to link it to what unit it had been associated with, so no effective degradation was recorded.”[272]As a result, combat aircraft were sometimes diverted from scheduled ATO targets of clear operational significance to attack “flex” targets of highly dubious tactical, let alone operational or strategic, worth. Moreover, owing to the absence of any feedback mechanism, aircraft were often committed against targets that had already been successfully struck, forcing the CAOC either to re‑role aircraft on short notice or else to expose aircrews needlessly to enemy IADS threats a second time. For most of the air war, roughly half of General Short’s available surface‑attack sorties were committed against targets in the KEZ. Of those, a significant percentage were “flexed” in this haphazard manner.[273]

Weather was still another complicating factor in the effort against dispersed VJ forces. From the 15,000‑ft altitude floor above which NATO aircrews typically operated, the cloud cover over Kosovo was greater than 50 percent for more than 78 percent of the air war’s duration. That allowed unimpeded strike operations on only 24 of the air war’s 78 days. The impact of these conditions on the flexible targeting effort was considerable. In all, 3,766 planned sorties, including 1,029 designated close air support sorties, had to be canceled because of weather.

Even on clear days, another factor preventing the kill box system from being as effective as it might otherwise have been was the tight rules‑of‑engagement regime that had been imposed after the Djakovica incident (see below), in which more than 60 ethnic Albanian refugees were reportedly killed in an attack by USAF F‑16s against what was thought to have been a VJ troop convoy. These restrictions had a far greater inhibiting influence on the effectiveness of NATO’s flexible targeting efforts than the oft‑cited 15,000‑ft altitude floor which NATO’s aircrews had been directed to observe. Unless an object of interest was clearly determined to be a valid military target, such as a VJ tank operating in the open, pilots had to get clearance for any attack from the CAOC, with General Short himself often making the decision after checking second sources like real‑time UAV video feed. Because of the delays created by these and similar hurdles, orbiting NATO aircraft often ran low on fuel before being cleared to drop their weapons and accordingly were forced to leave the area in search of a tanker.[274]

Last, and perhaps as decisive as any single other factor, VJ forces aggressively avoided making themselves easy targets for NATO air attacks. Indeed, digging in and hunkering down for defensive attrition warfare had lain at the heart of Yugoslav operational doctrine ever since the days of partisan operations against the Wehrmacht in World War II. Whenever General Clark would say, “You’ve got to get them in their assembly areas,” the reply typically was: “These guys aren’t assembling!”[275]RAF Harrier GR. Mk 7 pilots operating in kill boxes over Kosovo reported that “there was nothing moving around at all during the daytime,” adding that when Clark “got up and said knocking out five tanks was a good day for NATO, he [was] telling it straight. On some days we couldn’t find any tanks.”[276]Even with the aid of binoculars, the ground below often seemed devoid of life to NATO aircrews orbiting overhead at 15,000 ft. This was the predictable result of trying to engage an enemy who had no need to shoot, move, or expose his position, thanks to the absence of a credible NATO ground threat.

To be sure, there were some notable bright spots in NATO’s air effort against VJ forces in Kosovo. To cite one example, in those rare instances in which enemy armor and other targets exposed themselves to attack from the air, the upgraded AGM‑65G2 Maverick air‑to‑ground missile generally performed very effectively. The effectiveness rate for older Mavericks was lower, but still reportedly higher than 90 percent.[277]Also, both U‑2 imagery and pictures provided by the Navy’s F‑14 equipped with TARPS (Tactical Air Reconnaissance Pod System) later proved useful to the CAOC in what the Cohen‑Shelton after‑action report to Congress called “several” instances involving the rapid retargeting of NATO aircraft to new targets.[278]

To cite another notable example, the two Marine F/A‑18D squadrons that deployed to the former Warsaw Pact airfield at Taszar, Hungary, late in the air war played an active part in the effort against enemy forces in the KEZ.[279]For the first time on a large scale in combat, the F/A‑18D aircrews, along with NATO pilots flying other combat aircraft types, made heavy use of night‑vision goggles with compatible internal and external lighting modifications, thus enabling multi‑aircraft formations and simultaneous night bomb deliveries.[280]Some F/A‑18Ds also carried the internally mounted Advanced Tactical Aerial Reconnaissance System (ATARS). Still in operational evaluation as Allied Force began, the system provided digital, multispectral target images with its SAR and medium‑altitude electro‑optical (EO) imagery as a backup to pictures from other ISR sources, with a real‑time connection to ground receiver stations. It figured prominently in both targeting and BDA activities.[281]

In a typical night F/A‑18D flexible targeting mission (which might last as long as six hours, with four inflight refuelings), the C‑130 ABCCC would pass to orbiting Marine fighters the grid coordinates of a VJ artillery position detected by the TPQ‑36 and TPQ‑37 counterbattery radars attached to the U.S. Army’s Task Force Hawk in Albania. An airborne FAC in an OA‑10 would then illuminate the target location with flares and call in a two‑plane section of F/A‑18Ds to be available on short notice to attack it. In so doing, the OA‑10 FAC, in effect, performed reconnaissance by fire. When shot at in return, the FAC would determine the source of fire to be hostile, and the F/A‑18Ds would then be cleared to drop 500‑lb Mk 82 bombs on it, which would generally stop the artillery fire for the rest of the night.[282]It was said that the greatest frustration for all NATO aircrews flying combat missions was to be orbiting over the KEZ night after night, for as long as six hours interspersed with multiple inflight refuelings, only to be called in at long last by an airborne FAC and cleared to attack a reported VJ tank that was no longer there.[283]

Owing in large part to such operations, at least those that produced recognizable combat results, NATO’s effort to engage dispersed and hidden enemy forces in the KEZ was not a complete waste of time and assets. For one thing, VJ commanders knew all too well that as the weather began steadily improving with the onset of summer, any effort on their part to conduct large‑scale operations against either the KLA or civilian ethnic Albanians would put them at extremely high risk of being attacked. Moreover, General Short reported in late May that the newly focused attacks against the VJ’s 3rd Army in Kosovo were beginning to register discernible effects. He went on to predict that “if we do this for two more months, we will either kill this army in Kosovo or send it on the run.”[284]

Taken as a whole, however, NATO’s effort to attack enemy ground units in the KEZ was essentially a failure, the full extent of which became apparent only after the air war was over. To the very end, Short doubted that focusing exclusively, or even primarily, on elusive VJ forces in Kosovo would be enough to swing the desired outcome. He also placed little stock in claims emanating from NATO headquarters that the VJ was being progressively weakened by the air attacks. On that latter point, he observed that the only things that mattered were that army’s ability to move and its willingness to fight, and that both of those remained decidedly intact.[285]

In the first detailed official rundown of the air war’s accomplishments as Allied Force approached its midpoint, the limited effects of NATO’s bombing attempts against enemy forces in Kosovo were underscored by the frank admission that the VJ still retained 80 to 90 percent of its tanks.[286]Later, on May 19, NATO spokesman Major General Walter Jertz claimed more optimistically that one‑third of all VJ tanks and artillery in Kosovo had been destroyed.[287]As the bombing effort drew to a close, NATO was similarly claiming that it had taken out more than one‑quarter of the VJ’s tanks and APCs deployed in Kosovo. Britain’s chief of the defense staff, General Sir Charles Guthrie, further reported that more than 30 percent of the VJ’s artillery and mortar pieces had been destroyed by NATO attackers.[288]

In its final tally as Operation Allied Force ended, the U.S. Defense Department settled on 700 out of 1,500 tanks, APCs, and artillery pieces destroyed altogether in Kosovo.[289]More specifically, General Shelton announced in an early postwar briefing that NATO attacks had destroyed “around 120 tanks, about 220 armored personnel carriers, and up to 450 artillery and mortar pieces.” However, nothing like a matching number of hulks was found by allied inspectors after Allied Force ended. During their withdrawal, VJ troops took hundreds of tanks, artillery pieces, and APCs out of Kosovo. They also seemed spirited and defiant rather than beaten.[290]The VJ’s commander in chief, General Dragoljub Ojdanic, claimed after the war that only 524 Yugoslav soldiers had been killed, in marked contrast to NATO’s estimate of thousands.[291]

After the dust settled in early June, a preliminary NATO postmortem concluded that the air war had had almost no effect on VJ operations in Kosovo. In an after‑action briefing to senior Pentagon officials, the commander of Joint Task Force Noble Anvil, Admiral James Ellis, confirmed that NATO air operations were effective against VJ armor only after the KLA launched its offensive, forcing defending VJ troops to uncover and mass their armor and mechanized forces.[292]NATO initially claimed after the air war ended that it had disabled 150 of the estimated 400 VJ tanks in Kosovo. General Clark later scaled back that number to 110, after having determined that many tanks assumed to have been destroyed had, in fact, been decoys that the VJ had skillfully fielded in large numbers.[293]

Not only did the Serbs make successful use of tank decoys made out of tetra‑pak milk carton material, they also positioned wood‑burning stoves with their chimneys angled to make them look like artillery pieces. In some cases, water receptacles were found in the decoys, cleverly placed there to heat up under the sun to help replicate the infrared signature of a vehicle or hot artillery tube.[294]One source spoke of cockpit display videotapes showing targets with every appearance of being tanks collapsing instantly upon being hit. In addition, the Serbs made heavy and frequently effective use of smoke generators to protect targets against LGBs. After the air war ended, site‑survey teams that went in on the ground in Kosovo and interviewed witnesses discovered that VJ forces had buried many of their missile launchers, covered fuel trucks with rugs, and disguised tanks as haystacks and armored vehicles as trees.

The subsequent, and putatively definitive, after‑action report on Allied Force submitted to Congress by Secretary Cohen and General Shelton in the summer of 1999 claimed valid strikes on 93 enemy tanks, 153 APCs, 339 other military vehicles, and 389 artillery and mortar pieces.[295]Those downwardly revised estimates came on the heels of the findings by a munitions effectiveness assessment (MEA) team of 67 operators and intelligence experts, made up mostly of USAF officers, who went into Kosovo at Clark’s behest to comb the country, both by helicopter and on foot, in an on‑site survey of all actual DMPIs attacked. The team’s specific mission was to perform an assessment of attacks undertaken against mobile targets in the Presevo Valley region of Kosovo by cross‑referencing on‑scene observations and conversations with witnesses on the ground against available cockpit display videotapes, imagery intelligence, signals intelligence, human intelligence, and interviews with airborne FACs who had been operating near the target area at the time of the attacks.[296]

The team’s initial conclusion from that assessment was that “only a handful” of enemy tanks, APCs, and artillery pieces could be determined to have been catastrophically damaged by air attacks.[297]Although the team succeeded in investigating some 60 percent of NATO’s claimed hits on mobile targets in the KEZ, it confirmed only 14 tanks, 18 APCs, and 20 artillery pieces as destroyed for sure. A later assessment conducted by USAFE’s office of studies and analysis, using the team’s findings as one important input, reported 93 tanks and 153 APCs as having been struck altogether, the same numbers noted above that were cited later by Secretary Cohen and General Shelton. Many of those claimed hits, however, were validated by only a single source of evidence, such as a cockpit display videotape or an infrared event detected by DSP satellites.[298]In the later aftermath of Allied Force, on‑site surveys of bomb damage effects by KFOR observers and other inspectors further confirmed that NATO’s attacks against VJ forces had accomplished far less than had initially been assumed, notably including at Mount Pastrik.[299]

These seeming discrepancies led some air war critics to charge that NATO and the U.S. Defense Department were engaging in a blatant cover‑up of allied air power’s poor performance against VJ forces in Kosovo to avoid being embarrassed by the paltry numbers the inspection team had produced. That criticism turned out, however, to have been overblown for two reasons. First, the cover‑up charge was misdirected, in that it was based entirely on a leaked draft report by USAFE’s inspection team that went to Kosovo earlier in the summer of 1999. That draft report, dated August 3, 1999, and titled “Operation Allied Force: Munitions Effectiveness Assessment, Vol. II: Mobile Targets,” documented information collected in Kosovo and elsewhere by the MEA working group tasked with looking into mobile enemy targets. That effort was undertaken not to account for successful strikes, but rather to determine what equipment remained at the attacked sites. The freshest of the attacked sites visited was four weeks old, and some were only visited for the first time three months after the attacks.

All told, the USAFE team came across 14 tank carcasses and the hulks of 12 self‑propelled artillery vehicles, which could have looked like tanks from the air and been reported as such in post‑strike pilot mission reports. That added up to 26 confirmable “tanks” suffering sufficiently catastrophic damage from NATO air attacks to be written off and abandoned by departing VJ forces. Cross‑referencing pilot reports with corroborating evidence from other sources, the USAFE studies and analysis staff later documented presumed successful strikes on 93 tanks, 153 APCs, and 389 artillery pieces. It further documented another 60 instances of attacks on tanks that were believed to have been successful but that could not be validated because of the stringent criteria it had been given by SACEUR. As explained in SACEUR’s subsequent strike assessment briefing at NATO headquarters, 26 tanks could be categorized as “confirmed catastrophic kills,” based on physical information actually gathered on the ground in Kosovo. The remainder of the 93 reported tank kills were categorized as “assessed strikes,” which meant, in effect, that there were indications suggesting that a weapon may have hit a valid target.[300]

Air warfare professionals, notably including the USAF chief of staff, General Michael Ryan, have readily acknowledged since the end of Allied Force that the problems encountered by the operation’s flexible targeting effort outlined above reflected real challenges for the effective application of air power posed by such impediments as trees, mountains, poor weather, and an enemy ground force permitted the luxury of dispersing and hiding rather than concentrating to maneuver to accomplish its mission.[301]The Cohen‑Shelton report to Congress frankly admitted that the problems encountered with flexible targeting of VJ forces in Kosovo pointed up continued shortfalls in the nation’s ability to meet “the difficult challenge of rapidly targeting enemy forces and systems that can move and hide frequently.”[302]On that discomfiting point, U.S. and NATO defense officials had nothing whatever to hide and covered nothing up.

Second, and perhaps more important, although it was clearly essential for NATO to maintain constant pressure on VJ and MUP forces deployed in Kosovo and to bend every reasonable effort to suppress their freedom to operate at will against the ethnic Albanians, the majority of the combat sorties that SACEUR insisted be devoted to finding and attacking enemy forces in the KEZ arguably entailed a waste of munitions and other valuable assets. That perspective was pithily expressed by the vice chairman of the Joint Chiefs of Staff, USAF General Joseph Ralston, who later went on to replace Clark as SACEUR: “The tank, which was an irrelevant item in the context of ethnic cleansing, became the symbol for Serb ground forces. How many tanks did you kill today? All of a sudden, this became the measure of merit that had nothing to do with reality.”[303]When General Jumper, on being pressed later by reporters for an honest account of how many tanks NATO had actually destroyed, replied simply “enough,” he was telling the truth. The marginality of the tank issue to what really mattered in Allied Force was perhaps most convincingly explained by Brigadier General Daniel Leaf, commander of the 31st Air Expeditionary Wing at Aviano, when he declared in the immediate wake of the cease‑fire that “counting tanks is irrelevant. The fact is they withdrew, and while they took tanks with them, they returned to a country whose military infrastructure has been ruined. They’re not going to be doing anything with those forces for a long time.”[304]

True enough, a demonstrable record of effective performance by the attacks against VJ tanks may well have been regarded at the time as being of crucial importance toward vindicating SACEUR’s stress on attacks against dispersed and hidden enemy forces in Kosovo. Yet viewed in hindsight, the number of tanks taken out in the air war was, and remains, an issue of only scant pertinence to the operation’s ultimate outcome. Not only that, getting into the tank‑counting business in the first place made for a largely self‑inflicted wound by the Department of Defense, SACEUR, and NATO. In the end, all the to‑ing and fro‑ing over how many enemy tanks were taken out by NATO was mainly of academic interest, since air operations in the KEZ were, by all indications, not a determining factor affecting Milosevic’s ultimate decision to capitulate.[305]The KLA had been eliminated entirely as a tactical consideration by superior VJ strength. Moreover, notwithstanding more than two months of continual NATO bombing, the VJ lost few personnel to hostile fire, retained its command and control and resupply apparatus throughout the air effort, and continued to conduct ethnic cleansing forays until the last day of the air war, even though it did put itself at risk whenever its units exposed themselves to attack from the air. At bottom, NATO’s failure to perform better than it did against enemy ground units in the KEZ was as much a result of the strategy chosen by its leaders as it was of any inherent deficiencies in the air weapon. By ruling out before the fact even a ground threat, let alone any serious prospect of an early ground invasion, the Clinton administration and NATO ensured that air power would be stressed to the fullest when it came to attempts to engage fielded enemy forces.