Global Policy Forum

By Michael R. Gordon

In early 2003, as the clock ticked down toward the war with Iraq, C.I.A. officials met with senior military commanders at Camp Doha, Kuwait, to discuss their latest ideas for upending Saddam Hussein's government. Intelligence officials were convinced that American soldiers would be greeted warmly when they pushed into southern Iraq, so a C.I.A. operative suggested sneaking hundreds of small American flags into the country for grateful Iraqis to wave at their liberators. The agency would capture the spectacle on film and beam it throughout the Arab world. It would be the ultimate information operation.

Lt. Gen. David McKiernan, the commander of allied ground forces, quickly objected. To avoid being perceived as an occupying army, American forces had been instructed not to brandish the flag. The idea was dropped, but the C.I.A.'s optimism remained. The agency believed that many of the towns were "ours," said one former staff officer who attended the session. "At first, it was going to be U.S. flags," he said, "and then it was going to be Iraqi flags. The flags are probably still sitting in a bag somewhere. One of the towns where they said we would be welcomed was Nasiriya, where Marines faced some of the toughest fighting in the war."

Just as the intelligence about Iraq's presumed stockpiles of unconventional weapons proved wrong, so did much of the information provided to those prosecuting the war and planning the occupation. In a major misreading of Iraq's strategy, the C.I.A. failed to predict the role played by Saddam Hussein's paramilitary forces, which mounted the main attacks on American troops in southern Iraq and surprised them in bloody battles. The agency was aware that Iraq was awash in arms but failed to identify the huge caches of weapons that were hidden in mosques and schools to supply enemy fighters.

On postwar Iraq, American intelligence agencies underestimated the decrepit state of Iraq's infrastructure, which became a major challenge in reconstructing the nation, and concluded erroneously that Iraq's police had had extensive professional training. And while intelligence experts noted an insurgency in its catalog of possible dangers, it did not highlight that threat. The National Intelligence Council, senior experts from the intelligence community, prepared an analysis in January 2003 on postwar Iraq that discussed the risk of an insurgency in the last paragraph of its 38-page assessment. "There was never a buildup of intelligence that says: 'It's coming. It's coming. It's coming. This is the end you should prepare for,' " said Gen. Tommy R. Franks, the former head of the United States Central Command and now retired, referring to the insurgency. "It did not happen. Never saw it. It was never offered."

The Central Intelligence Agency has come under harsh criticism for its failings on Iraq's weapons and the Sept. 11 terror attacks, and critics have urged that it be overhauled as part of a broad reform of the nation's intelligence community. The agency declined requests for interviews for this article and declined to respond to written questions submitted to its chief spokesman.

Richard J. Kerr, a former deputy director who was asked by the agency to review its intelligence analysis on the Iraq war, said in an interview that much American intelligence on postwar Iraq was on the mark, particularly the assessment predicting the resentment of Iraqis if the United States did not transfer power quickly to a new Iraqi government. Still, he acknowledged some deficiencies. "Intelligence assessments on the likely Iraqi impatience with an extended U.S. presence and the role of the army in Iraqi society were particularly prescient," Mr. Kerr said. "The intelligence accurately forecast the reactions of the ethnic and tribal factions in Iraq," he said. "These positive comments, however, cannot gloss over the fact that Iraq revealed some serious systemic problems in the intelligence community. Collection was poor. Too much emphasis was placed on current intelligence and there was too little research on important social, political and cultural issues."

Trying to Catch Up

Despite more than a decade of antagonism between Saddam Hussein's government and the United States, the Bush administration was operating with limited information when it began to consider the invasion of Iraq. After the 1991 Persian Gulf war, collecting intelligence on Iraq was not always the top priority for American spy agencies, which were burdened by a multitude of potential crises and threats.

Iraq was considered a Tier 2 country. North Korea, in contrast, was Tier 1. As the agencies saw it, North Korea possessed an active nuclear weapons program and a large conventional army in striking range of South Korea and the American forces there. Iraq was seen more as a gathering threat. The months before the war were a scramble for more intelligence. The American military did its best to fill the gaps, using Predator drones, U-2 spy planes and other surveillance systems. The land forces command printed 100,000 maps of the southern Iraq oilfields, which the Marines were to secure. Detailed block by block analyses were prepared for downtown Baghdad.

Iraq, in intelligence parlance, was a "glass ball environment," meaning the weather was often conducive to collecting images from above. Much of the intelligence was derived from reconnaissance systems, not from operatives on the ground. With few spies inside Iraq, the agency relied on defectors, detainees, opposition groups and foreign government services, according to a Senate report. "Some critics have claimed during the prewar period, we did not have many Iraqi sources, " James L. Pavitt, former deputy director for operations for the agency, said in June in a speech to the Foreign Policy Association. "We certainly did not have enough," he said. "Until we put people on the ground in northern Iraq, we had less than a handful. As I mentioned before, the operating environment was tremendously prohibitive, and developing the necessary trust with those Iraqis who had access was extraordinarily difficult in light of the risks they faced. Once on the ground, however, our officers recruited literally dozens of agents - some of whom paid the ultimate price for their allegiance to us."

The C.I.A. inserted agents in the southern oil fields shortly before the war. American intelligence officers obtained the telephone numbers of Iraqi generals and called to encourage them not to fight. Fearful that the calls were a loyalty test by Saddam Hussein, some changed their numbers, which hindered their efforts to talk to each other when the war was under way. The United States gained a detailed understanding of Iraq's oil infrastructure and obtained a secret map of Iraq's Baghdad defense plan. The C.I.A. also helped debunk one threat that the military had worried about: the possibility that Mr. Hussein's government would flood the country to thwart an allied advance.

The agency, though, turned out to have a less clear understanding of what the United States would face once it invaded Iraq, or of Mr. Hussein's military strategy. In January 2003, the National Intelligence Council issued its assessment of what might happen after the dictator was ousted. The report cautioned that building democracy in Iraq would be difficult because of its authoritarian history. And it warned of the risk that the American forces would be seen as occupiers. "Attitudes toward a foreign military force would depend largely on the progress made in transferring power, as well as on the degree to which that force were perceived as providing necessary security and fostering reconstruction and a return to prosperity," it said. The report also noted that quick restoration of services would be important to maintain the support of the Iraqi public.

Broader Picture Was Missing

But the analysis was less prescient on other points. The study underestimated the fragile state of Iraq's infrastructure, suggesting it could be fixed quickly if it were not extensively damaged in the fighting. "Iraqis have restored their physical infrastructure quickly in previous wars," it stated. The United States chose not to attack the electrical grid, knowing that it would soon need to administer and reconstruct Iraq. But the electrical system collapsed from long neglect, and difficulties in restoring the service left much of the capital in darkness and aggravated residents' fears about crime.

In assessing potential threats, the intelligence report also gave far more weight to the possibility of score-settling among Iraqi ethnic groups than to an insurgency. The discussion of that prospect was remarkably brief. "The ability of Al Qaeda or other terrorist groups to maintain a presence in northern Iraq (or more clandestinely elsewhere) would depend largely on whether a new regime were able to exert effective security and control over the entire country," it noted. "In addition, rogue ex-regime elements could forge an alliance with existing terrorist organizations or act independently to wage guerrilla warfare against a new government or coalition forces."

Mr. Kerr, the former C.I.A. official, said the agency's regional experts were more concerned than the assessment by the National Intelligence Council about the potential threat of guerrilla attacks by paramilitary forces after Mr. Hussein's government was toppled, particularly if American troops stayed in Iraq for a significant period of time. But he acknowledged that the assessments did not anticipate the sort of virulent insurgency that Americans forces now face in Iraq. "They did believe there would be a fairly significant stay-behind group of Saddam loyalists and fedayeen that would attract outside support," he said. "But it would be stretching it to reach too far down this line. I could not justify saying that they predicted the war as it has developed."

Gaps Become Apparent

From the start of the war, it was clear that some of the intelligence was off. On March 19, 2003, for example, George J. Tenet, the director of central intelligence, told the White House that he had firm evidence that Mr. Hussein and his family were in a suburb near Baghdad known as Dora Farms. The Iraqi leader and his two sons were thought to be hiding in a concrete bunker; the C.I.A. provided exact coordinates.

Lt. Gen. Michael (Buzz) Moseley, the air war commander, who was at an air base in Saudi Arabia, quickly developed a plan for stealth fighters to drop satellite-guided bombs, followed by cruise missiles. The planes hit their targets. But when American forces got to Dora Farms after the fall of Baghdad, they discovered there was no underground bunker at that site, General Moseley said in an interview last year. The Iraqis responded to the attack by firing missiles at American forces in Kuwait. American intelligence learned that a small number of oil wells had been set on fire, so the land war was accelerated.

Senior military officers and intelligence analysts had expected that the Iraqi leader would center his defense in Baghdad, and planned for a decisive battle against his Republican Guard divisions and special military and paramilitary units in the capital. The American forces discovered in the first days of the war that the Iraqis had a different strategy. The Marines learned this the hard way. Task Force Tarawa, a Marine unit assigned to secure the bridges in eastern Nasiriya, was told that a C.I.A. source had reported that Iraq's 11th Infantry Division, which was to guard the bridges, would probably surrender. Convinced that Nasiriya would be a relatively easy fight, senior Marine commanders did not make any reconnaissance drones available.

The fight in Nasiriya turned out to be one of the toughest of the war. Thousands of paramilitary fighters, the Saddam Fedayeen, had taken up positions there and in the other southern cities, including Samawa and Najaf, determined to put down any Shiite rebellion and to draw the Americans into bloody bouts of urban warfare. In Nasiriya, the Marines' mission was complicated when the Army 507th Maintenance Battalion - made famous when Pfc. Jessica Lynch was taken prisoner - stumbled into the city. The Marines suffered 18 dead the first day, some by American fire, after it ran into hordes of Iraqi fighters. "All indications were that it would not be much of a fight, that the Iraqis were probably going to capitulate," recalled Joseph Apodaca, a retired lieutenant colonel who served as the intelligence officer for the task force that fought in Nasiriya. "After that contact in Nasiriya, I lost quite a bit of faith in national-level reporting."

Flawed intelligence led to other units' being caught by surprise, too. In Samawa, the Army's Third Squadron, Seventh Cavalry Regiment had been told, based on intelligence reports, to be prepared to conduct a parade to show solidarity with the inhabitants. Sgt. First Class Anthony Broadhead, who led a group of Bradley fighting vehicles and M-1 tanks into the city, was standing in the hatch of his tank and waving when the Iraqis responded by shooting. A fierce firefight between the soldiers and the paramilitary forces broke out. "The fighting that occurred in Samawa was not with conventional Iraqi forces but with Saddam Fedayeen and Baath Party members," noted Lt. Col. Terry Ferrell, the unit's commander. "In the intelligence summaries, we had heard about this type of enemy, but they had not been given any credit for being as tenacious and capable of fighting as they demonstrated not only in this battle, but in every other fight the squadron encountered."

The flawed information provided to the units in Nasiriya and Samawa were not the only lapses. American intelligence knew Iraq had huge quantities of conventional weapons, but did not realize that arms caches has been established in schools, hospitals and mosques as part of the strategy to turn the southern cities into bastions for the Saddam Fedayeen. "What intelligence did not reveal was the magnitude of the regime's weapons holdings," the First Marine Division noted in its after-action report. "Huge caches were hidden in every area of the country, but it was only after the division closed on those facilities that the full magnitude of the distribution of tons of weapons and ammunition throughout the country came to light." The failure of the American intelligence agencies to detect the paramilitary forces in the south made it harder to anticipate the potential for an insurgency, Colonel Apodaca said. "They are good at reaching into the higher levels of organizations, but those guys don't see clearly what is going on at the bottom," he said. An American general who asked not be identified because of the sensitivity of his position said: "I think it is safe to say we had an accurate picture of their forces in terms of their general capability and size. But we did not have a good sense of how they were intended to be used. We started out with a deficit of human intelligence, of sources inside."

Misreading the Consequences

Even in the last days of Mr. Hussein's government, some preliminary reports suggested that a guerrilla campaign could emerge once he was toppled. On April 5, 2003, a Defense Intelligence Agency task force said the Baathists had made plans to wage a protracted guerrilla war and would form a tactical alliance with Islamic jihadists. Their goal, the task force said, was to produce casualties so that the American public would push for United States forces to quit Iraq. On April 9, American intelligence agencies issued a "sense of the community" memo - their collective judgment - which concluded that Baath Party cadres, Iraqi security forces and paramilitary fighters were operating independently under longstanding orders. They could be expected to fight on until they were neutralized, Saddam Hussein was killed or senior Iraqi leaders whom they respected ordered them to stop fighting. Even then, the memo said, some would fight on.

Later, after the fall of Baghdad, American intelligence would learn more about preparations that had been made for a guerrilla campaign. The Iraq Survey Group, which was sent to Iraq primarily to search for evidence of unconventional weapons, uncovered some documents. The papers concerning Falluja, Iraq's most volatile city, identified storage areas for weapons caches and provided the names of 75 Saddam Fedayeen and 12 suicide volunteers who were expected to join in the fight. The battle for the future of Iraq has only intensified as the insurgency has become entrenched. It has now taken thousands of lives, crippled reconstruction, threatened election of a new Iraqi government and forced American troops to engage in a grueling guerrilla conflict. The C.I.A. and other intelligence services are deeply involved in gathering information to help subdue the rebels controlling some of Iraq's cities, trying to fill in the gaps that existed when the Americans invaded Iraq. "We understood their conventional force, their missiles programs, their air force," recalled Maj. Gen. James M. (Spider) Marks, now retired, who served as the chief intelligence officer for the land war command. "The elements of power which we could assess from a distance we assessed quite well. What we missed was the fine granularity that you get from a physical presence on the ground, by interacting with the Iraqi people over the years. Since 1991, we lost our finger on the pulse of the Iraqi people and built intelligence assessments from a distance. We did not appreciate the 'fear factor' and the grip that the regime had on the people."