BP PLC and other big oil companies based their plans for responding to a big oil spill in the Gulf of Mexico on U.S. government projections that gave very low odds of oil hitting shore, even in the case of a spill much larger than the current one. The government models, which oil companies are required to use but have not been updated since 2004, assumed that most of the oil would rapidly evaporate or get broken up by waves or weather. In the weeks since the Deepwater Horizon caught fire and sank, real life has proven these models, prepared by the Interior Department’s Mineral Management Service, wrong. Oil has hit 171 miles of shoreline in southern Louisiana, Mississippi, Alabama and northern Florida. Further, government models don’t address how oil released a mile below the surface would behave—despite years of concern among government scientists and oil companies about deep-water spills. BP’s efforts to contain the spill suffered a brief setback when an undersea robot hit the cap that’s channeling oil to the surface. BP was able to reinstall the cap Wednesday night. Separately, Interior Secretary Ken Salazar told lawmakers that he is reviewing how to re-draft a moratorium on new deep-water drilling in response to a federal judge’s ruling that struck down a blanket six-month drilling halt ordered by President Barack Obama last month. The Obama administration on Wednesday night asked the judge to delay his court ruling while the Justice Department appeals the decision, according to the Associated Press. The government’s optimistic forecasts reinforced the oil industry’s confidence in its spill-prevention technology, leading to decisions that left both oil companies and the government ill-prepared for the disaster that has unfolded in the Gulf since April 20. BP and government agencies responding to the spill have scrambled to assemble enough oil-containing boom and the ships and hardware needed to keep oil out of marshes and off beaches. Owen Kratz, chief executive officer of Helix Energy Solutions, one of the company’s working to contain the spill for BP, said Wednesday that the industry needs to have more oil containment equipment positioned to handle a blowout – instead of building containment systems after an accident. “We hope the best science will be used going forward in this model that MMS requires,” said a spokesperson for the American Petroleum Institute. The Obama administration has launched a major overhaul of the agency that regulates offshore oil and gas drilling in the wake of the Gulf spill. “Without question, we must raise the bar for offshore oil and gas operations,” a spokesperson for the Interior Department said Wednesday in response to questions about the spill models. BP has come under heavy fire from Congress and environmental groups for its lack of readiness to handle a worst-case spill. But that criticism has overlooked a key fact: BP was required by federal regulators to base its preparations on Interior Department models that were last updated in 2004. The government’s spill models have been at the center of years of debate among scientists that study oil spills. One study in the late 1990s used satellites to track almost 100 “drifters” set loose in the Gulf of Mexico to mimic floating oil. The paths of the drifting objects were compared with what the model predicted. After 30 days, the average discrepancy was 300 miles. “We have observed differences of some magnitude,” a 2003 paper said, summarizing the study. But the researchers, led by a team of scientists from the Interior Department’s MMS, concluded that the results were “neither surprising nor disappointing,” and “do not negate the utility” of the model. The scientists said the findings could lead to improvements in oil-spill modeling. Researchers have spent the past decade trying to improve modeling of oil spills. The biggest challenge: to update the models to reflect the new reality of deep-water oil drilling. Spills thousands of feet below the surface behave very differently than spills on the surface. Underwater currents, for example, can grab plumes of oil and transport them far from the scene of the initial spill, scientists say. Deep-water releases tend to break into smaller oil slicks, further complicating efforts to forecast where they’ll go. MMS said in early 2000, in a notice to lessees, that it planned to require oil companies operating in deep-water to use new oil-spill predictions specifically designed for deep water. That regulation never came into effect. Oil companies today still base their contingency plans on the government’s models, designed only for surface spills. In 2001, the then-head of the MMS environmental division wrote a paper that warned “the oil spill trajectory models currently used by the oil industry for the preparation of oil spill response plans may not be adequate for deep water.” Since then, MMS researchers have experimented with new models specifically designed to simulate deep-water oil spills. In 2005, after one such experiment, the MMS modeling team wrote in a paper that “spill response plans need to be upgraded” to deal with potential deep-water releases. But the models haven’t incorporated new deep-water simulations. Questions about the industry’s preparedness for a spill have come up repeatedly as Congress has investigated the response to the Gulf disaster. House lawmakers accused BP, Exxon Mobil Corp., Chevron Corp. and other companies last week of using “cookie cutter” contingency plans that contained numerous errors and omissions. Exxon Chief Executive Rex Tillerson pointed out that much of the company’s response plan “is prescribed by regulation, including the models that are used to project different scenarios for oil spills.” The MMS spill trajectory model is known as OSRA, an acronym for “oil spill risk analysis.” The model simulated currents and winds in the Gulf to calculate where oil slicks would travel over a period of three, 10, and 30 days. That model projected that a spill of oil on the surface in the Mississippi Canyon area, located 68 miles offshore, would have just an 11% chance of making landfall in Plaquemines Parish, La., after 30 days. In reality, Plaquemines, the area hardest hit by the current spill, got its first tar balls 22 days after the explosion. The bulk of the Gulf Coast, according to the model which projects spill trajectories for 30 days maximum, would not see oil reach shore even with a catastrophic offshore spill.