The department's risk assessment was "overly optimistic" about the chances of mistakes leading to a release of infectious material from the National Bio and Agro-Defense Facility planned for Manhattan, Kansas, according to a report by the National Research Council.

Congress asked the council, part of the National Academy of Sciences, to evaluate how accurately Homeland Security had assessed the risk in reports compiled in 2010 and early this year.

A second panel from the National Research Council is studying alternatives to the government laboratory. Its report is due out by the end of June.

Funding for the lab is on hold because of safety issues and its $1.14 billion cost.

Local residents, including some professors in the university town of Manhattan, have raised concerns over safety, and Representative Tim Bishop has questioned whether the money would be well spent given those concerns.

Scientists at the planned facility, known by the acronym NBAF, would study animal diseases introduced accidentally or by way of terrorism that could sicken livestock and humans. It would replace the aging Plum Island Animal Disease Center off Long Island, New York.

Animal illnesses studied at the lab would include the highly contagious foot-and-mouth disease and the Nepha and Hendra viruses that can spread to humans, swine fever and the Japanese encephalitis virus. The United States has not had an outbreak of foot-and-mouth disease since 1929.

Scientists and engineers on the panel declared the latest risk assessment was a substantial improvement over the 2010 report in terms of clarity and methods used, but is still "inadequate in critical respects."

The committee's role is not to recommend further risk assessment or whether NBAF should go forward, members told reporters in a conference call.

"That is within the purview of Congress," said committee chairman Gregory Baecher. "As of today they still don't have an adequate scientific basis for knowing the risks associated with NBAF."

Homeland Security estimated there is only a 1 in 46,000 chance each year of a pathogen being released from the plant through 142 possible events. The department concluded that earthquakes or tornadoes are the most likely cause for release.

But the panel faulted "the extremely low probabilities" of pathogen release from human error and said Homeland Security had exaggerated the likelihood of damage from an earthquake or major tornado, given the structural integrity of the building and the slim chances of those two disasters, especially an earthquake.

Nicole Stickel, a spokeswoman for Homeland Security, said in a statement on Friday that the panel agreed that NBAF is needed and supports its design.

She said the findings "will be incorporated into future plans and processes of operating the NBAF" and in analyzing risks.

Kansas' two Republican U.S. senators, Pat Roberts and Jerry Moran, said the evaluation showed the lab was needed and called on Homeland Security to release funding and begin construction.

(Editing by Andrew Stern and Xavier Briand)

• A Permanent Economic Emergency