Loux, who coordinates the state's official opposition to the project, suggested the computer model, known as the Total System Performance Assessment, will not meet the requirements for DOE to gain a repository license.
"The model is so complicated and so large, and takes so many computers to run it," Loux said. "It is fundamentally not capable of being checked by any third party."
The Yucca Mountain computer model is complex out of necessity to calculate repository safety over thousands of years, DOE spokesman Allen Benson said in response.
"The computer system in place allows for thousands of calculations to be made in a reasonable time at minimal cost," Benson said. "These calculations allow the evaluation of repository safety -- our primary concern -- in the manner required by the NRC."
The Energy Department will provide the NRC with tools to evaluate the model, Benson said.
"This is being done for the Nuclear Regulatory Commission, and they are the ones who have to be able to operate it," he said.
The performance assessment is the centerpiece of the Energy Department's bid for an NRC repository license.
By lodging its complaint, the state is laying the groundwork for a possible due process lawsuit on the grounds it was denied the ability to properly review the license application that DOE officials say they will file in June 2008.
"From our perspective, this is a huge deal," Loux said in an interview. "If nobody can review this thing, the application is unreviewable, and it can't go forward. I don't know if DOE has an idea of how this can be simplified. I don't think so."
Loux said the state bought a $10,000 software package designed to decipher the TSPA. Still, he said, "It is hard to imagine that we will be able to check DOE's work adequately" because input files "are likely to be extremely large."
The DOE network is modeled on a "Beowulf cluster," a collection of computers wired to Ethernet, a family of computer-networking technologies, and set up with software that enables them to operate as if they were a supercomputer.
In his letter, Loux asked Klein to investigate the issue.
"It now appears to pose the most critical issue vis a vis the transparency of DOE's work," Loux said.