The U.S. State Department on Friday corrected several errors it made in a key study evaluating the impact of the proposed Keystone XL pipeline, including an understatement of how many people could be killed on railroad tracks if the project were rejected and oil traffic by rail increased.

The department said, however, these corrections had “no impact” on the integrity of the conclusions of the January report, which played down potential environmental consequences of TransCanada Corp’s Canada-to-Texas project.

The Obama administration has not yet decided whether to approve the project.

The January report determined that blocking the controversial pipeline could increase oil train traffic and lead to an additional 49 injuries and six deaths per year, mostly by using historical injury and fatality statistics for railways.

That finding was a small element of a broader examination of how building the pipeline could impact climate change, endangered species, quality of life and other issues.

But the report mistakenly used a forecast for three months of expected accidents rather than full-year figures, officials said. The correct estimate of deaths should be roughly four times as large – between 18 and 30 fatalities per year.

Officials also revised a footnoted reference to how much electricity would be needed to power pumping stations along the route of the pipeline that would link Canada’s oil sands region to Texas refineries.

Running at something less than full capacity, the pumping stations would not require as much electricity – and so tax power plants less – than originally reported.

Revising that footnote has no impact on the State Department’s estimation of expected greenhouse gas emissions tied to the pipeline, a spokesperson said.

“It is common practice to publish an errata sheet that notes and corrects errors in voluminous technical documents such as environmental impact statements,” the State Department said.

“The Department has reviewed each of the items listed in the errata sheet and has determined that they have no impact on the integrity of, or the conclusions reached in, the (final report).”

The State Department also published several dozen public comments that had not been included in the roughly 2.5 million it received and previously disclosed. (Reporting by Patrick Rucker; Editing by Ros Krasny and Will Dunham)