Last week, the Ninth Circuit Court of Appeals agreed with the plaintiffs that the Environmental Impact Statement issued by the Bureau of Ocean Energy Management to support oil and gas leasing in the Chukchi Sea was flawed.  Although the decision was split and the Ninth Circuit’s track record on appeal is less than perfect, I think that they probably got it right.  Moreover, the flaws identified by the court provide a useful lesson to agencies in performing environmental analysis of probabilistic outcomes.

The issue on which BOEM got reversed was its use of a 1 billion barrel estimate of economically recoverable oil from the area subject to the lease.  BOEM chose the 1 Bbbl estimate on the ground that that was the lowest amount that could be present that would make development economically feasible.  It was also the amount assumed to be recovered from the first field within the lease area, even though BOEM acknowledged that, once one field were developed, additional development would be likely.  In fact, BOEM estimated that the range of recoverable oil was from 1 Bbbl to 29 Bbbl, and the “mean recoverable oil resource” was determined to be 12 Bbbl.  As the Court noted:

The mean estimate of economical oil production, at the center of the distribution curve, is by definition a more likely occurrence than is the lowest estimate of viable oil production.

The other significant point addressed by the Court was BOEM’s failure to consider any oil recovered other than from the first developed field.  The FEIS itself acknowledged that:

[w]hen the first project overcomes the cost, logistical, and regulatory hurdles, more projects are . . . likely to follow.

Finally, BOEM argued that, even if flawed, the error could be corrected once individual projects are assessed under NEPA.  The Court agreed that NEPA analysis can proceed stepwise under regimes such as here, where a lease is followed by licensing of individual projects.  However, the overall estimate of recoverable oil is by definition related to the overall impacts of the decision to lease, and not something that would be addressed in the individual licensing decisions.  Thus, the time to get it right was now.

Finally, I’ll note that agencies are not helped when internal emails by important decision-makers characterize the assumptions used as “subjective” and “entirely speculative.”