Industry estimates say it would cost utilities and their customers up to $100 billion to retrofit the [power] plants and save more fish, as environmental groups advocate.
But that might not be a tidy solution: the retrofits could leave less water for people.
The cost per plant is $1 billion, says the industry.
EPA's 316(b) website is here (comment period is open, mine is below). Here is the benefits analysis, full of benefits transfer.* And it looks like the EPA will be conducting a national choice experiment survey (PDF fact sheet) in order to develop benefit estimates.
*It is odd to see reference to one of your obscure papers, Whitehead (1993), in the table of contents. The EPA is using that paper to estimate the benefits of protecting sea turtles. The problem with using these older studies is that they don't account for the hypothetical nature of the willingness to pay exercise. As such, the benefits are likely overstated. Hypothetical bias adjustments began appearing after 1997 (with Champ et al.'s follow-up certainty study). Has there been a meta-analysis of CVM studies that includes a hypothetical bias variable?