Deconstructing the Dispute Over Ivanpah Bird Kill Numbers

In the wake of a recent Associated Press story on bird deaths at the Ivanpah Solar Electric Generating System, controversy has arisen over the actual numbers of birds being killed at Ivanpah by the plant's concentrated solar energy, a.k.a. "solar flux."

The AP story by Ellen Knickmeyer and John Locher, published Monday, fueled the new dispute with this sentence, early in the piece: "Estimates [of birds killed] per year now range from a low of about a thousand by BrightSource to 28,000 by an expert for the Center for Biological Diversity environmental group."

That's certainly a huge discrepancy: a factor of 28. How did BrightSource and the the Center for Biological Diversity (CBD) expert arrive at such diametrically opposed estimates?

Let's cover that 28,000 figure first. That number came from testimony offered to the California Energy Commission by ecologist K. Shawn Smallwood during hearing on the proposed Palen Solar Electric Generating System, which BrightSource would be building with partner Abengoa Solar in a joint venture called Palen Solar Holdings. Palen would have a design substantially similar to that of Ivanpah, but would be considerably larger.

That means that the emerging issue of bird mortality at Ivanpah is of supreme concern to the regulators deciding whether to allow Palen to be built, as the Palen project would likely be even more hazardous to birds.

In testimony offered to the commission on June 23, Smallwood discussed how he would estimate estimates of avian mortality at Ivanpah, based on recorded deaths at Ivanpah during April and May of this year. In that two-month period carcass surveyors recovered 183 birds from Ivanpah's fields of mirrored heliostats.

Smallwood pointed out that since many of the birds recovered were small; warblers, hummingbirds, and the like. If the prevalence of small species was representative of all birds killed at Ivanpah and not just the ones surveyors found, local scavengers such as ravens and kit foxes would likely have been able to remove many of the carcasses before they could be recovered.

Biologists call the unknown degree to which dead birds and other animals are eaten before they can be counted "scavenger bias." Smallwood suggested that the scavenger bias for Ivanpah's April and May mortality might be around 20 percent, meaning that ravens and foxes and such ate four carcasses for each one found.

If that estimate was true, said Smallwood, the actual number of dead birds in the areas surveyed for carcasses would run somewhere around 473 birds per month, or about five times the average monthly death toll for April and May. And since only one fifth of the facility's heliostat fields are being systematically surveyed by Ivanpah's contracting biologist, that 473 birds per month would need to be multiplied by five again, which would give us a figure of 2,365 birds per month across the entire range of heliostat fields.

Multiply that monthly estimated total by 12, and you get a figure of 28,380 birds potentially killed at Ivanpah in a year of operation.

That's a lot of birds, and Smallwood agreed in his testimony that his calculations were based on a few assumptions that he could not verify completely, mainly because the data on both Ivanpah bird mortalities and the methodology Ivanpah biologists use to collect it just isn't hard enough to make confident predictions. As Smallwood said:

Critics of those with concerns about Ivanpah's effect on wildlife have been quick to jump on the "back of the napkin" comment. In his testimony, Smallwood provides a detailed explanation at the methodology behind his own preliminary calculations. Is that methodology solid? Answering that question is, frankly, above ReWire's pay grade. Take a figure accompanying Smallwood's explanation of how he adjusted avian mortality estimates from a 1986 study of the now-defunct Solar One project near Barstow, for instance:

We here at ReWire aren't even sure we could fit that on the back of a napkin. Curious readers with post-graduate groundings in the biological sciences can take a look at his testimony and decide for themselves.

At any rate, that's where the AP's figure of "28,000 birds per year" comes from.

In response to the predictable uproar generated by AP's glancing coverage of a story ReWire has been flogging since early 2013, project designer BrightSource -- whose corporate survival depends on acceptance of its power tower technology -- leapt into damage control mode.

"Let's be clear," wrote BrightSource Vice President Joe Desmond in an August 19 blog post on the company's site, "no one disputes that certain levels of concentrated solar flux present a risk to birds."

Desmond continued, "In fact, Ivanpah reported 321 avian fatalities between January and June 2014, of which 133 were related to solar flux."

He then offered by way of comparison a bulleted list of other ways birds die in greater numbers, including as many as three billion killed by cats and around a billion crashing into buildings.

In a companion blog post, Desmond reiterated the company's line:

Desmond is careful to include the word "reported" along with the number 321, but some of BrightSource's supporters haven't been so careful.

Why is that "reported" so important? Precisely because of the potential sources of error cited by Smallwood.

Even if Smallwood's estimates of bias from scavenging and partial surveys are too pessimistic, the fact remains that that bias does exist. In order for the 321 "reported" bird deaths to be the total avian mortality at Ivanpah in the first half of 2014, we'd have to assume that surveyors collected every last bird that died at Ivanpah.

And that would mean that ravens and foxes and other scavengers didn't find a single carcass, and that all the birds that died thoughtfully made sure to land in the fifth of the heliostat field that was being searched.