On January 14, 2015, a commissioning test at the Crescent Dunes solar power project near Tonopah, Nevada, concentrated reflected sunlight in the air above the plant’s tower. Birds flying through that region were reported burned and killed. Contemporary accounts put the toll at about 115; later reporting based on federal agency records put it at roughly 130.

That variation belongs near the beginning of the story. The headline uses the widely reported 115 figure, but the surviving public accounts do not establish a single, exact count of identified carcasses. They describe observations of birds encountering an intensely illuminated region during a particular test.

The hazard came from overlapping reflected sunlight in the airspace above a concentrating solar power plant.

What the reports say about that January test

In Henry Brean’s March 2015 reporting for the Pahrump Valley Times, biologists were said to have counted roughly 115 birds burned during the test. SolarReserve spokeswoman Mary Grikas described mirrors held in standby for several hours, directing their light above the central tower.

An April 2015 report in Renewable Energy World, which included an interview with SolarReserve chief executive Kevin Smith, put the number of mirrors involved at 3,000. It described them aimed into a tightly concentrated ring of air above the receiver.

The full installation was larger. The SolarPACES project database lists 10,347 heliostats, the movable mirror assemblies that track the Sun, and a tower height of 195 meters, approximately 640 feet. The incident therefore involved a subset of the mirror field rather than all of its reflectors.

Why some accounts give a higher toll

Phil Taylor’s August 2016 investigation for E&E News drew on emails, monitoring material and other records released by the Bureau of Land Management. It described roughly 130 birds believed killed in about four hours on January 14.

The report quoted an email from Stantec’s Josh Vittori describing monitoring biologists seeing flashes and smoke as birds entered a bright region about 1,200 feet above the ground. That height refers to the illuminated airspace, not the 640-foot tower itself. The email also said that no carcasses had been found beneath the tower and that identifying species at that distance was difficult.

These are observations interpreted as fatalities, rather than 115 individually recovered birds.

The two reported totals should not be combined or averaged into a newly precise number. Nor does their difference justify discarding the incident. The defensible description is a reported loss of more than 100 birds during a short commissioning test, with uncertainty around the exact tally and species involved.

Mirrors and stored heat explain the engineering

Crescent Dunes was designed to turn concentrated sunlight into heat and then electricity. As the U.S. Energy Information Administration explained in March 2016, mirrors direct sunlight onto a receiver at the tower, heating molten salt. The hot salt passes through a heat exchanger to produce steam, which drives a turbine.

Thermal storage allows the collected heat to be used later. The plant’s design paired a nominal 110-megawatt generating capacity with storage intended to support up to ten hours of generation without direct sunlight. Those are design characteristics, not a statement that the plant delivered that output continuously.

Ordinary photovoltaic panels work differently: solar cells convert light directly into electricity. They do not aim thousands of reflected beams at a common tower. The January event consequently describes a hazard of this concentrating arrangement and its operation; it cannot supply a bird-mortality rate for every technology called solar power.

Standby still requires somewhere to send the light

Taking mirrors off the receiver does not stop them reflecting sunlight. In the January configuration, their reflected beams overlapped above the tower instead. The choice of aiming points determined where concentrated energy accumulated while the system was waiting or being tested.

Smith told Renewable Energy World that testing was stopped after the problem was recognized. The company described a revised approach that spread standby aiming points across a much larger area, reducing the concentration at any one location. That was an operational change to the mirror controls.

Spreading the reflected energy over more area reduces its concentration. Establishing how much that change reduced bird mortality requires observations after implementation, under documented operating conditions. A company’s explanation of a correction and a measured long-term ecological outcome are different pieces of evidence.

Counting deaths takes more than watching the tower

The U.S. Geological Survey’s 2016 report, Mortality Monitoring Design for Utility-Scale Solar Power Facilities, by Manuela Huso, Thomas Dietsch and Chris Nicolai, explains why survey design matters. Searches can miss animals outside the searched area, carcasses removed by scavengers and remains overlooked by observers.

Those problems affect the difference between a raw count and an estimate of total mortality. A short observation period also cannot represent every season or operating condition. Species identification matters because the same number of deaths can have different implications for populations of different sizes.

The January observations document a specific failure during testing. Assessing the effectiveness of the revised standby procedure would require comparable monitoring before and after the change, with search effort and detection limits reported. A usable follow-up record would state the search area, observation periods and corrections for missed remains alongside the reported deaths.