China counts wild pandas. Nobody believes its tally

American zoos point to a rise in the wild panda numbers — 1,864 at last count — as proof that the tens of millions of dollars they have spent on conservation have hit their mark.

But that population number comes from a 2013 Chinese government tally that no one in the conservation world actually believes — including the zoos themselves.

Zoos worldwide rent pairs of pandas from China for up to $1.1 million a year, then breed them in the hope of releasing future generations into the wild.

It is one of the world’s signature conservation programs. We spent months investigating it and found that few pandas had been released and that individual pandas had been injured, burned and even killed by aggressive artificial breeding.

Zoos have glossed over the harms and declared the program a success, often by pointing to the uptick in panda numbers.

But a 2006 research proposal from the National Zoo in Washington declared that the census methodology had “never been examined systematically and scientifically for its accuracy” and had “no quantitative data” to back it up.

Another document from 2012 described a “lack of sound data on giant panda numbers in nature.”

“The number is, of course, not fully reliable,” said Jianguo Liu, an ecologist at Michigan State University and EEB core faculty member who studies panda habitat in China.

China’s panda surveys, carried out about once a decade, are ambitious, multiyear affairs led by the national forestry bureau. In the most recent round, surveyors fanned out across steep terrain with GPS devices and topographic maps, sometimes taking local farmers as guides.

They searched for panda droppings, then analyzed them for bamboo fragments.

Pandas are solitary creatures that stick to particular forest areas, called home ranges. They have distinctive bite marks. They are also prolific poopers.

By analyzing bite marks that remain on the bamboo fragments, the thinking goes, surveyors can estimate how many pandas live nearby.

In a 2010 blog post, Yu Changqing, a scientist who was a survey adviser, criticized the methodology and questioned its organization.

Millions were allocated to administrative expenses, while experts were sidelined. He wrote that “the giant panda survey is a scientific matter that should have been left to scientific researchers.” He added, “The final result was really regrettable.”

The National Zoo proposed alternatives, calling the bite-mark approach “labor intensive and controversial.”

Experts say there are better options.

“DNA is definitely the gold standard,” said David Garshelis, a co-chairman of the Bear Specialist Group at the International Union for Conservation of Nature, a wildlife monitoring group.

Chinese experts said they had trouble finding enough fresh feces for consistent DNA analysis.

China’s forestry bureau arrives at a precise number — down to the last digit. It won’t say exactly how it does that.

American zoos and the World Wildlife Fund help finance the last census, but even they don’t know.

“They didn’t provide any of the information to understand how they did the analysis,” said Colby Loucks, a World Wildlife Fund vice president.

The forestry bureau did not answer our questions about the census. It said our reporting was inaccurate but did not elaborate.

“We just don’t have access” to timely data, said Qiongyu Huang, a wildlife biologist at the Smithsonian. “It’s kind of a bummer. We could have done a lot more research.”

Pandas are political. Censuses are political. A panda census is extremely political.

In the 1980s, some officials in western China worried that revealing the existence of pandas would lead to logging bans, so scientists believe they underreported or denied the presence of pandas.

Other pressures may lead to overreporting. For some poor areas in western China, pandas can bring in government funding.

These competing interests make it hard to know the truth. “There are so many human interventions,” said Lung Yuan Chih, a director of the Taiwan Human-Animal Studies Institute.

In 2016, the International Union for Conservation of Nature declared that pandas were no longer endangered.

The group’s experts believed that, even if the 1,864 count wasn’t exact, panda numbers were trending upward and their habitat was growing, said Dr. Garshelis. He said that China provided the group with some data, like the estimated number of pandas in each area, along with the locations of the droppings analyzed.

Annalisa Meyer, a National Zoo spokeswoman, said that, while “scientific discourse is expected,” the international community has reached a consensus that the panda population is increasing. She said that more habitat had been protected, and that researchers had observed wild pandas moving into those new areas.

Oddly enough, China’s forestry bureau disputed the I.U.C.N. decision, fearing that it would remove pandas from the spotlight, Dr. Garshelis said.

But China, like American zoos, now says the new designation proves the program’s success.

“Ultimately,” Dr. Garshelis said, “they became convinced by their own numbers.”

Read the story in The New York Times.