In a study published Wednesday in the journal Science Advances, Kock and his colleagues identify weather as the culprit. The 2015 mass mortality event, like two previous incidents in the 1980s, coincided with unusually high temperatures and humidity during the spring calving season, they report.

“The animals were in a sort of foggy soup, and it looks like that bacteria naturally occupying the tonsils were woken up by this environmental factor,” Kock said.

Tracing the saigas’ killer was no easy task. Their habitat is so remote, and the disease so swift, it was impossible to capture a healthy member of an infected herd for comparison. Instead, the researchers from Europe and Kazakhstan tracked the ailing animals up to 1.5km a day, performing examinations of diseased and deceased individuals where they fell. The only known survivors were in small herds far from the main group, difficult to pin down amid the vast landscape.

After necropsies revealed P. multicoda as the proximate cause of death, the researchers considered a range of factors that might have made the antelopes so vulnerable to a normally benign bacterium. They tested for pathogens that could have increased the danger and looked for evidence of nutritional deficiencies that would have weakened the animals’ immune system. They even looked at whether fuel from rocket launches at Kazakhstan’s Baikonur Cosmodrome triggered the outbreak.

Yet only three variables showed a strong relationship with the 2015 die-off and similar events in 1988 and 1981: average maximum daily relative humidity, average minimum daily temperature and average maximum daily dew-point temperature.

The scientists are still sorting out the mechanism by which these weather fluctuations changed the behaviour of the bacterium. They are also quick to point out that a connection to weather doesn’t mean the die-off can be blamed on climate change.

But climate change is expected to alter the weather in the saigas’ native habitat in ways that could make them more vulnerable to outbreaks like the one in 2015. In the past 15 years, scientists have recorded a 10-degree increase in central Kazakhstan’s median May temperatures, Kock said – exactly the time of year when the animals gather for spring calving.

“We know the trends are moving in that direction,” he said. “It’s pretty likely that we’ll get other events like this again.”

Conservationists would be powerless to stop such an outbreak; there is no way to give so many saiga a vaccine for haemorrhagic septicaemia, the disease caused by P. multicoda. Figuring out what happened in Kazakhstan could help researchers understand and predict other die-offs, however.

“Animals and their microbiomes have evolved within certain limits over millions of years, and now suddenly … we are shifting the environmental conditions rapidly outside that envelope,” Kock said. As he and his colleagues hypothesise, “you’ve now got bacteria responding quite quickly, while the host doesn’t have the capacity to change. If the immune system doesn’t adapt to that quicker change in behaviour of the pathogen, that’s how the balance is lost.”

The ramifications could be severe. “If this starts happening on a wider scale,” Kock continued, “it’s potentially catastrophic for livestock, for conservation, and ultimately for all mammals.”