Thanks to newly developed statistical methods, researchers can now use data from smaller surveys to scale wildlife population forecasts across borders.
The breakthrough — detailed Monday in the journal PNAS — will allow biologists and wildlife managers to precisely model apex predator populations at unprecedented scale.
A growing body of research has highlighted the importance apex predators play in maintaining the health and balance of Earth's ecosystems. The research has made the protection of these large species all the more important, but because big carnivores tend to spread out across the landscape, modeling population dynamics and distribution of apex predators remains difficult.
Statistical methods developed by researchers at the Norwegian University of Life Sciences promises to make the task more manageable — bringing a landscape-level approach to wildlife modeling.
"The way we tend to study populations is a bit like looking at an elephant through a microscope," Richard Bischof, study author and NMBU ecologist, said in a news release. "We can understand fine details but find it difficult to make out the entire shape."
Thanks to a combination of improved genetic sequencing technologies and statistical modeling techniques, researchers can monitor and analyze populations across much broader scales — relying only on collected samples of feces, urine or hair.
Over the last 20 years, researcher in Norway and Sweden have been working together to amass a giant database of DNA samples from brown bears, grey wolves and wolverines across Scandinavia.
Deploying novel statistical methods, researchers were able to use the database to construct a detailed distribution map of the three different predator populations across all of Scandinavia.
The new modeling techniques don't simply offer a snapshot in time, but can be frequently updated with new data to detect population changes.
"Wildlife surveys rarely detect every individual," said Bischof. "So, to estimate population size, we cannot simply count the number of animals for which DNA is found. Our models correct for this."
