A small patch of Siberia is holding onto the carbon trapped in its soil after horses were reintroduced to the area, new research has found. There is a long way to go to confirm this work is applicable everywhere, but modeling based on observations made there suggests we might have a way to prevent melting permafrost from unleashing runaway warming.

As the Arctic thaws, bacteria have started to break down organic carbon trapped in the permafrost for thousands of years, releasing it as methane and carbon dioxide. There is great uncertainty about the rate at which this will happen, but we can't ignore the possibility it could overwhelm efforts to cut back on our own emissions of greenhouse gasses. One factor we have only just begun to explore is how much grazing animals influence the rate of release.

One key thing has changed in the Arctic since the last time the world was this warm. The mammoths are gone and bison and horses have become rare. For this reason, Russian scientists Sergey and Nikita Zimov have tried reintroducing animals into an area of Siberia they have named Pleistocene Park. Together with Professor Christian Beer of the University of Hamburg, the Zimov's have published a paper in Scientific Reports demonstrating how much difference these animals make.

Siberian winters are notoriously cold. However, snow is a good insulator, keeping the upper soil in ungrazed areas to around -10ºC (14ºF). When summer comes, that rises quickly to well above freezing and the warmth penetrates deep enough to disturb the carbon monster that lies sleeping below.

However, as hoofed animals like horses and reindeer walk on the snow, they stamp it down, reducing its insulating effect. Using measurements at Swedish sites with and without reindeer, the authors show this makes a big difference to how deep the melt reaches in summer. Where the grazers find a hint of grass beneath, they may remove the snow altogether, allowing the bitter cold to seep so much deeper that even a warm summer cannot get rid of it.

When the Park's founders grazed 114 herbivores on a 1 square kilometer area (45 per square mile), the average snow depth halved. Beer calculated that even in a scenario of high global warming from other sources, this would would reduce nearly half the rise in tundra soil temperatures, cutting carbon release by 80 percent.

"It may be utopian to imaging resettling wild animal herds in all the permafrost regions of the Northern Hemisphere," Beer said in a statement. "But the results indicate that using fewer animals would still produce a cooling effect. What we've shown here is a promising method for slowing the loss of our permanently frozen soils, and with it, the decomposition and release of the enormous carbon stockpiles they contain."

Recently, Oxford University's Dr Marc Macias-Fauria used the Zimovs' work to model the number of horses and bison that would be needed to “rewild” Siberia, Alaska, and northern Canada, returning ecosystems to something like their natural state.

Macias-Fauria's work focused on grazing animals' capacity to convert other ecosystems to grassland, which at Arctic latitudes stays cooler than the alternatives. He concluded that five bison, seven to eight horses, and 15 reindeer per square kilometer should be sufficient to return the tundra to its natural state.

Beer's work shows that even when areas are already grasslands, trampling from herbivore hooves can keep the tundra colder still.