The natural world is a complex place. So complex, in fact, that it can be hard to tell whether something—be it a shark, flower, or bacterium—is good or bad, or even helpful. The same is true for a wildfire. After reviewing the evidence, ecologists writing in the journal Science say they’re still trying to figure out the complex relationship between wildfires and biodiversity.

In the 1990s, scientists hypothesized that pyrodiversity, or exposure to different types of fire, benefited an ecosystem’s biodiversity in the long run. Some fires appear more often or burn more intensely, thereby clearing out a range of habitats and making way for new species. The black-backed woodpecker, for example, was more likely to move into an area that had experienced severe burns. Many studies have supported that hypothesis. One found that sections of Yosemite National Park with higher pyrodiversity also had a significantly higher diversity of plants and their pollinators.

But other studies have found just the opposite. Biodiversity in Australia’s eucalyptus forests decreased as pyrodiversity went up, as safe plant habitats were harder to come by in the dry woodlands.

L.T. Kelly and L. Brotons. 2017. Science.

The authors of the new review found no tidy answers. Whether fire helps or hurts, they say, depends on the ecosystem itself. And the nature of fire’s effect on biodiversity is more than just a philosophical or theoretical question. In some places, the authors write, putting out wildfires will be essential; in other places, at other times, it could be better to let them burn.

“There is a need to further develop fire management approaches that, while supported by ecological theory, are better tailored to local conditions,” they write. “Interdisciplinary approaches involving ecologists, climate and fire modelers, scenario planners, and social scientists will help to ensure that we better understand and use fire to promote biodiversity.”