Zebras drink water from a pond in Tanzania. (Nariman El-Mofty/Associated Press)

Tim Searchinger and Lyndon Estes are, respectively, a research scholar and associate research scholar at Princeton University’s Woodrow Wilson School for Public and International Affairs.

In recent years, as governments have begun to protect tropical forests because of the carbon they store, other vast tropical ecosystems have come under increasing threat. These are the wet savannas, which have a mix of grasses, trees and shrubs and receive enough rainfall to grow crops. Farmers are converting these areas to croplands at vast rates, and global studies by the World Bank, the United Nations’ Food and Agricultural Organization and others often assume that, because they are not forests, savannas are environmentally expendable. In a new scientific paper, we find that this just isn’t so — at least if the world cares about climate change and biodiversity.

Climate change treaties have committed countries to work together to protect tropical forests and their carbon, but not necessarily savannas. That low level of respect has created the misperception that the world has a vast reservoir of lands it can call on for human needs at low environmental cost. Policies designed to promote biofuels rely on large estimates of available land that, in turn, rely on assumptions that wet tropical savannas can produce biofuel crops without much downside. Some African governments and foreign investors have used this assumption to justify the lease of vast savanna areas to grow staple crops for export. Brazil, which has aggressively and admirably moved to reduce deforestation in the Amazon in recent years, has seen an increase of agricultural clearing of the Cerrado, a tropical savanna with extraordinary biodiversity.

Such open-access policies are based on the reasonable view that savannas are less ecologically valuable than dense tropical forests. But less valuable is not the same as low value. If you had to choose, you might prefer to be shot in the leg rather than the chest, but you’d try quite hard to avoid getting shot in either place.

We analyzed the environmental costs of converting savannas to cropland relative to the foods or biofuels they might produce. Even though they are not forests, the African savannas — half of the world’s remaining savannas — hold a lot of carbon in trees, shrubs, grasses and soils. We found that, even with optimistic estimates, converting more than half of Africa’s wet savannas for cellulosic ethanol would increase the level of greenhouse gases in the atmosphere for more than 50 years, while less than 1 percent of the savannas can be expected to produce cellulosic ethanol in ways that meet European climate standards. Additionally, only 2 percent to 11 percent of these savannas have realistic potential to produce corn or soybeans while releasing significantly less carbon per ton of crop than average global croplands.

They are also richer biologically than many think. We found that the average numbers of mammal and bird species using these savannas almost match those of the world’s dense tropical forests, and many live nowhere else on Earth. Overall, the diversity of animals in three-quarters of Africa’s wet savannas exceeds that of three-quarters of the rest of the world. Warm, wet parts of the globe tend to develop great biodiversity, whether they are forests or not.

Do these findings mean such lands should never be touched? That would be neither feasible nor desirable. Africa’s food needs are growing rapidly, and its population is expected to double by 2050. Sub-Saharan Africa already imports up to a fifth of its staple crops. To avoid becoming even more dependent, the region will probably have to triple its food production. Even if its yields progress well, farmers will still have to convert tens of millions of acres and release billions of tons of carbon in the atmosphere to feed their fellow Africans. One lesson from this analysis is the need for better science and policies to focus the inevitable land conversion on those savannas with the lowest carbon and biodiversity costs for each ton of crop they are likely to produce.

But the key lesson for the rest of the world is that Africa has no environmentally low-cost surplus land that can be used to supply the world either staple foods or biofuels. If we care about climate change and biodiversity — and we should — the world has to make a concerted effort to hold down the demand for new cropland even as the world is on course to need 70 percent more food by 2050.

Governments therefore need to help farmers produce more food on existing cropland, not only in Africa but throughout the world. The world’s wealthier nations must reduce food waste and their consumption of meat, which is land-intensive. And governments have to stop making the land-use challenge worse by demanding biofuels that would require hundreds of millions of acres to supply only a small share of global energy.