Bontioli Natural Reserve Savannas Are Losing Carbon

Savannas hardly resemble carbon vaults. Their trees are widely spaced, their grasses can remain brown for months, and compared with dense tropical forest, the landscape may appear almost empty.

That assumption has consequences. Twenty years of satellite imagery from a protected stretch of West African dryland reveals another reality: these seemingly scrubby ecosystems stored substantial carbon and have been gradually losing it.

Dryland under pressure

The site lies in south-west Burkina Faso, West Africa, within the protected Bontioli Natural Reserve savanna.

This 47-square-kilometre refuge sits along the dry fringe of the Sahara. It has protected wildlife since 1957, although pressure from people has persisted throughout that time.

Environmental scientist Issaka Abdou Razakou Kiribou, of Haramaya University in Ethiopia, and colleagues set out to establish what 22 years of pressure had meant for carbon stored in this dryland. Their view came from above.

They used satellite images collected in 2000, 2010 and 2022. Software classified the area into land-cover categories, while established estimates supplied the typical carbon storage for each category.

What the maps showed

Over the 22-year period, the maps reveal one type of land replacing another. Wooded savanna, the denser and more tree-filled cover that formerly occupied around a third of the reserve, fell to about an eighth. Grassland and agricultural land expanded in its place.

Farmland was the most obvious gainer. Cropland rose from less than a tenth of the reserve to over a third, and this change accelerated: after 2010, land conversion occurred more than twice as quickly.

Replacing woody cover with crops is a well-established driver of carbon loss. One analysis places sub-Saharan Africa among the regions in which agricultural clearance has caused the greatest losses.

Measuring carbon decline in Bontioli Natural Reserve

Scientists already understood that converting trees to farmland releases carbon. This research provided a year-by-year record for a West African dryland, showing that carbon stored in the reserve's wooded savanna dropped by almost half.

By 2022, cropland contained the greatest overall carbon store of any land category. This was not because individual fields were carbon-rich, but because cropland covered such a large area: extensive land coverage, with little carbon in each patch.

The estimates were derived from the way vegetation appeared from space, rather than from weighing trees or sampling soil. Overall, the reserve's annual carbon store declined by around a quarter.

Carbon changes by the season

Carbon losses were not evenly distributed across the calendar. Between October and March, as the dry season desiccates the savanna and fires frequently pass through it, the land lost carbon. It regained some once April rains arrived: an annual back-and-forth pattern.

Fire is a natural part of savanna ecosystems, yet burning vegetation releases stored carbon into the atmosphere. Land clearance and burning produce a substantial portion of carbon emissions, with one report estimating that land-use emissions exceed a tenth of the global total.

Carbon distribution also changed across the dryland. It was relatively even in 2000, but by 2022 it had become fragmented into patches. The lowest-carbon areas corresponded with places where agriculture had advanced furthest into tree cover.

A surprising signal

The team then found a result that ran counter to expectations. When they compared plant growth with temperature and rainfall, growth was lower during warmer, wetter periods rather than higher. In a savanna, researchers would usually anticipate the reverse.

The researchers are cautious not to draw too much from the result. Temperature and rainfall together accounted for roughly half of these fluctuations, indicating that something meaningful is occurring. However, they cannot yet explain why warmth and rain might restrict growth.

There are several straightforward possibilities. High temperatures can place plants under stress even where water is available, while rainfall arriving at an unsuitable time may contribute little to sustained growth. Resolving the question will require field research.

Looking towards 2070

To look ahead, the researchers modelled the reserve's climate through to 2070 under both a low-emissions pathway and a high-emissions pathway. Each produced the same overall direction.

The reserve is expected to become warmer, with temperatures rising by up to about 1.1°C, while rainfall decreases.

Greater heat and reduced water availability would mean more carbon leaving the landscape. On the high-emissions pathway, by 2070 the reserve may lose about half again as much carbon as it would under the lower-emissions scenario.

This outcome is not inevitable. The lower-emissions pathway limits the reserve to a slower, more consistent decline. Other African drylands face the same tension between farming and carbon storage, as a study from Ethiopia also found.

What this could change

The study puts a figure on what had previously been an assumption. Within this reserve, the expansion of farmland was associated with a measurable decline in stored carbon, year after year and location by location.

Maps of this kind can identify the parts of a reserve that are losing carbon, allowing restoration teams to focus on the most affected areas instead of estimating where action is needed. A carbon sink once regarded as too dry to warrant protection begins to seem worth safeguarding.

There is also the unresolved finding to investigate. The signal that heat and rainfall reduce greenness rather than support it is one that future dryland climate models should be designed to detect.

For now, the message is clear: a dry, scrub-covered landscape stores more carbon than it appears to, and clearing it returns that carbon to the atmosphere.