What is it about?
This study revisits a long standing question about West Africa's "forest islands", small, dense patches of forest sitting inside open savanna, and asks whether their soil and vegetation are genuinely different from the surrounding land, or just appear that way depending on how closely you look. Working across paired forest-island and savanna sites in Nigeria, Ghana, and Burkina Faso, the researchers took two different approaches at once: comparing just the topsoil (0–20 cm, the depth most studies stop at), and comparing full soil profiles down through deeper layers. The results depended heavily on which approach was used. Looking only at topsoil, forest islands had higher average organic carbon and nitrogen than savannas, but the difference wasn't statistically solid. Looking at the full depth profile told a clearer story: forest island soil held significantly more organic carbon at every depth, though carbon dropped off with depth at a similar rate in both ecosystem types, especially past 50 cm. A broader statistical analysis found that ecosystem type explained about 13% of the overall variation in soil chemistry, modest but real. Vegetation told an even clearer story than soil: forest islands had noticeably larger tree stems but lower species diversity than the savanna, though how many trees and how much total plant material occupied the land were similar between the two. Curiously, soil properties and vegetation patterns barely tracked each other; only soil pH showed a real relationship, and even that was a negative one with plant diversity.
Featured Image
Why is it important?
A lot of prior forest island research, and a fair amount of tropical soil science generally, relies on topsoil only sampling because it's faster and cheaper. This study is a useful methodological check on that habit: it shows that topsoil alone would have led to the conclusion that forest islands and savannas aren't meaningfully different, while a full profile analysis reveals a real, consistent difference in carbon storage. That distinction matters for carbon accounting and conservation prioritization, since underestimating a forest island's true soil carbon value could mean underestimating its conservation worth. It's also a useful corrective for anyone assuming soil chemistry and vegetation structure move in lockstep; this data shows they mostly don't, which has implications for how these two types of ecological surveys should (or shouldn't) be used as stand-ins for one another
Perspectives
This paper picks up a thread I've been thinking about since our earlier forest island mineralogy work: two ecosystems can look statistically similar at the surface and still be genuinely different once you account for depth. The fact that topsoil alone would have understated the carbon difference is a finding I want to make sure gets attention, because it's a caution against a sampling shortcut that's extremely common in our field, including in some of our own earlier surveys. What stands out most to me, though, is the weak link between soil chemistry and vegetation structure. I would have expected richer, more carbon-dense soil to track more closely with tree diversity or density, and it largely didn't. That disconnect is worth digging into further, possibly through root-zone sampling or microbial community data, since it suggests something other than bulk soil fertility, perhaps historical land use, seed dispersal, or fire regime, is doing more work in shaping these forest islands than the soil chemistry alone can explain.
Dr Samuel Ayodele Mesele
International Institute of Tropical Agriculture
Read the Original
This page is a summary of: Forest islands in West African savannas: multi-scale evidence of soil and vegetation differentiation across soil profiles and ecosystems, Scientific African, June 2026, Elsevier,
DOI: 10.1016/j.sciaf.2026.e03425.
You can read the full text:
Contributors
The following have contributed to this page
