A study published in Nature in April 2026 puts biodiversity loss in sub-Saharan Africa at 24%. This estimate rests on the bii4africa dataset, fed by the expertise of 200 specialists for approximately 50,000 species of terrestrial vertebrates and vascular plants—a fraction of the continent’s likely several hundred thousand species. Measuring decline is one thing; knowing whether the measure says something accurate about what is being measured is another.

The essentials

  • The estimated 24% loss is calculated from the bii4africa dataset, based on the expertise of 200 specialists for approximately 50,000 species of terrestrial vertebrates and vascular plants, a fraction of the continent’s actual diversity.
  • In South Africa, the stewardship mechanism accounts for roughly 90% of the terrestrial expansion of protected areas since the Protected Areas Act was enacted, with 2.76 million hectares secured across 469 sites at the national scale.
  • The DRC holds a significant share of Africa’s biodiversity but is underrepresented in international databases like GBIF (Global Biodiversity Information Facility), which can limit the information available for conservation policy.
  • Conservation investments and policies tend to concentrate where data are available and documented, rather than according to ecological priorities alone.
  • Strengthening local data collection capacity, particularly in poorly documented regions, improves the reliability of global biodiversity metrics.

A global metric built on unevenly distributed data

This selection bias operates upstream of any calculation: data gaps can skew indicators by region and taxon. Global metrics can be limited by unequal data representation. Insufficient coverage of certain groups complicates assessment of their trends.

The figure of 24% strikes hard. It circulates in reports, diplomatic speeches, Kunming-Montreal commitments. But one must understand what it measures: the relative abundance of assessed species in their habitats, compared to a pre-industrial reference state. The calculation integrates regional lists of native taxa, spatial data on land use, and expert estimates of species responses.

The study mobilizes the knowledge of 200 experts for approximately 50,000 species of terrestrial vertebrates and vascular plants. This is substantial as a scientific effort. It is insufficient as a basis for claiming to measure the ecological health of a continent of 30 million km², the richest in biological diversity after the tropical forests of South America. Scientific estimates suggest a far higher number of species in sub-Saharan Africa when counting invertebrates, fungi, and microorganisms for which data remain limited.

The assessed species are not chosen at random. They are vertebrates, vascular plants, a few emblematic groups of invertebrates. Species for which researchers have published data, species for which institutions have funded inventories. The selection bias is mechanical: one assesses what one knows, one knows what one has the means to observe.

South Africa accounts for 90% of protected area expansion

The South Africa Biodiversity Stewardship Programme illustrates what institutional capacity enables. Since its launch, it has registered 2.76 million hectares of newly protected areas across 469 sites, involving private landowners and communities in long-term management. The approach is documented, sites are georeferenced, data are fed into international systems.

The programme accounts for approximately 90% of the terrestrial protected area expansion recorded in South Africa since the enactment of South African conservation law. This does not mean the rest of the continent protects nothing. It means other regions have less developed structures to record and report their actions.

Botswana, Namibia, Kenya have functional monitoring systems. Mozambique, the DRC, the Central African Republic operate under very different conditions. Forest rangers there are often few in number, inadequately equipped, active in areas where physical access remains difficult for several months a year. Reporting a patrol, let alone filling out an international database, requires infrastructure these countries do not possess at the necessary scale.

The DRC: 15 to 20% of African biodiversity, less than 2% of the data

The Congo forest is the planet’s second-largest lung. It harbors gorillas, okapis, Congo peafowl, but also thousands of species of trees, insects, and frogs that science has not yet named. The DRC holds a considerable share of African biodiversity.

In GBIF, the global platform that aggregates species occurrence data, the DRC is underrepresented relative to its ecological importance. The gap between ecological importance and presence in global data is not abstract. It has direct consequences for species distribution models, for the calculation of biodiversity loss indices, for conservation priority maps that donors use to allocate funding.

An ecosystem absent from the data risks being poorly accounted for in conservation decisions. The Clements et al. study shows that gaps in global data can limit models applied to Africa and proposes an approach based on local expertise. Where institutions are strong, data exist. Where they are weak, data are missing.

Conservation investments and policies tend to concentrate where data are available and documented, rather than according to ecological priorities alone. The same structural bias appears in other environmental domains: poorly measured carbon sinks are also poorly protected carbon sinks, as illustrated by the situation of European forests.

Contributions of the Clements et al. study, despite its limits

It would be inaccurate to reduce this study to its shortcomings. It is precisely because Clements and his coauthors mobilized the knowledge of 200 experts that their own analyses can point out the blind spots. The value of this work is twofold: it produces an estimate that is methodologically sound, and it makes visible the uncertainty surrounding that estimate.

The methodology of Nature imposes transparency on margins of error that major political announcements often erase. When a UNEP report cites a biodiversity loss figure, it generally does not cite the confidence interval nor the species absent from the calculation. The Clements study documents its own limits. This is a scientific rigor that policymakers have every interest in reading closely.

Sub-Saharan Africa has lost forest cover, wetlands, savanna. The species for which data are available show downward trends. The figure of 24% reflects a reality, even if imperfectly. The estimate of 24% carries uncertainties and could be affected by unquantified biases, without evidence that these necessarily lead to underestimation.

Monitoring as conservation infrastructure

This logic makes monitoring a precondition for protection, not an ancillary tool. Without sufficient presence data, prioritization models are less reliable, but other structured knowledge can help identify priorities. Strengthening local collection capacity thus acts upstream of any conservation decision: it can improve how ecosystems are considered in conservation trade-offs.

The practical issue is clear: improving data collection in central Africa and in the least-covered areas improves the relevance of any continental strategy. This strengthening takes several concrete forms.

African herbaria and natural history museums are an underexploited resource. Millions of specimens collected since the nineteenth century remain undigitized, stored in underfunded institutions. Mass digitization programs, funded by international donors, would make it possible to integrate several decades of historical data into GBIF within years.

Citizen science is advancing on the continent. Species survey applications operate in Kenya, Tanzania, South Africa. Extending these tools to less connected areas, training local agents in their use, represents a moderate-cost lever compared to large traditional scientific expeditions.

Partnership programs between African universities and international institutions are beginning to produce inventories in areas previously blank on maps. The University of Kinshasa and several specialized NGOs are working on herpetological inventories in the Congo Basin, with support from the Mohamed bin Zayed Species Conservation Fund. These efforts remain insufficient in volume, but they show that a decentralized model, rooted in local capacity, works.

The South Africa Biodiversity Stewardship Programme itself rests on a principle worth exporting: associating private landowners, local communities, and public authorities in formalized management agreements, with technical support from the state for documentation. This model has transformed South Africa’s capacity to report its advances to international systems. It could inspire adapted programs in other contexts.

Calibrating global conservation to the right priorities

The Kunming-Montreal framework sets a target of 30% of terrestrial and marine surfaces protected by 2030. This quantified target creates political pressure and measurable commitments. It presents a structural limit: countries with more developed administrative capacity can better record and report their protected surfaces.

The figure of 90% describes the share of stewardship in terrestrial protected area expansion in South Africa without allowing one to conclude there is a continental concentration of funding. This is a systemic bias that institutions like the GEF (Global Environment Facility) and major biodiversity funds have begun to recognize. The Kunming-Montreal Framework Fund includes preferential financing mechanisms for low-capacity countries, but their implementation remains underway.

The study shows that gaps in contextual data limit assessments and proposes a method based on local knowledge to better inform decisions. Several teams are working on Bayesian modeling methods to estimate the probable distribution of species in unsampled areas, based on data from better-documented similar areas. These estimates do not replace field data, but they make it possible to map blind spots and guide inventory priorities.

By 2035, the date for revising the global biodiversity framework, the reliability of African metrics will be tied to investments made to strengthen monitoring in poorly documented areas. Regions underrepresented in global databases risk being less well accounted for by policies based on these data.

The precision of tools—species distribution models, high-resolution satellite imagery, environmental DNA sequencing—is advancing fast. These technologies are beginning to reach prices accessible to African institutions. But technology alone does not bridge a deficit in human and institutional capacity. Training taxonomists, funding herbaria, connecting African universities to global data networks: this is what current analyses of African biodiversity make visible, in gaps.

Sources

  1. Daily Maverick, “From discovery to impact: mapping Africa’s biodiversity” (April 2026)
  2. Clements et al., Nature, April 2026, Study on biodiversity loss in sub-Saharan Africa (37,000 IUCN species)
  3. Global Biodiversity Information Facility (GBIF), African species occurrence data
  4. IUCN Red List, assessments of African species
  5. South Africa Biodiversity Stewardship Programme, data on protected area expansion (2.76 M ha, 469 sites)
  6. Mongabay, March 2026, report on biodiversity data gaps in central Africa