Restoration Flagships Need Comparable Evidence, Not Comparable Headlines

Three initiatives in Brazil, Saudi Arabia and the Sahel became United Nations World Restoration Flagships on 26 August. Together they commit to restore almost five million hectares by 2030. The recognition gives the initiatives visibility within the UN Decade on Ecosystem Restoration, but it also creates an analytical risk: three unlike programs can be compressed into one hectare total and one story of success.

Their differences are more useful than their sum.

Three restoration systems

Brazil’s Araticum Initiative brings more than 180 organizations together around the Cerrado, the biodiverse savannah that supplies water to eight of the country’s twelve major river basins. The program uses natural regeneration, direct seeding, native-species restoration and agroforestry. UNEP reports that more than 27,000 hectares had been placed under restoration by June 2026, against an ambition of two million hectares by 2030.

The Sahel Integrated Resilience Initiative works across Burkina Faso, Chad, Mali, Mauritania and Niger. Led by the World Food Programme, it combines land rehabilitation, soil and water management, livelihoods, nutrition and school meals. More than 335,000 hectares have been rehabilitated since 2018, while the program reports reaching more than four million people.

Saudi Arabia’s National Greening Program spans rangelands, forests, deserts, cities and coasts. UNEP reports one million hectares brought under restoration and 159 million trees planted by June 2026, against 2030 targets of 2.5 million hectares and 215 million trees. Its methods include natural regeneration, controlled grazing, wildlife reintroduction, rainwater harvesting and treated-wastewater use. Inclusion alongside the Cerrado and Sahel makes the comparison especially demanding: vegetation, water availability, land use, governance and the meaning of ecological recovery differ sharply across the three settings.

The flagship label does not erase those differences. It should make them easier to examine.

A hectare is a location, not yet a result

Restoration programs use hectare figures for several stages: land targeted, land enrolled, land treated and land showing improved condition. Those measures are not interchangeable. Brazil’s reported 27,000 hectares placed under restoration describe activity at an early fraction of its 2030 ambition. The Sahel’s 335,000 rehabilitated hectares reflect a longer operating record. A national pledge can define direction while saying little about the ecological state of any particular site.

The intervention also changes what a hectare means. Natural regeneration may involve protecting an area from pressure and allowing recovery. Direct seeding involves material inputs and follow-up. Agroforestry combines productive use with ecological functions. Half-moon earthworks or water-retention structures in the Sahel alter how scarce rainfall moves through degraded land. Counting all of them is legitimate only if the underlying categories remain visible.

Condition evidence is the next layer. The World Food Programme reports that satellite monitoring found improved soil and vegetation conditions at 75 percent of its resilience sites. That is more informative than an activity total because it observes a change associated with program locations. It still requires careful interpretation: the result concerns monitored sites, depends on the indicator and period used, and does not establish that every hectare has recovered in the same way.

Human outcomes belong inside the restoration record

Nature-based solutions are durable when institutions and livelihoods support the ecological mechanism. The Cerrado program depends on Indigenous Peoples, local communities, farmers, researchers, governments and seed networks. Its collection of more than 150 native plant species is not merely a biodiversity detail. It represents supply capacity for restoration and knowledge about what can be established where.

In the Sahel, land rehabilitation is connected to food production, school meals, market access and social cohesion. WFP reports that the share of participating households requiring humanitarian food assistance between harvests fell from 44 percent in 2020/2021 to 33 percent in 2024/2025. The finding does not prove that restoration alone caused the change; the program deliberately combines several interventions. It does show why ecological and social evidence should be interpreted together without collapsing one into the other.

Stewardship determines whether gains persist. Restored vegetation can be lost if grazing, cultivation, fire, tenure conflict or funding changes. Monitoring therefore needs to follow not only cover and soil condition but also who has authority to maintain the land, how benefits are distributed and whether local rules survive after external support recedes.

Recognition should create a learning system

UNEP and FAO say flagship progress will be monitored through the Framework for Ecosystem Restoration Monitoring. The opportunity is to create comparison without false uniformity. A useful record would identify the baseline, intervention, land status, ecological indicators, social outcomes, finance and maintenance responsibility for each program.

That structure would allow decision-makers to ask precise questions. Which approaches recover vegetation under particular rainfall and land-use conditions? Where does natural regeneration outperform active planting? How do seed systems constrain scale? Which livelihood arrangements reduce renewed degradation? What monitoring can be maintained by national and local institutions after program finance changes?

The answers will not produce a single restoration recipe. They can produce transferable judgment about conditions and failure modes.

World Restoration Flagships are intended to demonstrate that recovery can operate at scale. Recognition is valuable when it directs attention and investment toward initiatives with serious records. It becomes misleading if a target is reported as restored land, if early vegetation change is treated as durable ecosystem function, or if social participation is counted without examining authority and benefit.

The three new flagships should therefore be read as distinct restoration systems entering a shared learning framework. Their credibility will grow not when their figures look alike, but when each figure can be traced to what changed on the ground.