The savanna illuminated by dawn sunlight, in Kruger National Park, Limpopo province, South Africa, in September 2022. (Javier Puig Ochoa), CC BYA recent United Nations report has stated the world will likely overshoot the 1.5 C target by 2030 set out in the 2015 Paris Climate Agreement. Tree planting has become one of the most appealing symbols of climate action. It can be easily accounted for, visible and easy to communicate: more trees means more carbon being absorbed, a cooler planet. Everyone gets the point. Global initiatives now promise restoration on a vast scale, including the Bonn Challenge goal of restoring 350 million hectares of degraded and deforested land by 2030.But a tree is not automatically a climate solution, and land without dense tree cover is not necessarily degraded. When tree planting targets natural grasslands, shrublands and savannas, well-meaning efforts can replace one ecosystem with another, reduce water availability and threaten livelihoods. In 2007, the African Union launched the Great Green Wall Initiative; a plan to stop southern expansion of the Sahara Desert by planting trees and restoring landscapes across 22 countries from Senegal to Djibouti. Yet, almost 20 years later, the initiative has had mixed results.What truly works is simple: plant trees for restoration in the right ecosystems.Drylands are active ecosystemsDrylands, including grasslands, shrublands and savannas, are often characterized by what they seem to lack: water, dense vegetation or agricultural potential. Yet they cover about 40 per cent of the Earth’s surface, support billions of people and play an important role in the year-to-year variability of the global carbon cycle — changes in the carbon absorbed and released from one year to the next.Dryland ecosystems have evolved around low and highly variable rainfall, recurrent drought and disturbance. Native grasses, shrubs, scattered trees and biological soil crusts are not signs of an ecosystem needing restoration. They are parts of a functioning biome.That distinction matters because global restoration targets can identify these naturally open ecosystems as candidates for tree cover. Ecologists have warned that planting trees to expand into old-growth grasslands and savannas can damage biodiversity and ecosystems rather than restore them. Planting trees may result in more woody cover, but the landscape itself may register ecological and diversity loss.Why tree planting in drylands can backfireFirstly, trees use water. In some locations, appropriate native trees can provide shade, habitats, food and carbon storage. The problem is large-scale forestation that ignores the water balance. Planting ill-suited trees in water-limited ecosystems can substantially reduce the volume of water flowing through rivers or streams and create a trade-off between carbon storage and water supply. The trade-off can affect groundwater levels, downstream ecosystems, farming and household water security.Secondly, carbon is not the only component to bear in mind. Open vegetation and light-coloured soils often reflect more sunlight than a dark tree canopy. Converting them to forest lowers an area’s albedo — a measure of how much sunlight a surface reflects — and increases how much solar energy is absorbed. Research shows that albedo change can offset a substantial share of the climate benefits expected from additional tree cover, with especially important consequences in some drylands. Counting tonnes of carbon while ignoring this biophysical effect can exaggerate the benefit of tree planting.Forms of restoration that work in drylandsFarmer-managed natural regeneration can protect and manage native trees and shrubs that resprout from existing roots, stumps or seeds. In Niger, farmers’ regeneration of white acacia — a thorny tree native to parts of Africa and West Asia — and other useful species helped regreen approximately five million hectares without relying on mass seedling campaigns. Because farmers select and manage the regrowth, the approach can be adapted to crops, livestock and locally valued species.In Burkina Faso, farmers have also revived and improved zaï, which entails planting pits that collect runoff and concentrate manure or compost around crops. Research on soil and water conservation shows how zaï and related practices can rehabilitate degraded land and improve production under difficult rainfall conditions. Here, restoration is not defined as making the landscape look like a forest. It is defined by recovering soil function, vegetation and livelihoods.Other places may require protecting native grass cover, managing grazing pressure, restoring seasonal mobility, controlling invasive species, conserving biological soil crusts or assisting the recovery of locally appropriate shrubs and trees. No single technique works for all; you just need to do what works for your situation. That is precisely the point.Focus on restoring ecosystems, not planting treesTree planting can be valuable in deforested ecosystems and where locally appropriate species, water availability, land tenure and long-term ownership align. Growing trees alongside agricultural crops or livestock and allowing plants to naturally regenerate can also be effective in dryland farms. The mistake is turning those successes into a universal rule that equates more trees with a better climate.Governments and donors should ask whether a project restores the natural ecosystem, protects water and retains native biodiversity before campaigning for tree planting.Drylands are not failed landscapes. They are distinctive ecosystems that store carbon, sustain biodiversity and support people. We must not force forests where they do not belong and need to start valuing the life already adapted to thrive there.Joy Kehinde Adebisi receives funding from Western University Graduate Research Funds.