.

Drylands are defined as regions where rainfall is low relative to the atmosphere’s ability to draw water from soil and plants. Expanding boundaries of these regions are frequent concerns in climate change considerations. While they can support large populations and ecosystems, dryland expansion can worsen water scarcity, land degradation and desertification.

Many projections of future dryland expansion assume warming simply raises evaporation, but a new study suggests the process is more complex. The study, published in Environmental Research Letters, incorporates the reduced evapotranspiration that plants exhibit under increased carbon dioxide (CO₂) levels and finds that previous estimates of dryland expansion may be off.

Estimating current and future dryland area

Previous estimates of Earth’s total dryland area have varied, ranging from roughly 37% to 47% of land area, depending on the data and methods used. These estimates typically use the aridity index, which is defined as the ratio of precipitation to potential evapotranspiration (PET), where PET is the maximum amount of water that would evaporate from the soil and transpire from plants if an unlimited water supply were available. The data used to calculate rainfall and PET vary among studies, however.

The authors of the new study write, “Many of these assessments, however, relied on climate data prior to 2010, whereas global climate conditions have changed substantially over the past decade. Recent drought intensification and land-atmosphere feedbacks further suggest that some regions may already have shifted toward drier conditions. Because dryland boundaries are defined by aridity index (AI), which integrates precipitation and PET, changes in water supply or atmospheric evaporative demand can alter the spatial extent of drylands. Therefore, the current spatial distribution of global drylands needs to be reassessed using updated climate data.”

The issue extends to future assessments as well. The authors note that projected dryland expansion in conventional aridity-index studies has varied from about 4% to 10% and 11% to 23% of land area under different greenhouse gas scenarios. However, these models usually leave out changes in plant evapotranspiration under high CO₂ levels.

Weaker expansion when including plant responses to CO₂

The researchers sought to better quantify the current and future distribution of global drylands and identify hotspots of increasing drying and wetting under three different emissions scenarios by incorporating plant responses to CO₂. They combined historical climate observations with projections from 19 major climate models. They used multiple classes of dryland, from semi-arid to hyper-arid, and compared a standard evaporation calculation with one adjusted for the effects of rising CO₂ on plant water use.

Their results indicated that drylands covered 38.58% of the world’s land, excluding Antarctica, during the period 1994–2023, with semi-arid zones accounting for the largest percentage of global drylands. Hyper-arid zones occurred in Algeria, Libya and Saudi Arabia. The team also says global dryland area increased overall from 1960 to 2023, though the pace slowed in more recent decades.

Projected spatial distribution of global drylands and associated climate zones under different emission scenarios. Credit: Environmental Research Letters (2026). DOI: 10.1088/1748-9326/ae9b29

The new future projections indicate that even under the highest-emissions pathway, drylands are projected to cover about 40.28% of land by 2100, which corresponds to around 2.37 million square kilometers (915,000 square miles). Under the lowest-emissions pathway, the model shows a dryland cover of 39.31%, or 1.09 million square kilometers (421,000 square miles). Leaving out CO₂ plant effects raises the projected dryland share from 40.28% to 42.11%—a difference of about 2.44 million square kilometers (942,000 square miles) under a high-emissions scenario. Australia, eastern Brazil and parts of northern Africa stand out as persistent or emerging drying risk regions, while parts of northern and northwestern China show continuing wetting trends.

Net changes vs. regional changes

Although the net global dryland expansion is predicted to be more limited in this model, the study found substantial spatial reorganization of drying and wetting hotspots across global drylands. For example, under the SSP585 emissions scenario, transitioning areas shifting between drying and wetting hotspots reach 13.0% of land area, in which drying-to-wetting transitions actually dominate, covering 11.4%.

“The limited net increase in global dryland area does not imply weak regional change. Under SSP585, the global dryland fraction increases by only 1.7 percentage points relative to the 1994–2023 baseline, but climate-zone transitions affect 14.75% of the global terrestrial area. Changes in total dryland extent summarize only the net outcome of spatial gains and losses, whereas climate-zone redistribution captures shifts in dryland boundaries and subtype transitions. Thus, a limited net change in global dryland area can coexist with marked regional reorganization,” the study authors explain.

The team notes that climate-model disagreement is greatest at the edges of drylands, where small changes can shift a region between categories, and that it does not fully include irrigation, groundwater pumping, land-use change or other human water demands. Future studies may improve estimates by testing multiple evaporation methods and more detailed land-atmosphere models and combining climate projections with vegetation change, soil moisture, groundwater, irrigation and land management.

..

.