From Drought Propagation to Dryland Expansion: The Role of Land Feedbacks in Spreading Aridity

Half a century after Apollo 17's iconic "Blue Marble" photograph, depicting our Earth's life-sustaining hydrosphere, concerns about the future of this hydrosphere have intensified. Climate change is thought to be shifting ecosystems toward drier and more hostile conditions over land, threatening biodiversity and human resilience. Global warming accelerates evaporation, yet precipitation trends remain uncertain, leading to projections of overall desertification and dryland expansion. The IPCC highlights potential catastrophic risks, especially for subhumid ecosystems, and stresses the urgent need for understanding the mechanisms driving this dryland expansion. However, our time series of reliable observations are not sufficiently long to study this slow, creeping process at the global scale with sufficient accuracy. To bridge this knowledge gap, we propose to study the parallels between short-term drought propagation and long-term dryland expansion, hypothesising that the physical mechanisms underlying both are the same.

Specifically, we focus on a critical feedback from drying soils that has proven crucial for drought spatiotemporal propagation: as prolonged dry events decrease land evaporation, both atmospheric humidity and the likelihood of rainfall are further reduced. Simultaneously, drying soils release more sensible heat into the atmosphere, amplifying temperatures, reducing rainfall efficiency and often triggering compound heatwaves. Together, these feedbacks perpetuate drought conditions, reducing rainfall, both locally and downwind, and thus exacerbating droughts' spatial and temporal extent. Using a Lagrangian atmospheric model and four decades of reanalysis data, we confirm that droughts and heatwaves can self-propagate through these land–atmosphere interactions.

Interestingly, this same process may also drive dryland self-expansion over multi-decadal periods. Our findings suggest that nearly half of the 5.2 million km² of humid land that became drylands in the past four decades did so due to dryland self-expansion via land–atmospheric feedbacks. Existing drylands warmed and dried the air flowing towards downwind subhumid regions, decreasing rainfall and increasing potential evaporation there, causing their eventual transition into drylands. These results may help in predicting the broad impacts of dryland expansion, including disruptions to carbon sequestration, nutrient cycling, and land productivity. Identifying self-expansion hotspots enables targeted interventions in land-use and ecosystem management to mitigate dryland growth. Conservation in upwind drylands can slow down this process, while prioritizing vulnerable downwind regions for strategies like restoring vegetation and soil health can preserve their biodiversity and curtail their aridification. Furthermore, our findings highlight the need for improved climate models to predict future ecosystem transitions and emphasize the relevance of land feedbacks to understanding paleoclimatic tipping points.