Satellite data collected over more than 40 years show that dryland ecosystems — which cover about 40% of Earth's land and support over 2 billion people — are experiencing more extreme fluctuations in vegetation. A study published in Nature Climate Change found that while rising levels of atmospheric carbon dioxide have boosted plant growth during wet years, creating a long-term "greening" trend, this trend hides dangerous year-to-year variability that makes ecosystems more vulnerable during dry periods. Researchers from the University of Arizona, led by Wen Zhang, analyzed satellite data to track changes in the leaf area index — a measure of vegetation activity and productivity. Their findings reveal that about 80% of the world's drylands are becoming more unstable, with the extremes of vegetation growth becoming more pronounced over time. The study suggests that this instability is likely due to the combined effects of higher atmospheric carbon dioxide and unpredictable rainfall patterns. While elevated CO2 levels can make plants more efficient in using water, leading to increased leaf growth, this also means that larger vegetation requires more resources, making plants more susceptible to stress during dry periods. The implications of this instability are significant for both agriculture and ecology in dryland regions. Rain-fed farming and pasture forage production — which are vital for livestock and rangeland management — are becoming more unpredictable. Bill Smith, a senior author of the study, explained that rangeland managers, who depend on stable forage availability for planning, are facing new challenges. David Moore, another co-author, pointed out that the large swings in plant productivity could be an early warning of broader ecological changes, signaling that dryland ecosystems are under stress and losing their ability to recover from disturbances. The study also evaluated 13 major global vegetation models and found that none accurately predicted the observed increase in year-to-year variability in dryland vegetation. These models generally assume that dryland plants respond to changes in CO2 and rainfall in predictable ways, but they fail to account for how plant responses are changing over time. This gap in modeling has major consequences for predicting how ecosystems will respond to future climate change. Smith emphasized that without accurate models, projections of global ecological changes over the next 50 years remain uncertain. The study aims to encourage new research focused on improving the representation of drylands in Earth system models to better understand and predict future environmental changes.