Recent decline in the global land evapotranspiration trend due to limited moisture supply

• Published in: Nature (London) 2010-10, Vol.467 (7318), p.951-954
• Main Authors: Jung, Martin, Reichstein, Markus, Ciais, Philippe, Seneviratne, Sonia I., Sheffield, Justin, Goulden, Michael L., Bonan, Gordon, Cescatti, Alessandro, Chen, Jiquan, de Jeu, Richard, Dolman, A. Johannes, Eugster, Werner, Gerten, Dieter, Gianelle, Damiano, Gobron, Nadine, Heinke, Jens, Kimball, John, Law, Beverly E., Montagnani, Leonardo, Mu, Qiaozhen, Mueller, Brigitte, Oleson, Keith, Papale, Dario, Richardson, Andrew D., Roupsard, Olivier, Running, Steve, Tomelleri, Enrico, Viovy, Nicolas, Weber, Ulrich, Williams, Christopher, Wood, Eric, Zaehle, Sönke, Zhang, Ke
• Format: Article
• Language: English
• Subjects: 704/106/242; 704/106/694; 704/844/2739; Animal and plant ecology; Animal, plant and microbial ecology; Artificial Intelligence; Atmosphere; Atmosphere - chemistry; Biological and medical sciences; Climate change; Climate effects; Climate variability; Climatology; Climatology. Bioclimatology. Climate change; Earth Sciences; Earth, ocean, space; Ecosystem services; El Nino; Environmental aspects; Environmental Sciences; Estimates; Evapotranspiration; Exact sciences and technology; External geophysics; Fresh Water - analysis; Fundamental and applied biological sciences. Psychology; Global Changes; Global climate; Global Warming - statistics & numerical data; History, 20th Century; History, 21st Century; Humanities and Social Sciences; Humidity; Hydrologic cycle; letter; Meteorology; multidisciplinary; Plant Transpiration - physiology; Remote sensing; Reproducibility of Results; Science; Science (multidisciplinary); Sciences of the Universe; Seasons; Soil - analysis; Soil moisture; Solar energy; Synecology; Temperature; Terrestrial ecosystems; Trends; Uncertainty; Volatilization; Water Cycle
• ISSN: 0028-0836; 1476-4687
• DOI: 10.1038/nature09396

A question of water supply An acceleration of the global hydrological cycle, evapotranspiration included, is regarded as a key indicator of the impact of global warming on Earth's system. Evapotranspiration refers to the water that moves from Earth's land surface to the atmosphere through the combined effects of evaporation and plant transpiration. Martin Jung and colleagues use a data-driven machine-learning technique and a suite of process-based models to show that, between 1982 and 1997, evapotranspiration increased steadily with global warming. But since 1998, the increasing trend has flattened, probably as a result of limitations in soil-moisture supply in the Southern Hemisphere — particularly Africa and Australia. It remains to be seen whether this is part of a natural climate variation or a climate-change signal in which land evapotranspiration becomes more supply-limited in the long term. Climate change is expected to intensify the global hydrological cycle and to alter evapotranspiration, but direct observational constraints are lacking at the global scale. Now a data-driven, machine-learning technique and a suite of process-based models have been used to show that from 1982 to 1997 global evapotranspiration increased by about 7.1 millimetres per year per decade. But since 1998 this increase has ceased, probably because of moisture limitation in the Southern Hemisphere. More than half of the solar energy absorbed by land surfaces is currently used to evaporate water 1 . Climate change is expected to intensify the hydrological cycle 2 and to alter evapotranspiration, with implications for ecosystem services and feedback to regional and global climate. Evapotranspiration changes may already be under way, but direct observational constraints are lacking at the global scale. Until such evidence is available, changes in the water cycle on land—a key diagnostic criterion of the effects of climate change and variability—remain uncertain. Here we provide a data-driven estimate of global land evapotranspiration from 1982 to 2008, compiled using a global monitoring network 3 , meteorological and remote-sensing observations, and a machine-learning algorithm 4 . In addition, we have assessed evapotranspiration variations over the same time period using an ensemble of process-based land-surface models. Our results suggest that global annual evapotranspiration increased on average by 7.1 ± 1.0 millimetres per year per decade from 1982 to 1997. After that, coin