Multi-source analysis of recent changes in global terrestrial evapotranspiration

With rising temperatures, we can expect significant changes in the terrestrial water cycle (TWC).  Evapotranspiration (ET) represents a significant component of the TWC, linking the water, energy, and carbon cycle of the land and atmosphere. Although recent reviews showed that ET has been increasing and even accelerating since the 1980s due to an increase in the LAI (Yang et al., 2023), some studies suggest that ET might be declining (Kim et al., 2021). So, is the increase of ET product dependent? Where in the world do we find disagreement?

We processed 13 global ET products derived from reanalysis, remote sensing, synthesis, and land surface models thereby representing a wide variety of available data. For 2000-2019, we analyzed the ET slope per grid (0.25 deg 0.25 deg), and meaningful regions including biomes, land cover classes, Koeppen-Geiger regions, elevation classes, evaporation quantiles, and IPCC reference regions. Using indices for dataset similitude and concurrence, we created probability maps, which allow us to pinpoint hotspots of uncertainty and regions with a high likelihood of change.

We confirm that ET has increased for 37 % of the terrestrial land from 2000-2019. However, the direction of change in ET for 36 % of the global land area was uncertain with various products showing significant (p > 0.05) negative and positive trends. The spatial distribution of uncertainty varies greatly spatially. For example, over 60% of the area of mangroves and tropical/subtropical moist broadleaf forests and over 40 % of the area of tropical/subtropical, flooded, and montane grass- and shrublands resulted in uncertain ET changes. Some IPCC reference regions (NWS, CAF, SAM, and NSA) resulted in over 70 % of the area in uncertain ET changes. This indicates that estimating changes in ET is still product dependent.

By pinpointing the regions in which the ET products disagree on the magnitude and direction of change, we can lay the ground for the further improvement of TWC estimates.  On the other hand, dataset consensus can help to increase the credibility of hydrological and climate model evaluations and attribution studies. Overall, there is an urgent need to further constrain ET.   

 

Kim, S., Anabalón, A., & Sharma, A. (2021). An Assessment of Concurrency in Evapotranspiration Trends across Multiple Global Datasets. Journal of Hydrometeorology, 22(1), 231–244. https://doi.org/10.1175/JHM-D-20-0059.1

Yang, Y., Roderick, M. L., Guo, H., Miralles, D. G., Zhang, L., Fatichi, S., Luo, X., Zhang, Y., McVicar, T. R., Tu, Z., Keenan, T. F., Fisher, J. B., Gan, R., Zhang, X., Piao, S., Zhang, B., & Yang, D. (2023). Evapotranspiration on a greening Earth. Nature Reviews Earth & Environment, 4(9), Article 9. https://doi.org/10.1038/s43017-023-00464-3