Research centre established by:

KTH, Stockholm University and RISE
Diagram showing green and blue water fluxes: left side illustrates water movement in a landscape; right side shows world maps indicating regions of green water flux (top) and blue water flux (bottom) related to global water balance.

One Earth – Global patterns in water flux partitioning: Irrigated and rainfed agriculture drives asymmetrical flux to vegetation over runoff Graphical

A new study in the scientific journal ‘One Earth’ uses Machine Learning to model and show how the balance of freshwater flows varies across the Earth’s land surface. The results can be used to plan our future land and water use. The balance is particularly vulnerable under current and future climate change, as well as increasing food needs for a growing population.

Diagram shows green and blue water fluxes. Left: a mountain cross-section with green (vegetation) and blue (water bodies) arrows, labelled water balance. Right: world maps of green and blue water flux distribution.

The paper “Global patterns in water flux partitioning: irrigated and rainfed agriculture drives asymmetrical flux to vegetation over runoff,” by Digital Futures faculty member Gia Destouni and postdoc Daniel Althoff at Stockholm University, has just been published in One Earth.

Water on land is partitioned into a green flux that mainly sustains terrestrial ecosystems and societal food security and a blue flux that mainly sustains aquatic ecosystems and societal water security. The study shows that the green flux is mostly greater than the blue around the world. Expanding crops and irrigation to feed a growing human population will shift more blue water into green, exacerbating the blue water vulnerability to future climate change.

“These results are important to better plan and adapt our land use and balance between different freshwater flows to meet different water needs in society,” says Daniel Althoff from the Department of Physical Geography, Stockholm University.

Picture of Gia Destouni

“The global machine learning model we have developed in the study is a promising tool for predicting the availability of blue and green water under different future climate and land use scenarios around the world,” says Gia Destouni, Professor of Hydrology, Hydrogeology and Water Resources at the Department of Physical Geography, Stockholm University.

Gia is also an Affiliated Professor of Engineering Hydrology at KTH Royal Institute of Technology and the Editor-In-Chief of Water Resources Research at the American Geophysical Union.

Althoff D., Destouni G., Global patterns in water flux partitioning: Irrigated and rainfed agriculture drives asymmetrical flux to vegetation over runoff, One Earth, 6(9), 1246-1257, 2023. https://doi.org/10.1016/j.oneear.2023.08.002

Link to Graphical abstract, short description, highlights and Science for Society summary

Find out more in this article at Stockholm University: https://www.su.se/institutionen-for-naturgeografi/nyheter/viktigt-att-vi-lär-oss-balansera-mellan-olika-behov-av-sötvatten-1.674086

Contact:
Daniel Althoff
daniel.althoff@natgeo.su.se

Georgia (Gia) Destouni
georgia.destouni@natgeo.su.se 

More news

09/10/2026

How and why the Nobel Prize is awarded: Sten Grillner shares insights from neuroscience A...

From the classroom to the wider public: Digital Futures welcomed visitors during Nobel Calling 2026

09/10/2026

On 7 October, Digital Futures opened its doors to school students and the general public,...

A panel of four people seated on a stage at the Nobel Prize Museum in front of an audience.

Does AI make us stupid? Experts explore how AI is changing human abilities

06/10/2026

As artificial intelligence becomes part of research, education, workplaces and everyday life, an important question...

French business delegation explores Sweden’s collaborative innovation model

02/10/2026

On 29 September, Digital Futures welcomed Business Sweden and a delegation of French industrial leaders...