Climate scientists compare global water cycle models for the first time

A team of international researchers has completed a coordinated comparison of climate models that include water isotope data, finding that combining results from several models gives a closer match to real-world observations than any single model on its own.

Water isotopes are versions of the water molecule that contain heavier forms of hydrogen and oxygen. Scientists use these variations as markers to track how water moves through the atmosphere and how it changes phase from vapour to rain or snow. Over the past twenty years, research groups around the world have developed climate models that simulate these isotopes, but differences in the way the models were set up made it hard to compare their results directly.

To address this problem, the researchers launched the Water Isotope Model Intercomparison Project, known as WisoMIP. Eight advanced isotope-enabled climate models were run using the same atmospheric circulation data from the ERA5 reanalysis, and identical sea surface temperatures and sea ice conditions, for the period from 1979 to 2023. This uniform setup allowed the team to compare how each model simulated the three-dimensional distribution of water isotopes day by day.

The analysis showed that when the output from all eight models is averaged, the combined result more closely matches observed isotope patterns in precipitation, atmospheric water vapour, and snow than any individual model. The average also reproduced the broad patterns of oxygen isotopes in precipitation and identified specific regions where models still differ widely from each other.

Because water isotope signatures are preserved in ice cores, coral skeletons and tree rings, and because they can be measured in modern precipitation and atmospheric water vapour, the results help bridge observational data with climate model simulations. The dataset from WisoMIP sets a reference point that scientists can use to evaluate and improve isotope-enabled climate models, which may reduce uncertainty in future climate assessments.

The research was published in the Journal of Geophysical Research: Atmospheres by Hayoung Bong and colleagues, with support from institutions including the Institute of Industrial Science at the University of Tokyo and research agencies in Japan, Norway and the United States.