Water Isotope Model Intercomparison Project (WisoMIP)

WisoMIP is a coordinated intercomparison of water isotope-enabled atmospheric general circulation models (AGCMs). It is the first project in which multiple isotope-enabled AGCM simulations have been designed and conducted under a common experimental protocol.

By combining simulations from multiple models, the WisoMIP ensemble helps identify robust features of the global water cycle and can provide more reliable estimates than any individual simulation.

WisoMIP is the successor to the Stable Water Isotope Intercomparison Group (SWING). WisoMIP Phase 1 serves as the continuation of the SWING intercomparison effort and replaces the originally anticipated SWING3 phase.

On this page: Overview · Participating Models · Data Access · Reference — Legacy Archive: SWING1 · SWING2

Phase 1: Present-Day Climate with Reanalysis Nudging

In WisoMIP Phase 1, Working Group 1 focuses on simulations of the present-day climate from 1979 to 2023 using isotope-enabled AGCMs, each nudged toward ERA5 reanalysis data. The eight participating models include all isotope-enabled AGCMs worldwide capable of simulating water isotopes while being nudged globally toward reanalysis circulation fields. The ensemble therefore covers the full set of models capable of conducting the WisoMIP Phase 1 experiment.

Six global maps in two columns. Left: multi-model mean delta-18-O in precipitation,
surface humidity, and total column water vapor. Right: the corresponding inter-model
spread, largest over Antarctica and the high northern continents.

Figure 1. Multi-model mean annual δ18O (left) and inter-model spread (right) for δ18O in precipitation (top), surface humidity (middle), and total column water vapor (bottom), 1979–2023.

Taylor diagram plotting standardized deviation against Pearson correlation for each
model and reanalysis, with four numbered variables and the WisoMIP ensemble mean marked
by black stars.

Figure 2. Taylor diagram comparing the WisoMIP models and reanalyses for surface temperature (1), precipitation (2), δ18O of precipitation (3), and deuterium excess of precipitation (4). Black stars mark the WisoMIP ensemble mean.

By prescribing identical winds, sea surface temperatures, and sea ice conditions, the experiment isolates differences in model physics related to the water cycle and its isotopic behavior, while controlling for variability in atmospheric dynamics.

The ensemble mean generally agrees more closely with observations than most individual models because biases and outlying behavior among models partially cancel. The resulting nudged ensemble can serve as one benchmark for isotope-enabled model development, comparisons with satellite retrievals, and applications as an “isotope reanalysis” product, ranging from the investigation of responses to major modes of climate variability to paleoclimate reconstruction.

Full details are provided in Bong et al. (2026).

Participating Models

Native resolutions are given as longitude × latitude × vertical levels. All models are additionally provided interpolated to a common 128 × 64 grid on 17 pressure levels (1000, 925, 850, 700, 600, 500, 400, 300, 250, 200, 150, 100, 70, 50, 30, 20, 10 hPa).

Model Native resolution Principal points of contact
iCAM5 144 × 96 × 17 Qinghua Ding, Jiang Zhu
iCAM6 288 × 192 × 17 Richard P. Fiorella, Jiang Zhu
ECHAM6-wiso 192 × 96 × 26 Martin Werner, Alexandre Cauquoin
GISS-E2.1 144 × 90 × 17 Allegra N. LeGrande, Hayoung Bong
IsoGSM3 188 × 94 × 17 Kei Yoshimura, Hayoung Bong
LMDZ6 144 × 143 × 79 Camille Risi, Cécile Agosta
MIROC5-iso 128 × 64 × 17 Atsushi Okazaki, Hayoung Bong
NICAM-WISO 180 × 90 × 18 Masahiro Tanoue

Data Access

The WisoMIP Phase 1 archive (v1, 26 June 2025, 180 files, approximately 96 GB) is hosted on the NASA Center for Climate Simulation (NCCS) data portal and is available for direct download:

Files are netCDF, named following the pattern Total.<MODEL>_<PRODUCT>.nc, where <MODEL> is one of CAM5, CAM6, ECHAM, GISS, GSM (for IsoGSM3), LMDZ, MIROC, NICAM, or ENSEMBLE.

Direct Downloads

Each link is a single netCDF file containing all 28 variables. Monthly and Yearly are time series spanning 1979–2023. The remaining products are climatological means over that period. All fields are on the common 128 × 64 grid at 17 pressure levels, except Native grid, which preserves each model's own resolution and so varies in size from model to model.

Model Monthly DJF MAM JJA SON Yearly Annual mean Native grid
Ensemble mean 2.5 GB 5 MB 5 MB 5 MB 5 MB 207 MB 5 MB N/A
iCAM5 2.5 GB 5 MB 5 MB 5 MB 5 MB 207 MB 5 MB 4.2 GB
iCAM6 2.5 GB 5 MB 5 MB 5 MB 5 MB 206 MB 5 MB 16.7 GB
ECHAM6-wiso 2.5 GB 5 MB 5 MB 5 MB 5 MB 206 MB 5 MB 8.1 GB
GISS-E2.1 2.5 GB 5 MB 5 MB 5 MB 5 MB 206 MB 5 MB 3.9 GB
IsoGSM3 2.5 GB 5 MB 5 MB 5 MB 5 MB 206 MB 5 MB 5.5 GB
LMDZ6 2.5 GB 5 MB 5 MB 5 MB 5 MB 207 MB 5 MB 25.6 GB
MIROC5-iso 2.5 GB 5 MB 5 MB 5 MB 5 MB 206 MB 5 MB 2.5 GB
NICAM-WISO 2.5 GB 5 MB 5 MB 5 MB 5 MB 207 MB 5 MB 5.1 GB

Monthly Climatology

Climatological mean for each calendar month over 1979–2023.

Model Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Ensemble mean 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB
iCAM5 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB
iCAM6 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB
ECHAM6-wiso 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB
GISS-E2.1 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB
IsoGSM3 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB
LMDZ6 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB
MIROC5-iso 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB
NICAM-WISO 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB 5 MB

Variables

Every file contains all 28 variables listed below, so you can choose a file by the time period you need rather than by variable. Twenty-one of the variables are single-level maps on the latitude-longitude grid. The remaining seven vary with height (d18O, dD, dexcess, ta, ua, va, and hus) and have one extra vertical dimension, named p, which holds the 17 pressure levels listed above.

Name Units Description
Isotopes: precipitation
d18Opδ18O of precipitation
dDpδD of precipitation
dexcesspDeuterium excess of precipitation
Isotopes: evaporation
d18Oeδ18O of evaporation
dDeδD of evaporation
dexcesseDeuterium excess of evaporation
Isotopes: surface specific humidity
d18Osδ18O of specific humidity (surface)
dDsδD of specific humidity (surface)
dexcesssDeuterium excess of specific humidity (surface)
Isotopes: specific humidity on pressure levels
d18Oδ18O of specific humidity (pressure level)
dDδD of specific humidity (pressure level)
dexcessDeuterium excess of specific humidity (pressure level)
Isotopes: vertically integrated water vapor
d18Owδ18O of vertically integrated water vapor
dDwδD of vertically integrated water vapor
dexcesswDeuterium excess of vertically integrated water vapor
Water cycle and meteorology
prmm/dayPrecipitation (surface)
evmm/dayEvaporation (surface)
husskg/kgSpecific humidity (surface)
huskg/kgSpecific humidity (pressure level)
prwkg/m2Atmospheric column precipitable water
tas°CTemperature (surface)
ta°CTemperature (pressure level)
pshPaPressure (surface)
uam/su-wind (pressure level)
vam/sv-wind (pressure level)
iuqkg/m/sVertically integrated moisture u-flux
ivqkg/m/sVertically integrated moisture v-flux
pref%Precipitation efficiency, (pr/qvsum) × 100

Note: WisoMIP covers the atmosphere. For water isotope data in the ocean, see the Global Seawater Oxygen-18 Database.

Reference

Bong, H., et al., 2026: Water Isotope Model Intercomparison Project (WisoMIP): Present-day climate. J. Geophys. Res. Atmos., 131, e2025JD044985, doi:10.1029/2025JD044985. An open-access version is available from the ESS Open Archive.

Credits

Acknowledgment is given to Hayoung Bong and Allegra N. LeGrande for compiling and processing the WisoMIP Phase 1 data, managing the archive at NASA GISS, and analyzing the model results, to Sylvia Dee for coordinating the project, to all participating modeling groups listed above, and to the providers of the global ground-based and satellite observations used for model evaluation and analysis.

Contact

Please contact Dr. Allegra N. LeGrande or Dr. Gavin Schmidt if you have any questions or comments about these data.