We present here the first results, for the preindustrial and mid-Holocene climatological periods, of the newly developed isotope-enhanced version of the fully coupled Earth system model MPI-ESM, called hereafter MPI-ESM-wiso. The water stable isotopes
The hydrogen and oxygen atoms that compose the water molecule have several natural stable isotopes. This results in several forms of the water molecule called water stable isotopologues (hereafter designated by the term “water isotopes”), the most common being
However, the quantitative translation of past isotope signals recorded in natural archives to climate variables and their interpretation remains challenging because of the numerous and complex processes involved: changes in evaporation conditions and moisture sources, in atmospheric transport pathways, or in the seasonality of the precipitation. For example, using the spatial relationship between the
One way to improve our understanding of the mechanisms controlling the water isotope distribution linked to the variations of climate is to use general circulation models (GCMs) with explicit diagnostics of water stable isotopes. These complex models consider the numerous physical processes that influence the isotopic composition of the different water bodies in the Earth's climate system. Since the pioneering work of
When simulating different climates or evolving climate conditions, it is essential to describe in a coherent way the numerous links and feedbacks between the different natural reservoirs (atmosphere, land–vegetation, ocean) and to minimize the prescription of unknown boundary conditions (e.g., sea surface temperatures). For paleoclimate isotope applications, it means that it is necessary to simulate the water isotopes in a full hydrological cycle system, not only in the atmosphere or in the ocean components. With the gain in performance of supercomputers, it is now possible to model the water isotopes in fully coupled atmosphere–ocean GCMs. In the past decade, such models have been used to examine the internal variability and the forced response to orbital and greenhouse gas forcing for modern and mid-Holocene (6 ka) climates (Goddard Institute for Space Studies (GISS) ModelE:
The mid-Holocene (6k) is one of the PMIP4–CMIP6 (Paleoclimate Modeling Intercomparison Project – Coupled Model Intercomparison Project) past climates to evaluate the performance of the coupled GCMs
In this paper, we present the first results of a new isotope-enhanced version of the fully coupled Max Planck Institute for Meteorology Earth System Model (MPI-ESM)
For this study, we have implemented the water stable isotopes in the Earth system model MPI-ESM
To explicitly simulate both
The atmospheric general circulation model ECHAM6 has been developed on the basis of ECHAM5 (
The land surface model JSBACH
As a part of the coupled model MPI-ESM, the Hydrological Discharge (HD) model
The ocean component, MPIOM, has remained unchanged, except for the adaptations to high-resolution grids (
The coupling procedure between the atmosphere and the ocean in MPI-ESM, via the OASIS3 coupler
For this study, we have used the MPI-ESM-LR configuration (LR is for low resolution). The atmospheric component ECHAM6 was run at an approximately 1.875
Two different experiments were performed: one for the preindustrial period (PI) corresponding to the climate conditions at 1850 CE and one for the mid-Holocene 6000 years ago (6k). For the preindustrial climate, MPI-ESM has been continued from a standard PI simulation, i.e., without isotopes included, which has been run over 1000 years (Christian Stepanek, personal communication, 2019) using identical PI boundary conditions. In an analogous way as
At the end of the simulations, the global mean 2000 m deep salinity changes by less than 0.002 practical salinity units (psu) over 100 years, and the globally averaged
To evaluate our model, we used different datasets including isotopic measurements in precipitation, ocean water, ice cores and continental speleothems. We give a brief summary below.
For the modern isotopic content of precipitation, we use the Global Network of Isotopes in Precipitation (GNIP) database, available through the International Atomic Energy Agency (
To evaluate the modeled PI isotope values in the ocean, we use the Goddard Institute for Space Studies (GISS) global seawater oxygen-18 database
Since the pioneering work of
Selected ice core records and their geographical coordinates, reported PI values of
References:
Furthermore, we also use the SISAL (Speleothem Isotope Synthesis and Analysis) dataset (version 1b:
Figure
In Fig.
Figure
In a similar way as for the atmospheric part, we compare our simulated
In Fig.
Scatter plots of
The modeling of the deuterium-excess signal is challenging for GCMs. For the North Atlantic and Arctic Ocean region, the spatial structure of the marine boundary layer water vapor isotopic composition, which greatly influences the d-excess signal in precipitation, seems to be poorly simulated by the models
The modeled d-excess
MPI-ESM-wiso overestimates the deuterium excess in the ocean surface and underestimates the deuterium excess in precipitation, especially for the highest observed values. However, the modeled linear relationship between the deuterium excess in water vapor above the ocean surface (d-excess
Global distribution of simulated and observed annual mean d-excess values in precipitation
Before analyzing the 6k isotopic anomalies, we check that our modeled 6k
Simulated annual, boreal winter (DJF) and boreal summer (JJA) changes in 2 m temperature
One of the characteristics of the 6k climate is the enhanced seasonal contrast in the Northern Hemisphere due to changes in the insolation, giving rise to warmer Northern Hemisphere summers (Fig.
Even if the changes in temperature and precipitation amount are modest compared to periods like the LGM, they leave imprints on
Next, we compare our simulated 6k
Concerning the changes in d-excess
Figure
Modeled annual mean
The classical use of water isotopes to reconstruct past variations of climate implies that the modern spatial relationship between isotopes and climate variables, such as surface temperature, precipitation rate and salinity, can be taken as a surrogate for the temporal isotope–climate gradient at a given site. Such temporal relationships can be calculated from our model results. In Sect.
In the same way as previous studies
Correlation coefficients
The correlation values between modeled monthly anomalies of
In a similar way to the atmospheric relationships, we assess the temporal gradients between monthly anomalies of
Correlation coefficients
The global spatial relationship between climate variables and water isotopes does not change significantly between our simulated mean climate states of 6k and PI. For example, we find a mean spatial
For numerical reasons, robust
The
The
Spatial distribution of the calculated temporal 6k–PI
Vertically integrated water vapor transport over Antarctica during austral summer for PI
According to our model results, the Indian and African monsoons are enhanced during the mid-Holocene compared to the preindustrial period (Sect.
Figure
Spatial distribution of calculated temporal 6k–PI
We conclude that the reconstruction of past salinity through the isotopic content in sea surface waters can be complicated for regions with strong ocean dynamics (North Atlantic Ocean), variations in sea ice regimes (Arctic and Southern Ocean) or significant changes in the freshwater budget (Bay of Bengal), giving an extremely variable relationship between
In this study, we present the first simulations of the fully coupled model MPI-ESM, enhanced with water isotope diagnostics. The water isotopes have been implemented in all the components of the model (atmosphere, dynamic vegetation, hydrological discharge, ocean–sea ice), and the related isotope masses of
The modeled changes in temperature and precipitation rate during the mid-Holocene compared to the preindustrial period are consistent with previous PMIP results, with a warmer Northern Hemisphere summer and enhanced African and Indian monsoons. One great advantage of enabling MPI-ESM to model water stable isotopes is the possibility to directly compare available isotopic measurements with our climate simulations. We find a fair agreement between our modeled 6k isotopic anomalies and the observations from ice cores and speleothems. MPI-ESM-wiso simulates higher
In numerous previous paleoclimate studies, one of the main assumptions for using water isotopes to study past climate variations is that the modern spatial isotope–climate variable relationships can be used as a surrogate for the temporal gradients at any specific site. In this study, we focused especially on the variability of these relationships during and between two distinct periods of the Holocene. For that, we have analyzed the modeled temporal isotope–climate gradients (i) at an interannual timescale in our PI simulation and (ii) between the mean 6k and PI climates. For the
The focus of this study on the mid-Holocene and preindustrial climates was a first step toward studying the isotope–climate relationship under different warm climate conditions by MPI-ESM-wiso model simulations. Future studies will investigate the hydrological cycle variability for other interglacial periods, including the LIG, and for a transient Holocene experiment.
The MPI-ESM model code is available under a version of the MPI-M software license agreement (
The supplement related to this article is available online at:
AC developed the model code, designed the experiments and performed the simulations with the help of MW. AC and MW analyzed the model outputs. AC wrote the paper with contributions from all coauthors.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Paleoclimate data synthesis and analysis of associated uncertainty (BG/CP/ESSD inter-journal SI)”. It is not associated with a conference.
This work was supported by the German Federal Ministry of Education and Research (BMBF) as a Research for Sustainability initiative (FONA) through the PalMod project (FKZ: 01LP1511B). All simulations were performed at the German Climate Computing Center (DKRZ). We thank Jonathan Holmes and one anonymous referee for their useful suggestions that helped to improve this paper. We thank Christian Stepanek for his help on experiment design and the fruitful discussions. We acknowledge Dirk Barbi for his help for debugging, installing and running the model.
The article processing charges for this open-access publication were covered by a Research Centre of the Helmholtz Association.
This paper was edited by David Thornalley and reviewed by Jonathan Holmes and one anonymous referee.