Nonequilibrium Fractionation During Ice Cloud Formation in iCAM5: Evaluating the Common Parameterization of Supersaturation as a Linear Function of Temperature

Abstract

Supersaturation with respect to ice determines the strength of nonequilibrium fractionation during vapor deposition onto ice or snow and therefore influences the water isotopic composition of vapor and precipitation in cold environments. Historically, most general circulation models formed clouds through saturation adjustment and therefore prevented supersaturation. To match the observed isotopic content, especially the deuterium excess, of snow in polar regions, the saturation ratio with respect to ice (Si) was parameterized, usually by assuming a linear dependence of Si on temperature. The Community Atmosphere Model Version 5 (CAM5) no longer applies saturation adjustment for the ice phase and thus allows ice supersaturation. Here, we adapt the isotope-enabled version of CAM5 to compute nonequilibrium fractionation in ice and mixed-phase clouds based on Si from the CAM5 microphysics and use it to evaluate the common parameterization of Si. Our results show a wide range of Si predicted by the CAM5 microphysics and reflected in the simulated deuterium excess of Antarctic precipitation; this is overly simplified by the linear parameterization. Nevertheless, a linear function, when properly tuned, can reproduce the average observed relationship between D and deuterium excess reasonably well. However, only the model-predicted Si can capture changes in microphysical conditions under different climate states that are not due to changes in temperature. Furthermore, parametric sensitivity tests show that with the model-predicted Si, water isotopes are more closely tied to the model microphysics and can therefore constrain uncertain microphysical parameters. Plain Language SummaryThe concentration of oxygen and hydrogen isotopes of water depends on meteorological processes such as evaporation from the ocean and cloud formation. Water isotope concentrations measured in ice cores and other natural archives are therefore used to reconstruct Earth's past climate. In Antarctica, isotope concentrations strongly depend on the meteorological conditions during ice and mixed-phase cloud formation, especially the supersaturation with respect to ice. Isotope-enabled climate models, which are often used to support interpretations of isotope measurements, commonly prescribe supersaturation with respect to ice as a linear function of temperature. It is unknown how much this simplification affects the representation of water isotope concentrations in the models. Here we use a recently developed isotope-enabled climate model, which uses a physically based calculation of supersaturation, to evaluate the linear parameterization used in other models. We find that, even though it is less physically realistic, the linear parameterization can represent the average water isotope concentrations reasonably well. We also evaluate how model parameters related to supersaturation affect water isotopes. Our results suggest that water isotopes can potentially be used to improve climate models.

Greenland temperature and precipitation over the last 20 000 years using data assimilation

“…Particularly valuable would be transient simulations that include water isotopes as prognostic variables, which allow for direct assimilation of water isotope ratios (Steiger et al, 2017;Okazaki and Yoshimura, 2017), rather than the use of an explicit proxy system model between temperature and δ 18 O. Recent work shows significant improvements to the realism of water-isotope enabled models in the polar regions (Nusbaumer et al, 2017;Dütsch et al, 2019;Cauquoin et al, 2019;Okazaki and Yoshimura, 2019), and longer simulations, once available, should allow us to further improve upon the results we have presented here. In principle, our…”

Enhancing Understanding of the Hydrological Cycle via Pairing of Process‐Oriented and Isotope Ratio Tracers

“…Nonetheless, the mean absolute error in precipitation improves by >0.5‰ from iCAM5 to iCAM6 (Table 4). Moreover, as with precipitation δ 18 O, iCAM6 exhibits an improved representation of the seasonal variations in precipitation d-excess (Figure 5d) relative to iCAM5, though both models strongly underestimate Luz et al, 2009;Quade et al, 2018), supersaturation effects during condensation (e.g., Ciais & Jouzel, 1994;Dütsch et al, 2019;Jouzel & Merlivat, 1984), or sub-cloud hydrometeor evaporation and re-equilibration (e.g., . We focus on the iCAM6 simulation nudged to the ERA5 reanalysis for the remainder of our analysis.…”

“…Together, these changes have allowed for higher supersaturation, reducing a bias that had been present in CAM5 (e.g., Kay et al, 2016). Higher supersaturation amounts and kinetic fractionation during ice nucleation has been linked to an increase in d-excess over Antarctica, and in wintertime isotope ratios in the high northern latitudes (e.g., Dütsch et al, 2019;Jouzel & Merlivat, 1984;Lowenthal et al, 2016).…”

Stable Isotopic Variability of Daily Precipitation from Yungui Plateau of Southwest China: Insights from Moisture Source and Rainout Effect

“…The recycling estimates varies in the space due to the heterogeneity of the land cover type (e.g., vegetation and water body), which largely contribute to the atmospheric moisture via evaporation and transpiration (Jasechko et al 2013). Thus, long-term monitoring of precipitation in the monsoon regions is needed in the future, which may be helpful to quantify the two recycling components in the complex monsoon system and provide validation of predictions from the existing models within reasonable uncertainty (Dütsch et al 2019;Hu et al 2019).…”