ICON‐A, the Atmosphere Component of the ICON Earth System Model: I. Model Description

Self Cite

104

We evaluate the new icosahedral nonhydrostatic atmospheric (ICON‐A) general circulation model of the Max Planck Institute for Meteorology that is flexible to be run at grid spacings from a few tens of meters to hundreds of kilometers. A simulation with ICON‐A at a low resolution (160 km) is compared to a not‐tuned fourfold higher‐resolution simulation (40 km). Simulations using the last release of the ECHAM climate model (ECHAM6.3) are also presented at two different resolutions. The ICON‐A simulations provide a compelling representation of the climate and its variability. The climate of the low‐resolution ICON‐A is even slightly better than that of ECHAM6.3. Improvements are obtained in aspects that are sensitive to the representation of orography, including the representation of cloud fields over eastern‐boundary currents, the latitudinal distribution of cloud top heights, and the spatial distribution of convection over the Indian Ocean and the Maritime Continent. Precipitation over land is enhanced, in particular at high‐resolution ICON‐A. The response of precipitation to El Niño sea surface temperature variability is close to observations, particularly over the eastern Indian Ocean. Some parameterization changes lead to improvements, for example, with respect to rain intensities and the representation of equatorial waves, but also imply a warmer troposphere, which we suggest leads to an unrealistic poleward mass shift. Many biases familiar to ECHAM6.3 are also evident in ICON‐A, namely, a too zonal SPCZ, an inadequate representation of north hemispheric blocking, and a relatively poor representation of tropical intraseasonal variability.

Section: Models Experiments and Datamentioning

confidence: 99%

ICON‐A, The Atmosphere Component of the ICON Earth System Model: II. Model Evaluation

Crueger

Giorgetta

Brokopf

et al. 2018

Self Cite

104

“…Three transpose AMIP‐style experiments have been performed with the ICON‐A GCM with prescribed SSTs and sea ice coverage from observations at a resolution of R2B4 (approximately 139 km). ICON‐A also has a terrain following hybrid sigma height grid following Leuenberger et al (), with 47 full vertical levels, with layer thicknesses ranging from 40 m in the lowermost layers to ∼300 m at 1 km, and ∼1 km at 10 km (Giorgetta et al, ).…”

Section: Models and Experimentsmentioning

confidence: 99%

A Prospectus for Constraining Rapid Cloud Adjustments in General Circulation Models

Nam

Kühne

Salzmann

et al. 2018

Self Cite

Rapid cloud adjustments are an important component of the atmosphere's total response to increased CO2 concentrations. Unfortunately, scientific understanding of rapid shortwave cloud adjustments is rather poor. State‐of‐the‐art 5th Coupled Model Intercomparison Project models showed large uncertainty in regard to rapid cloud adjustments. This study determines whether large‐eddy simulations may, in principle, be used as a reference, thanks to their ability to resolve cloud dynamics and thermodynamics, to constrain rapid shortwave cloud adjustments in general circulation models. This is an open question since large‐eddy models can only be run over limited domains, for a short period of time, and are influenced by boundary conditions. Using the Icosahedral Non‐hydrostatic global climate model—Atmospheric component (ICON‐A), we examine shortwave rapid cloud adjustments over central Europe, which is found to be representative of shortwave rapid cloud adjustments over Northern Hemispheric global continents in the 5th Coupled Model Intercomparison Project models. This work finds (i) a couple of days of simulation is sufficient to get a clear signal in the net top‐of‐atmosphere radiative balance to emerge after a 4xCO2 perturbation and (ii) use of present‐day meteorological and CO2 concentrations for boundary conditions in global simulations is not an issue for short lead times, up to ∼36 hr. We also found that atmospheric processes influencing shortwave rapid cloud adjustments over central Europe are largely thermodynamically driven changes in local cloud dynamics and are rather independent of the synoptic‐scale and circulation effects on short timescales (<2 days). These results imply that high‐resolved large‐eddy simulations over a limited area can be instructive for assessing and constraining global rapid cloud adjustments.

“…We use an idealized numerical model of the tropical atmosphere, the ICON‐Spherical Limited Area Model (Müller, ; Müller et al, ). The model is based upon the atmospheric component of the ICOsahedral Non‐hydrostatic atmosphere model, ICON‐A (Giorgetta et al, ).…”

Section: Methodsmentioning

confidence: 99%

Self‐Aggregation of Convection in Spatially Varying Sea Surface Temperatures

Müller

Hohenegger

2020

The phenomenon of self-aggregation of convection was first identified in convection-permitting simulations of radiative convective equilibrium, characterized by homogeneous boundary conditions and in the absence of planetary rotation. In this study, we expose self-aggregation of convection to more complex, nonhomogeneous boundary conditions and investigate its interaction with convective aggregation, as forced by large-scale variations in sea surface temperatures (SSTs). We do this by conducting radiative convective equilibrium simulations on a spherical domain, with SST patterns that are zonally homogeneous but meridionally varying. Due to the meridional contrast in SST, a convergence line first forms, mimicking the Intertropical Convergence Zone. We nevertheless find that the convergence line breaks up and contracts zonally as a result of the self-aggregation of convection. The contraction is significant, being here more than 50% of the original extent. The stability of the convergence line is controlled by the strength of the meridional circulation, which depends upon the imposed SST contrast. However, the process of self-aggregation, once it is initiated, is insensitive to the strength of the SST contrast. The zonal contraction is accompanied by a slight meridional expansion and a moistening of the high latitudes, where SSTs are low. The moistening of the high latitudes can be understood from the fact that the convective cluster intensifies and expands its moist meridional low-level outflow when it self-aggregates zonally. Overall, our results suggest that the Intertropical Convergence Zone may be unstable to the self-aggregation of convection, that self-aggregation may serve as a precursor to the formation of atmospheric rivers, and that longer convergence lines are more likely to exist in regimes with strong SST gradients.Plain Language Summary Storm clouds over the tropical oceans are found to organize in large systems. They tend to form where the temperatures of the sea surface (SSTs) are highest and that is most often along the equator. With idealized computer model experiments we investigate how different sea surface temperature patterns, along with the natural tendency of clouds to aggregate, control the properties of the storm cloud system. We find that a contrast in SSTs immediately acts to organize the storm clouds. In our simulations, this means the formation of a long line of clouds, oriented along the equator. With time however, this line breaks up and then contracts just by itself. Here we find that the spatial difference in SSTs controls the stability of the cloud system against the break up and contraction: The stronger the difference in SST, the more stable is the line of clouds. Also we find that the storm clouds export moisture only in one direction: from the equator to the higher latitudes.