Marco A. Giorgetta

Western boundary currents (WBCs) are hotspots of air–sea interaction where observations show strong local covariability between sea surface temperature (SST) and precipitation anomalies. We investigate how oceanic and atmospheric horizontal resolution, mesoscale SST variability, and the treatment of atmospheric convection control this coupling across the Gulf Stream, Kuroshio, Agulhas, and Brazil currents. Using a coordinated hierarchy of ICON and EC‐Earth3P model simulations spanning horizontal resolutions of about 5–100 km, either parameterizing or explicitly simulating atmospheric convection, complementary atmosphere‐only sensitivity experiments forced with mesoscale‐resolving or smoothed SSTs, and the CMIP6 ensemble, we assess how model design impacts the strength and spatial structure of SST–precipitation covariability. Resolving mesoscale SST variability is necessary to reproduce the observed localized precipitation responses, but a high horizontal resolution alone does not guarantee full realism. Overly broad coupling emerges in simulations with 5–25 km oceanic resolutions, while explicitly simulating atmospheric convection at 10 km atmospheric resolution produces weaker, more spatially confined precipitation responses than parameterized convection. Whereas mesoscale SST variability sets the potential for strong SST–precipitation coupling, atmospheric physics controls the efficiency of the precipitation response. Our results emphasize the importance of both increasing model resolution and improving its physics for representing realistic air–sea coupling in next‐generation climate simulations.

We show how reducing the fall speeds of rain and ice leads to reduced biases of Convectively Coupled Equatorial Waves (CCEWs) in convection‐resolving configurations of the Icosahedral Non‐hydrostatic Model (ICON). In particular, reducing the fall speeds of rain and ice leads to stronger activity over the wavenumbers and frequencies associated with convectively coupled Kelvin waves, inertio‐gravity waves, mixed Rossby‐gravity waves, equatorial Rossby waves, and the Madden‐Julian oscillation, and to weaker activity of tropical depressions. The changes in CCEW activity are mediated by changes in the frequency of mesoscale convection profiles. Reducing the fall speeds increases the frequency of bottom‐heavy and top‐heavy mesoscale convection profiles and decreases the frequency of mid‐heavy mesoscale convection profiles. In turn, changes to the mesoscale convective profiles are reflected in the ability of the model to develop the tilt and top‐heaviness of CCEWs. Reductions in the fall speeds also increase precipitation over the warm pool region, possibly allowing for a stronger equatorial forcing of convectively coupled Kelvin waves. Our findings suggest that the representation of CCEWs within ICON agrees with their observed multiscale structure.

Elevation-dependent warming (EDW) has been reported in observations and climate models, yet its magnitude and controlling mechanisms remain uncertain, particularly due to the complexity of mountain regions. However both theoretical studies and climate simulations indicate a reduction in lapse rates and enhanced tropospheric warming under climate change. In this study, we examine EDW, and its relationship to tropospheric warming, in atmosphere-only experiments. This is done through the comparison between a historical control simulation and a perturbed climate state driven by uniform 4K warming in the prescribed sea surface temperatures. These simulations were performed with the ICON model in its Sapphire configuration at ∼10 km horizontal grid spacing. This setup offers an improved representation of high-elevation terrain compared to common climate change simulations, key to adequate analysis of EDW, together with a strong free-tropospheric warming, important for understanding its role in shaping EDW. The simulation exhibits a robust, statistically significant increase in surface warming with elevation, ranging from ∼4.9 K below 500 m to almost 7 K above 5500 m, corresponding to a global EDW slope of 0.317 K km-1. Regional contrasts are most pronounced at low elevations, while at intermediate and high elevations the surface warming profiles converge toward the tropospheric warming profile. Seasonal variations suggest an influence from snow-related processes, yet the majority of the seasonal variability in surface warming can be explained by seasonal variations in tropospheric warming.A direct comparison of binned surface and tropospheric temperature changes at corresponding heights reveals a tight coupling, with small deviations possibly resulting from radiative processes near the surface. These results indicate that, under strong free-tropospheric warming, EDW can be approximated to first order by the vertical structure of tropospheric warming, with surface energy-balance processes largely providing secondary modulation. The sensitivity of this coupling to different forcing magnitudes and climate states warrants further investigation.

We show how in the Sapphire configuration of the Icosahedral Non-hydrostatic Model (ICON) the representation of Convectively Coupled Equatorial Waves (CCEWs) is sensitive to the fall speeds of rain and ice, and how reducing these fall speeds can lead to a better representation of CCEWs in ICON. In particular, reductions in the fall speeds of rain and ice lead to more active convectively coupled Kelvin, Inertio-Gravity, and Mixed Rossby-gravity waves and, at the same time, less active convectively coupled Equatorial Rossby waves and Tropical Depressions.We then explore how changes in these fall speeds upscale from the kilometer scales to the synoptic and planetary scales of CCEW, finding that this up-scaling is mediated by the frequency of occurrence of shallow, stratiform, and deep convection. Reducing the fall speed of rain and ice leads to increases in the frequency of occurrence of, respectively, shallow and stratiform convection profiles and, at the same time, leads to decreases in the frequency of occurrence of deep convection profiles. We argue that changes in these profiles are reflected in the ability of the model to develop the tilt and top-heaviness of CCEWs, which ultimately leads to their better representation.Our findings suggest a physical representation of CCEWs within ICON and provide further support for the classification of CCEWs into two distinct categories, a gravity wave group and a moisture mode group, each associated with distinct convective profiles and with distinct propagation mechanisms.

A defining attribute of global-storm resolving models is that modelling is replaced by simulation. In addition to overloading the word “model” this avails the developer of a much larger variety of tests, and brings about a richer interplay with their intuition. This has proven helpful in identifying and correcting many mistakes in global-storm resolving models that traditional climate models find difficult to identify, and usually compensate by “tuning.” It also means that storm-resolving models are built and tested in a fundamentally different way than are traditional climate models. In this talk I will review the development of ICON as a global storm resolving model to illustrate how this feature, of trying to simulate rather than model the climate system, has helped identify a large number of long-standing bugs in code bases inherited from traditional models; how this can support open development; and how sometimes these advantages also prove to be buggy.

Western Boundary Currents (WBCs) are key regions of air–sea interactions, where oceanic variability can strongly influence the atmospheric circulation and precipitation. Despite growing observational evidence of local covariability between SST and precipitation anomalies along these currents, climate models still differ markedly in their ability to represent this coupling. In particular, it remains unclear which elements of model resolution and physical parameterizations control the emergence, strength, and spatial organization of the SST–precipitation relationship.Here, we examine the sensitivity of local SST–precipitation covariability to oceanic and atmospheric resolution and to the representation of moist convection. We analyze a coordinated hierarchy of global simulations, including coarse-resolution CMIP6 models, eddy-permitting and eddy-resolving configurations of ICON and EC-Earth3P, a convection-permitting ICON experiment, and atmosphere-only simulations forced with mesoscale-resolving and smoothed SSTs. Using a consistent diagnostic framework across four major WBC systems, we assess how model design shapes both the amplitude and structure of the atmospheric response.Our results show that resolving mesoscale ocean variability is essential for reproducing a localized precipitation response to SST anomalies. However, increasing resolution alone does not guarantee realism: high-resolution configurations often produce overly broad coupling, while disabling the convective parameterization weakens the response despite fine grid spacing. These findings highlight the need for a physically consistent treatment of ocean mesoscale dynamics and atmospheric convection to capture realistic air–sea coupling along WBCs, with implications for simulating extratropical precipitation and storm-track variability.

We show how, in convection resolving configurations of the Icosahedral Non-hydrostatic Model (ICON), reducing the fall speeds of rain and ice leads to a stronger gravity wave group and a weaker moisture mode group of of Convectively Coupled Equatorial Waves (CCEWs). Reductions in these fall speeds upscale from the grid scale to the synoptic and planetary scales of CCEWs mediated by the frequency of different types of convection. Reducing the fall speeds leads to increases in the frequency of shallow and stratiform convection and to decreases in the frequency of deep convection. In turn, these changes in convection are reflected in the ability of the model to develop the tilt and top-heaviness of CCEWs, which ultimately leads to their better representation in ICON. Our findings suggest a representation of CCEWs within ICON that agrees with their observed multiscale structure and provides further support for the classification of CCEWs.

The radiation parameterization is one of the computationally most expensive components of Earth system models (ESMs). To reduce computational cost, radiation is often calculated on coarser spatial or temporal scales, or both, than other physical processes in ESMs, leading to uncertainties in cloud‐radiation interactions and thereby in radiative temperature tendencies. One way to address this issue is to emulate the radiation parameterization using machine learning (ML), which is typically faster and has good accuracy in high‐dimensional parameter spaces. This study investigates the development and interpretation of an ML‐based radiation emulator using the ICOsahedral Non‐hydrostatic model with the RTE+RRTMGP radiation code, which calculates radiative fluxes based on the atmospheric state and its optical properties. With a Bidirectional Long Short‐Term Memory architecture, which can account for vertical bidirectional auto‐correlation, we can accurately emulate shortwave and longwave heating rates with a mean absolute error of and , respectively. Further, we analyze the trained neural networks using Shapley Additive exPlanations and confirm that the networks have learned physically meaningful relationships among the inputs and outputs. It is worth noting that we observe that the local temperature is used as a predictive source for the longwave heating, consistent with physical models of radiation. For shortwave heating, we find that clouds reflect radiation, leading to reduced heating below the cloud. In contrast, an architecture that is not inspired by the underlying physics, such as a multilayer perceptron, tends to rely on spurious or less physically meaningful correlations to make its predictions.