Linjiong Zhou

We present the implementation of an inline three‐dimensional (3D) turbulent kinetic energy (TKE)‐based subgrid diffusion scheme in the Geophysical Fluid Dynamics Laboratory (GFDL) Finite‐Volume Cubed‐Sphere (FV3) dynamical core. This scheme addresses the need for a unified and physically grounded treatment of turbulence as FV3 is increasingly used across the gray zone and large‐eddy simulation scales. Building on the 3D TKE framework introduced in Zhu et al. (2025, https://doi.org/10.1038/s41612‐025‐01117‐6 ), we embed the full TKE budget calculations, including the 3D shear production, and TKE‐informed horizontal and vertical diffusion directly into the FV3 core. Key features include computation of 3D TKE shear production based on FV3 D‐grid winds, updating of TKE and diffusion coefficients within FV3 time‐stepping loop, and extension of TKE‐based horizontal diffusion to all dynamical quantities. This integration avoids reliance on external physics packages and improves consistency between resolved and subgrid processes. The inline 3D TKE scheme offers a robust alternative to numerical damping in FV3 and enables a physically based representation of 3D subgrid‐scale diffusion that is well suited for applications at grid spacings of ∼100 m and finer.

Mesoscale convective systems (MCSs) are critical components of global energy and water cycles and significantly contribute to extreme weather events. However, projecting future MCS behavior remains challenging due to the limitations of regional models and the inadequate representation of MCSs in coarser climate models. In this study, we use GFDL's global storm‐resolving model (GSRM), X‐SHiELD, to explore the response of global MCSs to both increased sea surface temperatures (SST) and elevated CO ₂ levels using three sets of unique two‐year‐long warming simulations. We find that SST warming leads to an increase in MCS occurrence over ocean regions while reducing it over land, whereas elevated CO ₂ results in an overall increase over ocean and land. When SST and CO ₂ increases are combined, their impacts on MCS changes are generally additive. Using stepwise multiple linear regression, we identify the key environmental drivers of these changes across five MCS hotspots, highlighting the regional variability in MCS responses. Furthermore, MCS‐associated precipitation and its contribution to total rainfall are shaped by changes in both MCS frequency and the precipitation intensity within each event. By utilizing the explicit MCS‐resolving capabilities of GSRMs, this study provides critical insights into future changes in MCS characteristics and their implications for global precipitation patterns.

Recent advances have allowed for integration of global storm resolving models (GSRMs) to a timescale of several years. These short simulations are sufficient for studying aggregated statistics of short-timescale and small spatial-scale phenomena; however, it is questionable what we can learn from these integrations about the large-scale climate response to perturbations. To address this question, we use the response of X-SHiELD (a GSRM) to uniform sea surface temperature warming and CO₂ increase in two-year integrations and compare it to similar CMIP experiments. Specifically, we assess the statistical meaning of having two years in one model outside the spread of another model or model ensemble. This is of particular interest because X-SHiELD shows a distinct response of the global-mean precipitation to uniform warming and the northern hemisphere jet shift response to isolated CO₂ increase. To estimate the probability of X-SHiELD’s and the CMIP models having different means, we take the approach of Bayesian inference. We derive a posterior distribution for the differences in the mean between X-SHiELD and the CMIP models taking into account the X-SHiELD values for the global-mean precipitation response to uniform warming and the response of the norther hemisphere jet latitude to isolated CO₂ increase. We find that the most probable value for the difference between X-SHiELD and the CMIP mean is larger than one standard deviation, representing both internal variability and inter-model spread of the CMIP models. We also find that there is an important base-state dependence for some large-scale metrics that, when taken into account, can qualitatively change the interpretation of the results. We note that a year-to-year comparison is meaningful due to the use of prescribed sea-surface-temperature simulations.

Observations reveal a clear difference in the intensity of deep convection over tropical land and ocean. This observed land‐ocean contrast provides a natural benchmark for evaluating the fidelity of global storm‐resolving models (GSRMs; global models with horizontal resolution on the order of kilometers), and GSRMs provide a potentially valuable tool for probing unresolved scientific questions about the origin of the observed land‐ocean contrast. However, land‐ocean differences in convective intensity have received relatively little attention in GSRM research. Here, we show that the strength of the land‐ocean contrast simulated by GSRMs is strongly sensitive to details of GSRM implementations, and not clearly governed by any of several hypothesized drivers of the observed land‐ocean contrast. We first examine DYAMOND Summer GSRM simulations, and show that only a subset produce a clear land‐ocean contrast in the frequency of strong updrafts. We then show that the use of a sub‐grid shallow convection scheme can determine whether or not the GSRM X‐SHiELD produces a clear land‐ocean contrast. Finally, we show that three putative drivers of the observed land‐ocean contrast (convective available potential energy, boundary layer depth, and microphysics) fail to explain why a land‐ocean contrast is present in X‐SHiELD simulations with sub‐grid shallow convection disabled. These results provide encouraging evidence that GSRMs can mimic the observed land‐ocean convective intensity contrast. However, they also show that their ability to do so can be sensitive to uncertain sub‐grid parameterizations, and suggest that existing theory may not fully capture drivers of the land‐ocean contrast simulated by some GSRMs.

Using the novel kilometer‐scale global storm‐resolving model Geophysical Fluid Dynamics Laboratory eXperimental System for High‐resolution prediction on Earth‐to‐Local Domains (X‐SHiELD), we investigate the impact of a 4 K increase in sea surface temperatures on Northern Hemisphere midlatitude cyclones, during the January 2020–January 2022 period. X‐SHiELD simulations reveal a poleward shift in cyclone tracks under warming, consistent with CMIP projections. However, X‐SHiELD's high resolution and explicit deep convection allowed for a detailed analysis of the warm and cold sectors, which are instead typically underrepresented in traditional CMIP models. Instead, compositing the 100 most intense midlatitude cyclones in the North Atlantic, we find that the warm sector exhibits statistically significant increases in wind speed and precipitation of up to 15% locally per degree of warming, while changes in the cold sector are less pronounced. This study demonstrates X‐SHiELD's potential to provide a realistic‐looking perspective into the evolving risks posed by midlatitude cyclones in a warming climate.

The Geophysical Fluid Dynamics Laboratory (GFDL)’s System for High‐resolution prediction on Earth‐to‐Local Domains (SHiELD) model typically uses the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS) analyses to initialize its medium-range global forecasts. Both initial condition (IC) and forecast model have an impact on model prediction skills. The quality of the IC is partially determined by the model short-range forecast used as first guess in data assimilation. A data assimilation (DA) system has been developed for the global SHiELD to demonstrate the prediction skills of the model initialized from its own analysis. The DA system largely leverages the advanced DA techniques used in GFS and assimilates all the observations assimilated in GFS. Compared to the SHiELD forecasts initialized from GFS analysis, SHiELD forecast skill is significantly improved by using its own analysis. Tremendous improvement was found in the Southern Hemisphere with positive impact lasting up to 10 days. The DA system is also useful in identifying and understanding model errors. The most noticeable model error detected by the DA system originates from the TKE-EDMF boundary layer scheme. The model error leads to insufficient ensemble spread, which could not be fully addressed by the multiplicative inflation and stochastic physics schemes used in the system. Including two versions of the TKE EDMF scheme in the ensemble can alleviate the systematic model error, which further improves forecast skills. The use of the interchannel correlated observation errors for Infrared Atmospheric Sounding Interferometer (IASI) and Cross-track Infrared Sounder (CrIS) was also investigated, which improves the forecast skill up to day 5 and further reduces the impact of the model error in the marine stratocumulus region. Further understanding of the model error associated with the TKE-EDMF scheme will be presented.

As the atmosphere gets warmer, it is expected to hold more water vapor, thereby fueling stronger storms. At the same time, the condensation of this vapor increases the combined load of liquid water and ice aloft, forcing convection to do more work to lift water to the level where it precipitates. This takes away from the generation of kinetic energy, thereby creating a “brake” on atmospheric motions. The evolution of this precipitation “brake” with warming determines the magnitude of future storm intensification, with important societal implications. The new generation of kilometer-scale climate models is capable of projecting this evolution. In this presentation, we show how the NOAA/GFDL X-SHiELD experimental global storm-resolving model can be used to estimate the total mechanical work done by convection and the work done to lift water which is then subsequently dissipated by friction during precipitation. The statistics are computed in year-long simulations of the present climate and of a 4K warmer climate.We find that the ratio of kinetic energy generation vs work spent to lift water is respectively 30% vs 70% of the total mechanical work done by convection on global average, with a relative stability across regions and in the present vs future climate.Moving beyond regional averages, when we organize the space by decreasing values of dissipation, we find that the ratio of work spent to lift water to total mechanical work strongly increases in the most convective percentiles, that is, most of the work done by convection is used to lift water in the extremes, showing that water loading strongly opposes kinetic energy generation. We also find that the total work done by convection, the work spent to lift water and the precipitation-induced dissipation all increase similarly with warming in the most convective percentiles. This suggests that, as the Earth warms, the updrafts tend to “kill” themselves in situ from increased water loading instead of generating a response at larger scale.

We introduce a 6.5‐km version of the Geophysical Fluid Dynamics Laboratory (GFDL)'s System for High‐resolution prediction on Earth‐to‐Local Domains (SHiELD). This global model is designed to bridge the gap between global medium‐range weather prediction and global storm‐resolving simulation while remaining practical for real‐time forecast. The 6.5‐km SHiELD represents a significant advancement over GFDL's flagship global forecast system, the 13‐km SHiELD. This global model features a holistically‐developed scale‐aware suite of physical parameterizations, stepping into the formidable convective “gray zone” of resolutions below 10 km. Comparative analyses with the 13‐km SHiELD, conducted over a 3‐year hindcast period, highlight noteworthy improvements across global‐scale, regional‐scale, tropical cyclone (TC), and continental convection predictions. In particular, the 6.5‐km SHiELD excels in predicting considerably finer‐scale convective systems associated with large‐scale frontal systems and extratropical cyclones. The predictions of global temperature, wind, cloud, and precipitation are significantly improved in this global model. Regionally, over the contiguous United States and the Maritime Continent, substantial reductions in prediction biases of precipitation, cloud cover, and wind fields are also found. In the mesoscale realm, the model demonstrates prominent improvements in global TC intensity and continental convective precipitation prediction: biases are relieved, and skill is higher. These findings affirm the superiority of the 6.5‐km SHiELD compared to the current 13‐km SHiELD, which will advance weather prediction by successfully addressing both synoptic weather systems and specific storm‐scale phenomena in the same global model.