Search for: All records

Abstract Machine learning (ML) models are successful with weather forecasting and have shown progress in climate simulations, yet leveraging them for useful climate predictions needs exploration. Here we show this feasibility using Neural General Circulation Model (NeuralGCM), a hybrid ML-physics atmospheric model developed by Google, for seasonal predictions of large-scale atmospheric variability and Northern Hemisphere tropical cyclone (TC) activity. Inspired by physical model studies, we simplify boundary conditions, assuming sea surface temperature (SST) and sea ice follow their climatological cycle but persist anomalies present at the initialization time. With such forcings, NeuralGCM can generate 100 simulation days in ~8 minutes with a single Graphics Processing Unit (GPU), while simulating realistic atmospheric circulation and TC climatology patterns. This configuration yields useful seasonal predictions (July–November) for the tropical atmosphere and various TC activity metrics. Notably, the predicted and observed TC frequency in the North Atlantic and East Pacific basins are significantly correlated during 1990–2023 (r=~0.7), suggesting prediction skill comparable to existing physical GCMs. Despite challenges associated with model resolution and simplified boundary forcings, the model-predicted interannual variations demonstrate significant correlations with the observation, including the sub-basin TC tracks (p<0.1) and basin-wide accumulated cyclone energy (p<0.01) of the North Atlantic and North Pacific basins. These findings highlight the promise of leveraging ML models with physical insights to model TC risks and deliver seamless weather-climate predictions. 

Abstract Wind energy plays a critical role in mitigating climate change and meeting growing energy demands. However, the long-term impacts of anthropogenic warming on wind resources, particularly their seasonal variations and potential compounding risks, remain understudied. Here we analyze large-ensemble climate simulations in high-emission scenarios to assess the projected changes in near-surface wind speed and their broader implications. Our analyses show robust wind changes including a decrease of wind speed (i.e., stilling) up to ~15% during the summer months in Northern Midlatitudes. This stilling is linked to amplified warming of the midlatitude land and the overlying troposphere. Despite regional and model uncertainties, robust signals of warming-induced wind stilling will likely emerge from natural climate variations in the late 21st century under the high-emission scenarios. Importantly, the summertime wind stilling coincides with a projected surge in cooling demand, and their compounding may disrupt the energy supply-demand balance earlier. These findings highlight the importance of considering the seasonal responses of wind resources and the associated climate-energy risks in a warming climate. By integrating these insights into future energy planning decisions, we can better adapt to a changing climate and ensure a reliable and resilient energy future. 

Changing atmospheric circulations shift global weather patterns and their extremes, profoundly affecting human societies and ecosystems. Studies using atmospheric reanalysis and climate model data indicate diverse circulation changes in recent decades but show discrepancies in magnitude and even direction, underscoring the urgent need for validation with independent, climate-quality measurements. Here we show statistically significant changes in tropospheric circulation over the past two decades using satellite-observed, height-resolved cloud motion vectors from the Multi-angle Imaging SpectroRadiometer (MISR). Upper tropospheric cloud motion speeds in the mid-latitudes have increased by up to about 4 m s−1 decade−1. This acceleration is primarily because of the strengthening of meridional flow, potentially indicating more poleward storm tracks or intensified extratropical cyclones. The Northern and Southern Hemisphere tropics shifted poleward at rates of 0.42 ± 0.22 and 0.02 ± 0.14° latitude decade−1 (95% confidence interval), respectively, whereas the corresponding polar fronts shifted at rates of 0.37 ± 0.31 and 0.31 ± 0.21° latitude decade−1. We also show that the widely used ERA5 reanalysis winds subsampled to the MISR are in good agreement with the climatological values and trends of the MISR but indicate probable ERA5 biases in the upper troposphere. These MISR-based observations provide critical benchmarks for refining reanalysis and climate models to advance our understanding of climate change impacts on cloud and atmospheric circulations. 

The future risk of tropical cyclones (TCs) strongly depends on changes in TC frequency, but models have persistently produced contrasting projections. A satisfactory explanation of the projected changes also remains elusive. Here we show a warming-induced contraction of tropical convection delays and reduces TC formation. This contraction manifests as stronger equatorial convection and weaker off-equatorial convection. It has been robustly projected by climate models, particularly in the northern hemisphere. This contraction shortens TC seasons by delaying the poleward migration of the intertropical convergence zone. At seasonal peaks of TC activity, the equatorial and off-equatorial components of this contraction are associated with TC-hindering environmental changes. Finally, the convection contraction and associated warming patterns can partly explain the ensemble spread in projecting future TC frequency. This study highlights the role of convection contraction and provides motivation for coordinated research to solidify our confidence in future TC risk projections.