An official website of the United States government
Abstract Marine heatwaves have become more frequent and intense under anthropogenic warming, posing increasing threats to marine ecosystems and coastal societies, necessitating a better understanding of their mechanism and predictability. Here we show how ocean dynamics modulate marine heatwaves globally by comparing dynamic and slab ocean climate model simulations. We discover that ocean dynamics significantly promote marine heatwave intensity and duration in mid-to-high latitude oceans, as well as the eastern tropical Pacific where marine heatwaves are inherently linked to extreme El Niño events. Our mixed-layer heat budget analysis unravels that heat accumulation during marine heatwave episodes is strongly influenced by vertical mixing and horizontal transport processes, so that warm sea surface temperature extremes in dynamic ocean differ in magnitude and evolution rhythm from those in slab ocean. We further find robust multi-year potential predictability of marine heatwave in the North Atlantic with a dynamic ocean, owing primarily to the predictability of the Atlantic Meridional Overturning Circulation. Our findings emphasize the irreplaceable role of oceanic dynamics in marine heatwave evolution and predictability, with important implications for future climate extreme prediction and adaptation strategies.
more »
« less
This content will become publicly available on February 16, 2027
Deep Learning Atmospheric Models Reliably Simulate Out‐of‐Sample Land Heat and Cold Wave Frequencies
Abstract Deep learning (DL)–based general circulation models (GCMs) are emerging as fast simulators, yet their ability to replicate extreme events outside their training range remains unknown. Here, we evaluate two such models—the hybrid Neural General Circulation Model (NGCM) and purely data‐driven Deep Learning Earth System Model (DLESyM)—against a conventional high‐resolution land–atmosphere model (HiRAM) in simulating land heatwaves and coldwaves. All models are forced with observed sea surface temperatures and sea ice over 1900–2020, focusing on the out‐of‐sample period (1900–1960). Both DL models generalize successfully to unseen climate conditions, broadly reproducing the frequency and spatial patterns of heatwave and coldwave events during 1900–1960 with skill comparable to HiRAM. An exception is over portions of North Asia and North America, where all models perform poorly during 1940–1960. Due to excessive temperature autocorrelation, DLESyM tends to overestimate heatwave and coldwave frequencies, whereas the physics–DL hybrid NGCM exhibits persistence more similar to HiRAM.
more »
« less
- PAR ID:
- 10694388
- Publisher / Repository:
- AGU
- Date Published:
- Journal Name:
- Geophysical Research Letters
- Volume:
- 53
- Issue:
- 3
- ISSN:
- 0094-8276
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Most oceans over the globe have experienced surface warming during the past century, but the subpolar Atlantic is quite otherwise. The sea surface temperature cooling trend to the south of Greenland, known as the North Atlantic Warming Hole, has raised debate over whether it is driven by the slowing of the Atlantic Meridional Overturning Circulation. Here we use observations as a benchmark and climate models as a tool to demonstrate that only models simulating a weakened historical Atlantic overturning can broadly reproduce the observed cooling and freshening in the warming hole region. This, in turn, indicates that the realistic Atlantic overturning slowed between 1900 and 2005, at a rate of −1.01 to −2.97 Sv century−1(1 Sv = 106 m3 s−1), according to a sea-surface-temperature-based fingerprint index estimate. Particularly, the Atlantic overturning slowdown causes an oceanic heat transport divergence across the subpolar North Atlantic, which, while partially offset by enhanced ocean heat uptake, results in cooling over the warming hole region.more » « less
-
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 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 min with a single graphics processing unit 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 observed sub-basin TC tracks (p< 0.1) and basin-wide accumulated cyclone energy (ACE) (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.more » « less
-
Extreme weather events present some of the most severe natural threats to the electric grid, and accurate load forecasting during those events is essential for grid management and disaster preparedness. In this study, we evaluate the effectiveness of hybrid deep learning (DL) models for electrical load forecasting in the IEEE 118-bus system. Our analysis focuses on the Connecticut region during extreme weather. In addition, we determine multivariate models capable of multi-input and multi-output forecasting while incorporating weather data to improve forecasting accuracy. This research is divided into two case studies that analyze different combined DL model architectures. Case Study 1 conducts CNN-Recurrent (RNN, LSTM, GRU, BiRNN, BiGRU, and BiLSTM) models with fully connected dense layers, which combine convolution and recurrent neural networks to capture both spatial and temporal dependencies in the data. Case Study 2 evaluates Hybrid CNN-Recurrent models with a fully connected dense layer model that incorporates a flattening step before the recurrent layers to increase the temporal learning process. Based on the results obtained from our simulations, the hybrid CNN-GRU-FC (using BiGRU) model in Case Study 2 obtained the best performance with an RMSE of 9.112 MW and MAPE of 11.68% during the hurricane period. The Hybrid CNN-GRU-FC model presents a better accuracy of bidirectional recurrent models for load forecasting under extreme weather conditions.more » « less
-
Abstract Extreme heat events are a threat to human health, productivity, and food supply, so understanding their drivers is critical to adaptation and resilience. Anticyclonic circulation and certain quasi-stationary Rossby wave patterns are well known to coincide with heatwaves, and soil moisture deficits amplify extreme heat in some regions. However, the relative roles of these two factors in causing heatwaves is still unclear. Here we use constructed circulation analogs to estimate the contribution of atmospheric circulation to heatwaves in the United States in the Community Earth System Model version 1 (CESM1) preindustrial control simulations. After accounting for the component of the heatwaves explained by circulation, we explore the relationship between the residual temperature anomalies and soil moisture. We find that circulation explains over 85% of heatwave temperature anomalies in the eastern and western United States but only 75%–85% in the central United States. In this region, there is a significant negative correlation between soil moisture the week before the heatwave and the strength of the heatwave that explains additional variance. Further, for the hottest central U.S. heatwaves, positive temperature anomalies and negative soil moisture anomalies are evident over a month before heatwave onset. These results provide evidence that positive land–atmosphere feedbacks may be amplifying heatwaves in the central United States and demonstrate the geographic heterogeneity in the relative importance of the land and atmosphere for heatwave development. Analysis of future circulation and soil moisture in the CESM1 Large Ensemble indicates that, over parts of the United States, both may be trending toward greater heatwave likelihood.more » « less
