This content will become publicly available on May 1, 2027

Title: Quantifying Very Extreme Precipitation and Temperature Using Huge Ensembles Generated by Machine Learning–Based Climate Model Emulators
Abstract Weather extremes produce major impacts on society and ecosystems and are likely to change in likelihood and magnitude with climate change. However, very low-probability events are hard to characterize statistically using observations or even climate model output because of short records/runs. For precipitation, consideration of such events arises in quantifying probable maximum precipitation (PMP), namely, estimating extreme precipitation magnitudes for designing and assessing critical infrastructure. A recent National Academies report on modernizing PMP estimation proposed using very large climate model-based ensembles to estimate extreme quantiles, possibly through machine learning–based ensemble boosting. Here, we assess statistical aspects of such an approach for the contiguous United States using a huge ensemble (10 560 years) produced by a state-of-the-art emulator (ACE2) trained on ERA5 reanalysis. The results indicate that one can practically estimate very extreme precipitation and temperature quantiles, provided one uses appropriate statistical extreme value techniques. More specifically, the results provide evidence for 1) the use of threshold-exceedance methods with a sufficiently high threshold (necessary for precipitation) for reliable estimation, 2) the robustness of results to variation in extremes by season and storm type, and 3) the sufficiency of the ensemble for well-constrained statistical uncertainty. Our results also show that the emulator produces extremes outside the range of the ERA5 training data. While encouraging for emulators’ potential use for quantifying the climatology of extremes, more investigation is needed to assess whether emulators are fit for this purpose. Our focus is on how to use huge ensembles to estimate very extreme statistics; we expect the results to be relevant for future improved emulators. Significance StatementWeather extremes produce major impacts on society and ecosystems and are likely to change in likelihood and magnitude with climate change. Machine learning now provides methods to simulate weather for many hypothetical years at much lower computational expense than traditional models. This could provide the opportunity to estimate the magnitude of very low-probability events using statistical methods applied to the simulation output, for use in planning, risk assessment, and adaptation. Our work finds that when used appropriately, statistical methods designed to analyze extremes can reliably estimate the magnitude of extremely high precipitation and temperature events in the United States, under the critical assumption that the machine learning method is a good mimic of real-world weather.  more » « less
Award ID(s):
2311164
PAR ID:
10700087
Author(s) / Creator(s):
 ;  
Publisher / Repository:
American Meteorological Association
Date Published:
Journal Name:
Bulletin of the American Meteorological Society
Volume:
107
Issue:
5
ISSN:
0003-0007
Page Range / eLocation ID:
E1053 to E1070
Subject(s) / Keyword(s):
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract We present a lightweight, easy‐to‐train, low‐resolution, fully data‐driven climate emulator, LUCIE, that can be trained on as low as 2 years of 6‐hourly ERA5 data. Unlike most state‐of‐the‐art AI weather models, LUCIE remains stable and physically consistent for 100 years of autoregressive simulation with 100 ensemble members. Long‐term mean climatology from LUCIE's simulation of temperature, wind, precipitation, and humidity matches that of ERA5 data, along with the variability. We further demonstrate how well extreme weather events and their return periods can be estimated from a large ensemble of long‐term simulations. We further discuss an improved training strategy with a hard‐constrained first‐order integrator to suppress autoregressive error growth, a novel spectral regularization strategy to better capture fine‐scale dynamics, and finally an optimization algorithm that enables data‐limited (as low as 2 years of 6‐hourly data) training of the emulator without losing stability and physical consistency. Finally, we provide a scaling experiment to compare the long‐term bias of LUCIE with respect to the number of training samples. Importantly, LUCIE is an easy to use model that can be trained in just 2.4 hr on a single A‐100 GPU, allowing for multiple experiments that can explore important scientific questions that could be answered with large ensembles of long‐term simulations, for example, the impact of different variables on the simulation, dynamic response to external forcing, and estimation of extreme weather events, amongst others. 
    more » « less
  2. Earth system models, or simulators, are foundational for projecting climate change impacts, but their computational expense limits the number and diversity of simulations available. Machine learning-based emulators, statistical surrogates trained on simulator outputs, can replicate components of climate models at orders-of-magnitude lower cost, enabling ensembles and interpolation across scenarios. We argue that the next phase of climate modeling hinges on closer collaboration between simulator and emulator communities. We outline three priorities: (1) co-design of simulators and emulators so that experimental design, diagnostics, and data products support training, evaluation, and targeted simulation; (2) shared, machine learning-ready benchmarks with data partitions and metrics that emphasize physical fidelity; and (3) treating emulators as reliable software components with interfaces, documentation, and deployment pathways for sensitivity analyses, scenario exploration, and uncertainty decomposition. This perspective envisions emulators not as statistical shortcuts, but core tools that accelerate the pace of climate science. 
    more » « less
  3. Abstract An ensemble postprocessing method is developed to improve the probabilistic forecasts of extreme precipitation events across the conterminous United States (CONUS). The method combines a 3D vision transformer (ViT) for bias correction with a latent diffusion model (LDM), a generative artificial intelligence (AI) method, to postprocess 6-hourly precipitation ensemble forecasts and produce an enlarged generative ensemble that contains spatiotemporally consistent precipitation trajectories. These trajectories are expected to improve the characterization of extreme precipitation events and offer skillful multiday accumulated and 6-hourly precipitation guidance. The method is tested using the Global Ensemble Forecast System (GEFS) precipitation forecasts out to day 6 and is verified against the Climatology-Calibrated Precipitation Analysis (CCPA) data. Verification results indicate that the method generated skillful ensemble members with improved continuous ranked probabilistic skill scores (CRPSSs) and Brier skill scores (BSSs) over the raw operational GEFS and a multivariate statistical postprocessing baseline. It showed skillful and reliable probabilities for events at extreme precipitation thresholds. Explainability studies were further conducted, which revealed the decision-making process of the method and confirmed its effectiveness on ensemble member generation. This work introduces a novel, generative AI–based approach to address the limitation of small numerical ensembles and the need for larger ensembles to identify extreme precipitation events. Significance StatementWe use a new artificial intelligence (AI) technique to improve extreme precipitation forecasts from a numerical weather prediction ensemble, generating more scenarios that better characterize extreme precipitation events. This AI-generated ensemble improved the accuracy of precipitation forecasts and probabilistic warnings for extreme precipitation events. The study explores AI methods to generate precipitation forecasts and explains the decision-making mechanisms of such AI techniques to prove their effectiveness. 
    more » « less
  4. Key Points Extreme weather risk is inherently difficult to quantify because of data scarcity Subseasonal weather forecast ensembles are an untapped resource for computing statistics of extremes, done here by statistical weighting We characterize stratospheric extremes up to once in 500‐year events from 46‐day hindcast ensembles across 20 winters 
    more » « less
  5. Abstract Precipitation extremes will increase in a warming climate, but the response of flood magnitudes to heavier precipitation events is less clear. Historically, there is little evidence for systematic increases in flood magnitude despite observed increases in precipitation extremes. Here we investigate how flood magnitudes change in response to warming, using a large initial-condition ensemble of simulations with a single climate model, coupled to a hydrological model. The model chain was applied to historical (1961–2000) and warmer future (2060–2099) climate conditions for 78 watersheds in hydrological Bavaria, a region comprising the headwater catchments of the Inn, Danube and Main River, thus representing an area of expressed hydrological heterogeneity. For the majority of the catchments, we identify a ‘return interval threshold’ in the relationship between precipitation and flood increases: at return intervals above this threshold, further increases in extreme precipitation frequency and magnitude clearly yield increased flood magnitudes; below the threshold, flood magnitude is modulated by land surface processes. We suggest that this threshold behaviour can reconcile climatological and hydrological perspectives on changing flood risk in a warming climate. 
    more » « less