A5023857630- Meteorological Phenomena and Simulations
- Climate variability and models
- Atmospheric and Environmental Gas Dynamics
- Atmospheric Ozone and Climate
- Geophysics and Gravity Measurements
- Oceanographic and Atmospheric Processes
- Cryospheric studies and observations
- Computational Physics and Python Applications
- Tropical and Extratropical Cyclones Research
- Atmospheric aerosols and clouds
- Radiomics and Machine Learning in Medical Imaging
- Marine and coastal ecosystems
- Precipitation Measurement and Analysis
- Hydrological Forecasting Using AI
- Medical Imaging Techniques and Applications
- Atmospheric chemistry and aerosols
- Solar Radiation and Photovoltaics
- Energy Load and Power Forecasting
- Ionosphere and magnetosphere dynamics
- Climate Change Policy and Economics
- Species Distribution and Climate Change
- Coastal and Marine Dynamics
- Ocean Acidification Effects and Responses
- Scientific Computing and Data Management
- Advanced Measurement and Metrology Techniques
Allen Institute for Artificial Intelligence
2021-2024
University of Washington
2017-2024
Columbia University
2023
Vulcan (United States)
2020-2022
NOAA Geophysical Fluid Dynamics Laboratory
2021
University of Toronto
2013-2017
Abstract Due to limited resolution and inaccurate physical parameterizations, weather climate models consistently develop biases compared the observed atmosphere. Using FV3GFS model at coarse resolution, we propose a method of machine learning corrective tendencies from hindcast simulation nudged toward observational analysis. We show that random forest can predict nudging this with moderate skill using only state as input. This is then coupled FV3GFS, adding temperature, specific humidity...
Abstract Global atmospheric “ storm‐resolving ” models with horizontal grid spacing of less than 5 km resolve deep cumulus convection and flow in complex terrain. They promise to be reference that could used improve computationally affordable coarse‐grid global climate across a range climates, reducing uncertainties regional precipitation temperature trends. Here, machine learning nudging tendencies as functions column state is correct the physical parameterization temperature, humidity,...
Abstract Parameterization of subgrid‐scale processes is a major source uncertainty in global atmospheric model simulations. Global storm‐resolving simulations use finer grid (less than 5 km) to reduce this by explicitly resolving deep convection and details orography. This study uses machine learning replace the physical parameterizations heating moistening rates, but not wind tendencies, coarse‐grid (200 atmosphere model, using training data obtained spatially coarse‐graining 40‐day...
Abstract The degree of Hadley cell expansion under global warming will have a substantial impact on changing rainfall patterns. Most previous studies quantified changes in total tropical width, focused the Southern Hemisphere or considered each hemisphere's response to multitude anthropogenic forcings. It is shown here that exclusive CO 2 forcing, climate models predict twice as much relative Northern Hemisphere. This asymmetry present annual mean and all seasons except boreal autumn. robust...
Earth and Space Science Open Archive This preprint has been submitted to is under consideration at Geophysical Research Letters. ESSOAr a venue for early communication or feedback before peer review. Data may be preliminary.Learn more about preprints preprintOpen AccessYou are viewing the latest version by default [v1]Correcting weather climate models machine learning nudged historical simulationsAuthorsOliverWatt-MeyeriDNoah DominoBrenowitzSpencer KonciusClarkBrianHenniDAnnaKwaJeremy...
The impact of global warming–induced intertropical convergence zone (ITCZ) narrowing onto the higher-latitude circulation is examined in GFDL Atmospheric Model, version 2.1 (AM2.1), run over zonally symmetric aquaplanet boundary conditions. A striking reconfiguration deep tropical precipitation from double-peaked, off-equatorial ascent to a single peak at equator occurs under globally uniform +4 K sea surface temperature (SST) perturbation. This response found be highly sensitive SST profile...
Climate models are complicated software systems that approximate atmospheric and oceanic fluid mechanics at a coarse spatial resolution. Typical climate forecasts only explicitly resolve processes larger than 100 km any process occurring below this scale (e.g. thunderstorms) using so-called parametrizations. Machine learning could improve upon the accuracy of some traditional physical parametrizations by from global cloud-resolving models. We compare performance two machine models, random...
Abstract Bretherton et al. (2022, https://doi.org/10.1029/2021MS002794 ) demonstrated a successful approach for using machine learning (ML) to help coarse‐resolution global atmosphere model with real geography (a ∼200 km version of NOAA's FV3GFS) evolve more like fine‐resolution model, at the scales resolved by both. This study extends that work application in multiple climates and multi‐year ML‐corrected simulations. Here four (∼25 km) 2 year reference simulations are run FV3GFS...
Abstract One approach to improving the accuracy of a coarse‐grid global climate model is add machine‐learned (ML) state‐dependent corrections prognosed tendencies, such that evolves more like reference fine‐grid storm‐resolving (GSRM). Our past work demonstrating this was trained with short (40‐day) simulations GFDL's X‐SHiELD GSRM 3 km horizontal grid spacing. Here, we extend span full annual cycle by training and testing our ML using new year‐long simulation. corrective models are learning...
Existing ML-based atmospheric models are not suitable for climate prediction, which requires long-term stability and physical consistency. We present ACE (AI2 Climate Emulator), a 200M-parameter, autoregressive machine learning emulator of an existing comprehensive 100-km resolution global model. The formulation allows evaluation laws such as the conservation mass moisture. is stable 100 years, nearly conserves column moisture without explicit constraints faithfully reproduces reference...
Abstract This study examines the cause of spread extratropical circulation responses to inclusion atmospheric cloud radiative effects (ACRE) across general models. The ensemble Clouds On‐Off Klimate Intercomparison Experiment aquaplanet simulations shows that these include both equatorward and poleward shifts eddy‐driven jet varying magnitudes. These disparate occur despite relatively consistent response in tropics: a heating upper troposphere, which leads strengthening Hadley cell. It is...
Abstract Accurate precipitation simulations for various climate scenarios are critical understanding and predicting the impacts of change. This study employs a Cycle‐generative adversarial network (CycleGAN) to improve global 3‐hr‐average fields predicted by coarse grid (200 km) atmospheric model across range climates, morphing them match their statistical properties with those reference fine‐grid (25 simulations. We evaluate its performance on both target climates an independent ramped‐SST...
Abstract. Simulation software in geophysics is traditionally written Fortran or C++ due to the stringent performance requirements these codes have satisfy. As a result, researchers who use high-productivity languages for exploratory work often find hard understand, modify, and integrate with their analysis tools. fv3gfs-wrapper an open-source Python-wrapped version of NOAA (National Oceanic Atmospheric Administration) FV3GFS (Finite-Volume Cubed-Sphere Global Forecast System) global...
Cloud microphysical parameterizations in atmospheric models describe the formation and evolution of clouds precipitation, a central weather climate process. Cloud-associated latent heating is primary driver large small-scale circulations throughout global atmosphere, have important interactions with radiation. Clouds are ubiquitous, diverse, can change rapidly. In this work, we build first emulator an entire cloud parameterization, including fast phase changes. The performs well offline...
Abstract This study updates a body of literature that aims to separate atmospheric disturbances into standing and traveling zonal wave components. Classical wavenumber–frequency analysis decomposes longitude- time-dependent signals contributions from distinct spatial temporal scales. Here, an additional decomposition the spectrum components is described. Previous methods decompose power parts with no explicit allowance for covariance between two. provides simple method calculate variance...
Abstract Northern Hemisphere stratospheric polar vortex strength variability is known to be largely driven by persistent anomalies in upward wave activity flux. It has also been shown that attenuation and amplification of the stationary primary way which flux varies. This study determines structure interfere with climatological drive this variability. Using a recently developed spectral decomposition it fixed-node standing waves are drivers “linear interference” phenomenon. particularly true...
Abstract Coarse‐grid weather and climate models rely particularly on parameterizations of cloud fields, coarse‐grained fields from a fine‐grid reference model are natural target for machine‐learned parameterization. We machine‐learn the coarsened‐fine properties as function coarse‐grid state in each grid cell NOAA's FV3GFS global atmosphere with 200 km spacing, trained using 3 simulation modified version FV3GFS. The ML outputs fractional cover liquid ice condensate mixing ratios, inputs...
Abstract Can the current successes of global machine learning‐based weather simulators be generalized beyond 2‐week forecasts to stable and accurate multiyear runs? The recently developed AI2 Climate Emulator (ACE) suggests this is feasible, based upon 10‐year simulations with a network trained on output from physics‐based atmosphere model using grid spacing approximately 110 km forced by repeating annual cycle sea‐surface temperature. Here we show that ACE, without modification, can emulate...
Accurate precipitation simulations for various climate scenarios are critical understanding and predicting the impacts of change. This study employs a Cycle-generative adversarial network (CycleGAN) to improve global 3-hour-average fields predicted by coarse grid (200~km) atmospheric model across range climates, morphing them match their statistical properties with reference fine-grid (25~km) simulations. We evaluate its performance on both target climates an independent ramped-SST...
Abstract The distribution of temperatures in the wintertime polar stratosphere is significantly positively skewed, which has important implications for characteristics ozone chemistry and stratosphere–troposphere coupling. typical argument why temperature skewed that radiative balance sets a firm lower limit, while planetary wave driving can force much larger positive anomalies temperature. However, upward Eliassen–Palm (EP) flux also this suggests dynamics may play an role setting skewness...