skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Using deep learning to emulate and accelerate a radiative-transfer model
Abstract This paper describes the development of U-net++ models, a type of neural network that performs deep learning, to emulate the shortwave Rapid Radiative-transfer Model (RRTM). The goal is to emulate the RRTM accurately in a small fraction of the computing time, creating a U-net++ that could be used as a parameterization in numerical weather prediction (NWP). Target variables are surface downwelling flux, top-of-atmosphere upwelling flux ( ), net flux, and a profile of radiative-heating rates. We have devised several ways to make the U-net++ models knowledge-guided, recently identified as a key priority in machine learning (ML) applications to the geosciences. We conduct two experiments to find the best U-net++ configurations. In Experiment 1, we train on non-tropical sites and test on tropical sites, to assess extreme spatial generalization. In Experiment 2, we train on sites from all regions and test on different sites from all regions, with the goal of creating the best possible model for use in NWP. The selected model from Experiment 1 shows impressive skill on the tropical testing sites, except four notable deficiencies: large bias and error for heating rate in the upper stratosphere, unreliable for profiles with single-layer liquid cloud, large heating-rate bias in the mid-troposphere for profiles with multi-layer liquid cloud, and negative bias at lowzenith angles for all flux components and tropospheric heating rates. The selected model from Experiment 2 corrects all but the first deficiency, and both models run ~10 4 times faster than the RRTM. Our code is available publicly.  more » « less
Award ID(s):
2019758
PAR ID:
10422694
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
Journal of Atmospheric and Oceanic Technology
ISSN:
0739-0572
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; (, Journal of Atmospheric and Oceanic Technology)
    Abstract Radiative transfer (RT) is a crucial but computationally expensive process in numerical weather/climate prediction. We develop neural networks (NN) to emulate a common RT parameterization called the Rapid Radiative Transfer Model (RRTM), with the goal of creating a faster parameterization for the Global Forecast System (GFS) v16. In previous work we emulated a highly simplified version of the shortwave RRTM only—excluding many predictor variables, driven by Rapid Refresh forecasts interpolated to a consistent height grid, using only 30 sites in the Northern Hemisphere. In this work we emulate the full shortwave and longwave RRTM—with all predictor variables, driven by GFSv16 forecasts on the native pressure–sigma grid, using data from around the globe. We experiment with NNs of widely varying complexity, including the U-net++ and U-net3+ architectures and deeply supervised training, designed to ensure realistic and accurate structure in gridded predictions. We evaluate the optimal shortwave NN and optimal longwave NN in great detail—as a function of geographic location, cloud regime, and other weather types. Both NNs produce extremely reliable heating rates and fluxes. The shortwave NN has an overall RMSE/MAE/bias of 0.14/0.08/−0.002 K day−1for heating rate and 6.3/4.3/−0.1 W m−2for net flux. Analogous numbers for the longwave NN are 0.22/0.12/−0.0006 K day−1and 1.07/0.76/+0.01 W m−2. Both NNs perform well in nearly all situations, and the shortwave (longwave) NN is 7510 (90) times faster than the RRTM. Both will soon be tested online in the GFSv16. Significance StatementRadiative transfer is an important process for weather and climate. Accurate radiative transfer models exist, such as the RRTM, but these models are computationally slow. We develop neural networks (NNs), a type of machine learning model that is often computationally fast after training, to mimic the RRTM. We wish to accelerate the RRTM by orders of magnitude without sacrificing much accuracy. We drive both the NNs and RRTM with data from the GFSv16, an operational weather model, using locations around the globe during all seasons. We show that the NNs are highly accurate and much faster than the RRTM, which suggests that the NNs could be used to solve radiative transfer inside the GFSv16. 
    more » « less
  2. ; ; ; ; ; ; ; (, Atmospheric Measurement Techniques)
    Abstract. One-dimensional variational retrievals of temperature and moisture fields from hyperspectral infrared (IR) satellite sounders use cloud-cleared radiances (CCRs) as their observation. These derived observations allow the use of clear-sky-only radiative transfer in the inversion for geophysical variables but at reduced spatial resolution compared to the native sounder observations. Cloud clearing can introduce various errors, although scenes with large errors can be identified and ignored. Information content studies show that, when using multilayer cloud liquid and ice profiles in infrared hyperspectral radiative transfer codes, there are typically only 2–4 degrees of freedom (DOFs) of cloud signal. This implies a simplified cloud representation is sufficient for some applications which need accurate radiative transfer. Here we describe a single-footprint retrieval approach for clear and cloudy conditions, which uses the thermodynamic and cloud fields from numerical weather prediction (NWP) models as a first guess, together with a simple cloud-representation model coupled to a fast scattering radiative transfer algorithm (RTA). The NWP model thermodynamic and cloud profiles are first co-located to the observations, after which the N-level cloud profiles are converted to two slab clouds (TwoSlab; typically one for ice and one for water clouds). From these, one run of our fast cloud-representation model allows an improvement of the a priori cloud state by comparing the observed and model-simulated radiances in the thermal window channels. The retrieval yield is over 90%, while the degrees of freedom correlate with the observed window channel brightness temperature (BT) which itself depends on the cloud optical depth. The cloud-representation and scattering package is benchmarked against radiances computed using a maximum random overlap (RMO) cloud scheme. All-sky infrared radiances measured by NASA's Atmospheric Infrared Sounder (AIRS) and NWP thermodynamic and cloud profiles from the European Centre for Medium-Range Weather Forecasts (ECMWF) forecast model are used in this paper. 
    more » « less
  3. ; ; ; (, Journal of Advances in Modeling Earth Systems)
    Abstract We explore the importance of the life cycle of detrained tropical anvil clouds in producing a weak net cloud radiative effect (NCRE) by tropical convective systems. We simulate a horizontally homogeneous elevated ice cloud in a 2‐D framework using the System for Atmospheric Modeling cloud‐resolving model. The initially thick cloud produces a negative NCRE, which is later canceled by a positive NCRE as the cloud thins and rises. Turning off interactive cloud radiation reveals that cloud radiative heating and in‐cloud convection are fundamental in driving net radiative neutrality. In‐cloud convection acts to thin initially thick anvil clouds and loft and maintain thin cirrus. The maintenance of anvil clouds is tied to the recycling of water vapor and cloud ice through sublimation, nucleation, and deposition as air parcels circulate vertically within the cloud layer. Without interactive radiation, the cloud sediments and sublimates away, producing a large negative NCRE. The specification of cloud microphysics substantially influences the cloud's behavior and life cycle , but the tendency of the life cycle to produce compensating cloud radiative effects is robust to substantial changes in the microphysics. Our study shows that small‐scale processes within upper level ice clouds likely have a strong influence on the NCRE associated with tropical convective cloud systems. 
    more » « less
  4. ; ; ; ; (, Artificial Intelligence for the Earth Systems)
    Abstract Vertical profiles of temperature and dewpoint are useful in predicting deep convection that leads to severe weather, which threatens property and lives. Currently, forecasters rely on observations from radiosonde launches and numerical weather prediction (NWP) models. Radiosonde observations are, however, temporally and spatially sparse, and NWP models contain inherent errors that influence short-term predictions of high impact events. This work explores using machine learning (ML) to postprocess NWP model forecasts, combining them with satellite data to improve vertical profiles of temperature and dewpoint. We focus on different ML architectures, loss functions, and input features to optimize predictions. Because we are predicting vertical profiles at 256 levels in the atmosphere, this work provides a unique perspective at using ML for 1D tasks. Compared to baseline profiles from the Rapid Refresh (RAP), ML predictions offer the largest improvement for dewpoint, particularly in the middle and upper atmosphere. Temperature improvements are modest, but CAPE values are improved by up to 40%. Feature importance analyses indicate that the ML models are primarily improving incoming RAP biases. While additional model and satellite data offer some improvement to the predictions, architecture choice is more important than feature selection in fine-tuning the results. Our proposed deep residual U-Net performs the best by leveraging spatial context from the input RAP profiles; however, the results are remarkably robust across model architecture. Further, uncertainty estimates for every level are well calibrated and can provide useful information to forecasters. 
    more » « less
  5. ; (, Journal of Geophysical Research: Atmospheres)
    Abstract This study estimates the stratosphere–troposphere exchange (STE) of air masses and ozone concentrations averaged over 2007 to 2010 using the Modern Era Retrospective‐Analyses for Research and Applications 2 (MERRA2) and ERA5 reanalyses, and observations. The latter includes Microwave Limb Sounder (MLS) for ozone, MLS and Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC) for temperatures, and A‐Train measurements for diabatic heating. The extratropical downward ozone fluxes are 538 Tg year−1from the ERA5 reanalysis, 543 Tg year−1from the MERRA2 reanalysis, and 528–539 Tg year−1from the observations, consistent with previous studies. Previous studies, however, did not consider tropical upward ozone flux. Here we show that the tropical upward ozone flux is 183–193 Tg year−1, which compensates about 35% of the extratropical downward ozone fluxes and should not be neglected. After considering the tropical upward ozone flux, the global ozone STE is 346 Tg year−1from the ERA5 reanalysis, 360 Tg year−1from the MERRA2 reanalysis, and 336–346 Tg year−1from the observations. Those estimates (347 ± 12 Tg year−1) can be used as the contribution of ozone STE to the tropospheric ozone budget. We also investigate cloud radiative effects on the STE of air mass and ozone. At 380 K, cloud radiative effects enhance downward fluxes in the extratropics from both reanalyses and observation, but reduce and enhance upward fluxes in the tropics from reanalyses and observation, respectively. The discrepancy in the tropics is related to the tropical tropopause layer thin cirrus that is missing in the reanalyses. We find that cloud radiative effects enhance the global ozone STE by about 21%–29%. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email