scholarly journals Robustness of neural network emulations of radiative transfer parameterizations in a state-of-the-art general circulation model

2021 ◽  
Vol 14 (12) ◽  
pp. 7425-7437 ◽  
Author(s):  
Alexei Belochitski ◽  
Vladimir Krasnopolsky
Keyword(s):  
Neural Network ◽  
Radiative Transfer ◽  
General Circulation Model ◽  
Network Architecture ◽  
General Circulation ◽  
State Of The Art ◽  
Circulation Model ◽  
Physically Based ◽  
The Stability ◽  
Model Components

Abstract. The ability of machine-learning-based (ML-based) model components to generalize to the previously unseen inputs and its impact on the stability of the models that use these components have been receiving a lot of recent attention, especially in the context of ML-based parameterizations. At the same time, ML-based emulators of existing physically based parameterizations can be stable, accurate, and fast when used in the model they were specifically designed for. In this work we show that shallow-neural-network-based emulators of radiative transfer parameterizations developed almost a decade ago for a state-of-the-art general circulation model (GCM) are robust with respect to the substantial structural and parametric change in the host model: when used in two 7-month-long experiments with a new GCM, they remain stable and generate realistic output. We concentrate on the stability aspect of the emulators' performance and discuss features of neural network architecture and training set design potentially contributing to the robustness of ML-based model components.

scholarly journals Robustness of Neural Network Emulations of Radiative Transfer Parameterizations in a State-of-the-Art General Circulation Model

2021 ◽  
Author(s):  
Alexei Belochitski ◽  
Vladimir Krasnopolsky
Keyword(s):  
Neural Network ◽  
Radiative Transfer ◽  
Network Architecture ◽  
General Circulation ◽  
State Of The Art ◽  
Circulation Model ◽  
Set Design ◽  
Neural Network Architecture ◽  
Parametric Change ◽  
Model Components

Abstract. The ability of Machine-Learning (ML) based model components to generalize to the previously unseen inputs, and the resulting stability of the models that use these components, has been receiving a lot of recent attention, especially when it comes to ML-based parameterizations. At the same time, ML-based emulators of existing parameterizations can be stable, accurate, and fast when used in the model they were specifically designed for. In this work we show that shallow-neural-network-based emulators of radiative transfer parameterizations developed almost a decade ago for a state-of-the-art GCM are robust with respect to the substantial structural and parametric change in the host model: when used in two seven month-long experiments with the new model, they not only remain stable, but generate realistic output. Aspects of neural network architecture and training set design potentially contributing to stability of ML-based model components are discussed.

scholarly journals Solar-locked and geographical atmospheric structures inferred from a Venus general circulation model with radiative transfer

Icarus ◽  
2019 ◽  
Vol 321 ◽  
pp. 232-250 ◽  
Author(s):  
Masaru Yamamoto ◽  
Kohei Ikeda ◽  
Masaaki Takahashi ◽  
Takeshi Horinouchi
Keyword(s):  
Radiative Transfer ◽  
General Circulation Model ◽  
General Circulation ◽  
Circulation Model

Developing a data-driven ocean model

2020 ◽  
Author(s):  
Rachel Furner ◽  
Peter Haynes ◽  
Dan Jones ◽  
Dave Munday ◽  
Brooks Paige ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
General Circulation Model ◽  
General Circulation ◽  
Climate Models ◽  
Circulation Model ◽  
Machine Learning Algorithms ◽  
Data Driven ◽  
Geophysical Research ◽  
Weather And Climate

<p>The recent boom in machine learning and data science has led to a number of new opportunities in the environmental sciences. In particular, climate models represent the best tools we have to predict, understand and potentially mitigate climate change, however these process-based models are incredibly complex and require huge amounts of high-performance computing resources. Machine learning offers opportunities to greatly improve the computational efficiency of these models.</p><p>Here we discuss our recent efforts to reduce the computational cost associated with running a process-based model of the physical ocean by developing an analogous data-driven model. We train statistical and machine learning algorithms using the outputs from a highly idealised sector configuration of general circulation model (MITgcm). Our aim is to develop an algorithm which is able to predict the future state of the general circulation model to a similar level of accuracy in a more computationally efficient manner.</p><p>We first develop a linear regression model to investigate the sensitivity of data-driven approaches to various inputs, e.g. temperature on different spatial and temporal scales, and meta-variables such as location information. Following this, we develop a neural network model to replicate the general circulation model, as in the work of Dueben and Bauer 2018, and Scher 2018.</p><p>We present a discussion on the sensitivity of data-driven models and preliminary results from the neural network based model.</p><p> </p><p><em>Dueben, P. D., & Bauer, P. (2018). Challenges and design choices for global weather and climate models based on machine learning. Geoscientific Model Development, 11(10), 3999-4009.</em></p><p><em>Scher, S. (2018). Toward Data‐Driven Weather and Climate Forecasting: Approximating a Simple General Circulation Model With Deep Learning. Geophysical Research Letters, 45(22), 12-616.</em></p>

scholarly journals Impact of an improved shortwave radiation scheme in the MAECHAM5 General Circulation Model

2007 ◽  
Vol 7 (10) ◽  
pp. 2503-2515 ◽  
Author(s):  
C. Cagnazzo ◽  
E. Manzini ◽  
M. A. Giorgetta ◽  
P. M. De F. Forster ◽  
J. J. Morcrette
Keyword(s):  
Radiative Transfer ◽  
General Circulation Model ◽  
Spectral Resolution ◽  
General Circulation ◽  
Circulation Model ◽  
Shortwave Radiation ◽  
Transfer Model ◽  
Ozone Absorption ◽  
Zonal Winds ◽  
Radiation Scheme

Abstract. In order to improve the representation of ozone absorption in the stratosphere of the MAECHAM5 general circulation model, the spectral resolution of the shortwave radiation parameterization used in the model has been increased from 4 to 6 bands. Two 20-years simulations with the general circulation model have been performed, one with the standard and the other with the newly introduced parameterization respectively, to evaluate the temperature and dynamical changes arising from the two different representations of the shortwave radiative transfer. In the simulation with the increased spectral resolution in the radiation parameterization, a significant warming of almost the entire model domain is reported. At the summer stratopause the temperature increase is about 6 K and alleviates the cold bias present in the model when the standard radiation scheme is used. These general circulation model results are consistent both with previous validation of the radiation scheme and with the offline clear-sky comparison performed in the current work with a discrete ordinate 4 stream scattering line by line radiative transfer model. The offline validation shows a substantial reduction of the daily averaged shortwave heating rate bias (1–2 K/day cooling) that occurs for the standard radiation parameterization in the upper stratosphere, present under a range of atmospheric conditions. Therefore, the 6 band shortwave radiation parameterization is considered to be better suited for the representation of the ozone absorption in the stratosphere than the 4 band parameterization. Concerning the dynamical response in the general circulation model, it is found that the reported warming at the summer stratopause induces stronger zonal mean zonal winds in the middle atmosphere. These stronger zonal mean zonal winds thereafter appear to produce a dynamical feedback that results in a dynamical warming (cooling) of the polar winter (summer) mesosphere, caused by an increased downward (upward) circulation in the winter (summer) hemisphere. In addition, the comparison of the two simulations performed with the general circulation model shows that the increase in the spectral resolution of the shortwave radiation and the associated changes in the cloud optical properties result in a warming (0.5–1 K) and moistening (3%–12%) of the upper tropical troposphere. By comparing these modeled differences with previous works, it appears that the reported changes in the solar radiation scheme contribute to improve the model mean temperature also in the troposphere.

Tropical Tropospheric-Only Responses to Absorbing Aerosols

2012 ◽  
Vol 25 (7) ◽  
pp. 2471-2480 ◽  
Author(s):  
Geeta G. Persad ◽  
Yi Ming ◽  
V. Ramaswamy
Keyword(s):  
Black Carbon ◽  
General Circulation Model ◽  
General Circulation ◽  
Large Scale ◽  
Substantial Reduction ◽  
Circulation Model ◽  
Cloud Amount ◽  
Radiative Heating ◽  
Physically Based ◽  
Absorbing Aerosols

Abstract Absorbing aerosols affect the earth’s climate through direct radiative heating of the troposphere. This study analyzes the tropical tropospheric-only response to a globally uniform increase in black carbon, simulated with an atmospheric general circulation model, to gain insight into the interactions that determine the radiative flux perturbation. Over the convective regions, heating in the free troposphere hinders the vertical development of deep cumulus clouds, resulting in the detrainment of more cloudy air into the large-scale environment and stronger cloud reflection. A different mechanism operates over the subsidence regions, where heating near the boundary layer top causes a substantial reduction in low cloud amount thermodynamically by decreasing relative humidity and dynamically by lowering cloud top. These findings, which align well with previous general circulation model and large-eddy simulation calculations for black carbon, provide physically based explanations for the main characteristics of the tropical tropospheric adjustment. The implications for quantifying the climate perturbation posed by absorbing aerosols are discussed.

Sign in / Sign up

Export Citation Format

Share Document