Using deep learning to emulate and accelerate a radiative-transfer model

Using deep learning to emulate and accelerate a radiative-transfer model
复制标题

使用深度学习来模拟和加速辐射传输模型

DOI:
10.1175/jtech-d-21-0007.1
复制
发表时间:
2021
影响因子:
2.2
通讯作者:
Hagerty, Venita
Hagerty, Venita
中科院分区:
地球科学4区
文献类型:
--
作者:
Lagerquist, Ryan;Turner, David;Ebert-Uphoff, Imme;Stewart, Jebb;Hagerty, Venita

文献摘要

参考文献

被引文献

相似文献

本文描述了U-net++模型的开发,这是一种执行深度学习的神经网络,用于模拟短波快速辐射传输模型(RRTM)。我们的目标是在一小部分计算时间内准确地模拟RRTM,创建一个可用作数值天气预报(NWP)参数化的U-net++。目标变量是地面下降流通量、大气层顶上升流通量()、净通量和辐射加热率剖面。我们已经设计了几种方法来使U-net++模型知识导向,最近被确定为机器学习(ML)应用于地球科学的关键优先事项。我们进行了两个实验,以找到最佳的U-net++配置。在实验1中,我们在非热带网站上进行训练,并在热带网站上进行测试,以评估极端的空间泛化。在实验2中,我们在所有地区的网站上进行训练,并在所有地区的不同网站上进行测试,目的是创建用于NWP的最佳模型。从实验1中选择的模式显示了令人印象深刻的技巧,在热带测试站点,除了四个显着的缺陷:大的偏差和错误的加热率在平流层上部,不可靠的配置文件与单层液体云,大的加热率偏差在对流层中层的配置文件与多层液体云,和负偏差在低天顶角的所有通量分量和对流层加热率。从实验2中选择的模型纠正了除第一个缺陷之外的所有缺陷,并且两个模型的运行速度都比RRTM快约104倍。我们的代码是公开的。
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 nontropical 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, unreliablefor profiles with single-layer liquid cloud, large heating-rate bias in the midtroposphere for profiles with multilayer liquid cloud, and negative bias at low zenith 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 ~104times faster than the RRTM. Our code is available publicly.
ARM 程序中的光谱辐射测量和分析
DOI: --
发表时间: 2016
期刊:
影响因子: --
作者:
E. Mlawer;D. Turner
通讯作者: D. Turner
DOI: 10.1016/j.envsoft.2020.104856
发表时间: 2020-09
期刊: Environ. Model. Softw.
影响因子: --
作者:
M. Sadeghi;P. Nguyen;K. Hsu;S. Sorooshian
通讯作者: M. Sadeghi;P. Nguyen;K. Hsu;S. Sorooshian
DOI: 10.1175/mwr-d-18-0316.1
发表时间: 2019-08-01
影响因子: 3.2
作者:
Gagne, David John, II;Haupt, Sue Ellen;Thompson, Gregory
通讯作者: Thompson, Gregory
用于检测可能发生对流的区域的深度学习方法
DOI: --
发表时间: 2020
期刊:
影响因子: --
作者:
J. Stewart;C. Kumler;D. Hall;M. Govett
通讯作者: M. Govett
DOI: 10.1029/2018gl078202
发表时间: 2018-06-16
影响因子: 5.2
作者:
Gentine, P.;Pritchard, M.;Yacalis, G.
通讯作者: Yacalis, G.