Predicting atmospheric optical properties for radiative transfer computations using neural networks
Wageningen University and Research（瓦赫宁根大学及研究中心）；Cooperative Institute for Research in Environmental Sciences, University of Colorado at Boulder（科罗拉多大学博尔德分校环境科学合作研究所）；University of Colorado at Boulder（科罗拉多大学博尔德分校）；NOAA Physical Sciences Laboratory（美国国家海洋和大气管理局物理科学实验室）；SURFsara（SURFsara科研网络机构）
Menno A. Veerman, Robert Pincus, Robin Stoffer, Caspar van Leeuwen, Damian Podareanu, Chiel C. van Heerwaarden
备注：
13 pages,5 figures, submitted to Philosophical Transactions A
The radiative transfer equations are well-known, but radiation parametrizations in atmospheric models are computationally expensive. A promising tool for accelerating parametrizations is the use of machine learning techniques. In this study, we develop a machine learning-based parametrization for the gaseous optical properties by training neural networks to emulate a modern radiation parameterization (RRTMGP). To minimize computational costs, we reduce the range of atmospheric conditions for which the neural networks are applicable and use machine-specific optimised BLAS functions to accelerate matrix computations. To generate training data, we use a set of randomly perturbed atmospheric profiles and calculate optical properties using RRTMGP. Predicted optical properties are highly accurate and the resulting radiative fluxes have average errors within \SI{0.5}{\flux} compared to RRTMGP. Our neural network-based gas optics parametrization is up to 4 times faster than RRTMGP, depending on the size of the neural networks. We further test the trade-off between speed and accuracy by training neural networks for the narrow range of atmospheric conditions of a single large-eddy simulation, so smaller and therefore faster networks can achieve a desired accuracy. We conclude that our machine learning-based parametrization can speed-up radiative transfer computations whilst retaining high accuracy.
Virginia Military Institute（弗吉尼亚军事学院）；University of Virginia（弗吉尼亚大学）
As electro-optical energy from the sun propagates through the atmosphere it is affected by radiative transfer effects including absorption, emission, and scattering. Modeling these affects is essential for scientific remote sensing measurements of the earth and atmosphere. For example, hyperspectral imagery is a form of digital imagery collected with many, often hundreds, of wavelengths of light in pixel. The amount of light measured at the sensor is the result of emitted sunlight, atmospheric radiative transfer, and the reflectance off the materials on the ground, all of which vary per wavele…
TU Darmstadt（达姆施塔特工业大学）；Mila（米拉人工智能研究所）；Université de Montréal（蒙特利尔大学）；Environment and Climate Change Canada（加拿大环境与气候变化部）；McGill University（麦吉尔大学）
Numerical simulations of Earth's weather and climate require substantial amounts of computation. This has led to a growing interest in replacing subroutines that explicitly compute physical processes with approximate machine learning (ML) methods that are fast at inference time. Within weather and climate models, atmospheric radiative transfer (RT) calculations are especially expensive. This has made them a popular target for neural network-based emulators. However, prior work is hard to compare due to the lack of a comprehensive dataset and standardized best practices for ML benchmarking. To …
Changes in the atmospheric composition alter the magnitude and partitioning between the downward propagating solar and atmospheric longwave radiative fluxes heating the Earth's surface. These changes are computed by radiative transfer codes in Global Climate Models, and measured with high precision at surface observation networks. Changes in radiative heating signify changes in the global surface temperature and hydrologic cycle. Here, we develop a conceptual framework using an Energy Balance Model to show that first order changes in the hydrologic cycle are mainly associated with changes in s…
Recently, Smirnov published a paper (B. M. Smirnov, "Collision and radiative processes in emission of atmospheric carbon dioxide", 2018, J. Phys. D.: Appl. Phys., Vol. 51, No. 21, pp. 214004) which dismisses the role of increasing concentrations of anthropogenic CO$_2$ on global warming of planet Earth. We show that these conclusions are the consequence of two flaws in Smirnov's theoretical model which neglect the effects of the increased concentrations of CO$_2$ on the absorption of Earth's blackbody radiation in the 12-15$μ$m region. The influence of doubling the concentration of CO$_2$ in t…
Observations of brown dwarfs (BDs), free-floating planetary-mass objects, and directly imaged extrasolar giant planets (EGPs) exhibit rich evidence of large-scale weather. Cloud radiative feedback has been proposed as a potential mechanism driving the vigorous atmospheric circulation on BDs and directly imaged EGPs, and yet it has not been demonstrated in three-dimensional dynamical models at relevant conditions. Here we present a series of atmospheric circulation models that self-consistently coupled dynamics with idealized cloud formation and its radiative effects. We demonstrate that vigoro…
Thousands of exoplanets have been detected to date, and with future planned missions this tally will increase. Understanding the climate dependence on the planetary parameters is vital for the study of terrestrial exoplanet habitability. Using an idealized general circulation model with a seasonal cycle, we study the seasonal response of the surface temperature and Hadley circulation to changes in the orbital, rotational and radiative timescales. Analyzing the climate's seasonal response to variations in these timescales, we find a regime transition between planets controlled by the annual mea…
Growing observational evidence has suggested active meteorology in atmospheres of brown dwarfs (BDs) and directly imaged extrasolar giant planets (EGPs). In particular, a number of surveys have shown that near-IR brightness variability is common among L and T dwarfs. Despite initial understandings of atmospheric dynamics which is the major cause of the variability by previous studies, the detailed mechanism of variability remains elusive, and we need to seek a natural, self-consistent mechanism. Clouds are important in shaping the thermal structure and spectral properties of these atmospheres …
We present EOS, a procedure for determining the Outgoing Longwave Radiation (OLR) and top-of-atmosphere (TOA) albedo for a wide range of conditions expected to be present in the atmospheres of rocky planets with temperate conditions. EOS is based on HELIOS and HELIOS-K, which are novel and publicly available atmospheric radiative transfer (RT) codes optimized for fast calculations with GPU processors. These codes were originally developed for the study of giant planets. In this paper we present an adaptation for applications to terrestrial-type, habitable planets, adding specific physical reci…
Spectroscopic measurements at top of atmosphere are uniquely capable of attributing changes in Earth's outgoing infrared radiation field to specific greenhouse gasses. The Atmospheric Infrared Sounder (AIRS) placed in orbit in 2002 has spectroscopically resolved a portion of Earth's outgoing longwave radiation for over 17 years. Concurrently, atmospheric CO$_2$ rose from 373 to 410 ppm, or 28% of the total increase over pre-industrial levels. The Coupled Model Intercomparison Project Phase 6 (CMIP6) multi-model ensemble average predicts 0.477 Wm$^{-2}$ clear-sky longwave effective radiative fo…
The Net Exchange Formulation (NEF) is an alternative to the usual radiative transfer equation. It was proposed in 1967 by Green [1] for atmospheric sciences and by Hottel [2] for engineering sciences. Until now, the NEF has been used only in a very few cases for atmospheric studies. Recently we have developped a longwave radiative code based on this formulation for a GCM of the Mars planet. Here, we will present results for the Earth atmosphere, obtained with a Monte Carlo Method based on the NEF. In this method, fluxes are not addressed any more. The basic variables are the net exchange rates…
After the sunset, under calm and clear sky conditions, aerosol laden surface air-layer, cools radiatively to the upper atmosphere. Predominant effect of the radiative cooling on the vertical temperature profile extends to several hundred meters from the surface. This results in the development of a stable, nocturnal inversion layer. However, ground surface, owing to its higher thermal inertia, lags in the cooling process. Due to this about a meter thick air layer just above the ground can be (2-6)deg Celsius cooler than the ground. Thus, at the surface an unstable convective layer is present, …
The atmospheric circulation in Venus is well known to exhibit strong super-rotation. However, the atmospheric mechanisms responsible for the formation of this super-rotation are still not fully understood. In this work, we developed a new Venus general circulation model to study the most likely mechanisms driving the atmosphere to the current observed circulation. Our model includes a new radiative transfer, convection and suitably adapted boundary layer schemes and a dynamical core that takes into account the dependence of the heat capacity at constant pressure with temperature. The new Venus…
The formation, properties and lifetime of secondary organic aerosols in the atmosphere are largely determined by gas-particle partitioning coefficients of the participating organic vapours. Since these coefficients are often difficult to measure or compute, we developed a machine learning (ML) model to predict them given molecular structure as input. Our data-driven approach is based on the dataset by Wang et al. (Atmos. Chem. Phys., 17, 7529 (2017)), who computed the partitioning coefficients and saturation vapour pressures of 3414 atmospheric oxidation products from the master chemical mecha…
A Hybrid LSTM--Vision Transformer Architecture for Predicting HRRR Forecast Errors
Atmospheric Sciences Research Center, University at Albany, SUNY（纽约州立大学奥尔巴尼分校大气科学研究中心）；University of Oklahoma（俄克拉荷马大学）；State Weather Risk Communication Center, University at Albany, SUNY（纽约州立大学奥尔巴尼分校州天气风险沟通中心）
Aerosols sourced from combustion such as black carbon (BC) are important short-lived climate forcers whose direct radiative forcing and atmospheric lifetime depend on their morphology. These aerosols' complex morphology makes modeling their optical properties difficult, contributing to uncertainty in both their direct and indirect climate effects. Accurate and fast calculations of BC optical properties are needed for remote sensing inversions and for radiative forcing calculations in atmospheric models, but current methods to accurately calculate the optical properties of these aerosols are co…
A General Strategy for Physics-Based Model Validation Illustrated with Earthquake Phenomenology, Atmospheric Radiative Transfer, and Computational Fluid Dynamics
验证通常被定义为确定模型从其预期用途的角度准确表示真实世界的程度的过程。验证是至关重要的，因为行业和政府越来越依赖计算机模型的预测来证明他们的决策。在这篇文章中，我们调查的模型验证文献，并提出制定验证作为一个迭代的建设过程，模仿的过程中发生的隐含在科学家的头脑。因此，我们提供了一个正式的代表性的逐步建立的信任模型，从而取代丧失能力的索赔的不可能验证一个给定的模型的自适应过程的建设性近似。这种方法更适合于验证的模糊、粗粒度性质。我们的程序因素的冗余度与用于验证的实验的新颖性，以及模型预测的观察结果的程度。我们说明了新的方法，首先与成熟的量子力学可以说是最好的建立物理理论，然后与几个具体的例子，从我们的一些主要的科学利益：地震的元胞自动机模型，太阳辐射传输的异常扩散模型在多云的大气，计算流体动力学代码的里希特迈尔-梅什科夫不稳定性。这篇文章是Sornette等人[2007]的扩充版本，发表在2007年的美国国家科学院院刊上（doi：10.1073/pnas.0611677104），电子增刊网址为http://www.pnas.org/cgi/content/full/0611677104/DC1。Sornette et al. [2007]也可以在physics/0511219上以预印本形式获得。
