Stable a posteriori LES of 2D turbulence using convolutional neural networks: Backscattering analysis and generalization to higher Re via transfer learning
Authors:
Yifei Guan, Ashesh Chattopadhyay, Adam Subel, and Pedram Hassanzadeh
Abstract:
There is a growing interest in developing data-driven subgrid-scale (SGS) models for large-eddy simulation (LES) using machine learning (ML). In a priori (offline) tests, some recent studies have found ML-based data-driven SGS models that are trained on high-fidelity data (e.g., from direct numerical simulation, DNS) to outperform baseline physics-based models and accurately capture the inter-scale transfers, both forward (diffusion) and backscatter. While promising, instabilities in a posteriori (online) tests and inabilities to generalize to a different flow (e.g., with a higher Reynolds number, Re) remain as major obstacles in broadening the applications of such data-driven SGS models. For example, many of the same aforementioned studies have found instabilities that required often ad-hoc remedies to stabilize the LES at the expense of reducing accuracy.
Machine Learning Gravity Wave Parameterization Generalizes to Capture the QBO and Response to Increased CO2
Authors:
Zachary Espinosa, Aditi Sheshadri, Gerald Cain, Edwin Gerber, and Kevin DallaSanta
Abstract:
We present single-column gravity wave parameterizations (GWPs) that use machine learning to emulate non-orographic gravity wave (GW) drag and demonstrate their ability to generalize out-of-sample.
Discovery of interpretable structural model errors by combining Bayesian sparse regression and data assimilation: A chaotic Kuramoto-Sivashinsky test case
Authors:
Rambod Mojgani, Ashesh Chattopadhyay, and Pedram Hassanzadeh
Abstract:
Models of many engineering and natural systems are imperfect. The discrepancy between the mathematical representations of a true physical system and its imperfect model is called the model error. These model errors can lead to substantial differences between the numerical solutions of the model and the state of the system, particularly in those involving nonlinear, multi-scale phenomena. Thus, there is increasing interest in reducing model errors, particularly by leveraging the rapidly growing observational data to understand their physics and sources.
Observed and Modeled Mountain Waves from the Surface to the Mesosphere Near the Drake Passage
Authors:
Christopher G. Kruse, M. Joan Alexander, Lars Hoffmann, Annelize van Niekerk, Inna Polichtchouk, Julio T. Bacmeister, Laura Holt, Riwal Plougonven, Petr Šácha, Corwin Wright, Kaoru Sato, Ryosuke Shibuya, Sonja Gisinger, Manfred Ern, Catrin I. Meyer, and Olaf Stein
Abstract:
Four state-of-the-science numerical weather prediction (NWP) models were used to perform mountain wave- (MW) resolving hind-casts over the Drake Passage of a 10-day period in 2010 with numerous observed MW cases. The Integrated Forecast System (IFS) and the Icosahedral Nonhydrostatic (ICON) model were run at Δx ≈ 9 and 13 km globally. TheWeather Research and Forecasting (WRF) model and the Met Office Unified Model (UM) were both configured with a Δx = 3 km regional domain. All domains had tops near 1 Pa (z ≈ 80 km). These deep domains allowed quantitative validation against Atmospheric InfraRed Sounder (AIRS) observations, accounting for observation time, viewing geometry, and radiative transfer.
A QBO cookbook: Sensitivity of the Quasi-Biennial Oscillation to resolution, resolved waves, and parameterized gravity waves
Authors:
Chaim I. Garfinkel, Edwin P. Gerber, Ofer Shamir, Jian Rao, Martin Jucker, Ian White, Nathan Paldor
Abstract:
The most prominent mode of variability in the tropical stratosphere is the quasi-biennial oscillation (QBO), characterized by easterly and westerly winds alternating sign every 14 months. Only relatively recently have comprehensive models begun to simulate a QBO spontaneously, and even in these models the representation of the QBO typically suffers from biases. Here we elucidate the sensitivities of the QBO to a wide range of model parameters, and explore how these parameters affect the QBO behavior. We expect that these results will be helpful for improving the QBO in more comprehensive models.
Can we improve the realism of gravity wave parameterizations by imposing sources at all altitudes in the atmosphere?
Authors:
B Ribstein, C Millet, F Lott, A de la Cámara
Abstract:
Gravity waves are fluctuations in the atmosphere (seen in the temperature, wind velocity, and pressure fields) that transport energy and momentum from their sources in the troposphere and middle atmosphere to their sinks in the middle atmosphere. This way they exert a profound influence on the global circulation. Due to their relative small spatial scales, atmospheric general circulation models do not explicitly resolve these waves, and their effects on the circulation resolved by the model need to be parameterized. Parameterizations of gravity waves generated by fronts and flow imbalances typically assume that wave sources are at a certain vertical level in the troposphere, which is easy to implement but neglects the fact that these processes can occur at all altitudes in the atmosphere. In this study, we explore to which extent parameterizations of gravity wave due to fronts and flow imbalances can be improved by allowing waves to be emitted from all model levels. Our results show evidence of modest corrections of some model biases, and a clear improvement in the parameterized gravity waves energy spectra.
Mountain waves produced by a stratified shear flow with a boundary layer. Part III: Trapped lee waves and horizontal momentum transport
Authors:
Clément Soufflet, François Lott, and Bruno Deremble
Abstract:
The boundary layer theory for non-hydrostatic mountain waves presented in Part II is extended to include upward propagating gravity waves and trapped lee waves. To do so, the background wind with constant shear used in Part II is smoothly curved and become constant above a “boundary-layer” height which is much larger than the inner layer scale.
Mountain Waves Produced by a Stratified Shear Flow with a Boundary Layer. Part II: Form Drag, Wave Drag, and Transition from Downstream Sheltering to Upstream Blocking
Authors:
François Lott, Bruno Deremble, and Clément Soufflet
Abstract:
The nonhydrostatic version of the mountain flow theory presented in Part I is detailed. In the near-neutral case, the surface pressure decreases when the flow crosses the mountain to balance an increase in surface friction along the ground. This produces a form drag that can be predicted qualitatively. When stratification increases, internal waves start to control the dynamics and the drag is due to upward-propagating mountain waves as in Part I. The reflected waves nevertheless add complexity to the transition.
Seasonal and Latitudinal Variability of the Gravity Wave Spectrum in the Lower Stratosphere
Authors
Erik A. Lindgren, Aditi Sheshadri, Aurélien Podglajen, Robert W. Carver
Abstract:
Superpressure balloon data of unprecedented coverage from Loon LLC is used to investigate the seasonal and latitudinal variability of lower stratospheric gravity waves over the entire intrinsic frequency spectrum. We show that seasonal variability in both gravity wave amplitudes and spectral slopes exist for a wide range of intrinsic frequencies and provide estimates of spectral slopes in five latitudinal regions for all four seasons, in five different frequency windows.