3D volumetric tomography of clouds using machine learning for climate analysis

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: March 10, 2025

Abstract The prediction of climate has been a long-standing problem in contemporary science. One the reasons stems from gap ability to obtain 3D mapping clouds, especially shallow scattered clouds. These clouds are strongly affected by mixing processes with their surroundings, rendering internal volumetric structure highly heterogeneous. heterogeneous modulate incoming solar energy and outgoing long-wave radiation, thereby having crucial role system. However, is major challenge. Here, we combine machine learning space engineering enable, for first time, scatterers We employ ten nano-satellites formation simultaneously view same per scene different angles recover which derive statistics, including uncertainty. demonstrate this on real-world data. results provide key features predicting precipitation renewable energy.

Language: Английский

---

Satellite Retrievals Show Adiabatic Fraction of Marine Low Clouds Decreasing With Increasing Temperature and Height Above Cloud Base

Journal of Geophysical Research Atmospheres, Journal Year: 2025, Volume and Issue: 130(5)

Published: March 8, 2025

Abstract Cloud adiabatic fraction (f ad ) is an important metric that quantitatively characterizes the impact of atmospheric mixing on cloud thermodynamic properties. Due to lack vertical profiling water, previous studies variability f within clouds have been confined single scales. Our prior research achieved a breakthrough in large‐scale retrieval fraction, while it only provided two‐dimensional information leaving variation in‐cloud unquantified. In this study, utilizing global‐scale data derived from our research, we developed predictive function for global marine low‐cloud based geometric thickness and base temperature (CBT). This enabled us successfully estimate distributions across various conditions scale. The exhibits quadratic reduction top, which more pronounced at higher temperatures. Specifically, as CBT rises 2 24°C, diminishes 0.85 0.23. decreasing trend with increasing temperatures expected reduce albedo coverage potentially constituting positive radiative feedback mechanism.

Language: Английский

---

Distinct Mixing Regimes in Shallow Cumulus Clouds

Geophysical Research Letters, Journal Year: 2024, Volume and Issue: 51(2)

Published: Jan. 19, 2024

Abstract Understanding the nature of mixing between cloudy air and its surroundings is an important yet, open question. In this research, we use high‐resolution (10 m) bin‐microphysics Large Eddy Simulation a cumulus cloud, together with Lagrangian passive tracer tracking method, to study mixing. We analyze tracers as function their trajectories thermodynamic conditions they undergo inside outside cloud. Three main regimes (core, periphery, skin) are identified, each determining subset similar trajectories. These can be observed throughout cloud's lifetime provide evidence for presence undiluted core in shallow clouds. At dissipation stage, fourth regime identified: cloud‐top entrainment followed by downdrafts.

Language: Английский

---

The Role of the Toroidal Vortex in Cumulus Clouds' Entrainment and Mixing

Journal of Geophysical Research Atmospheres, Journal Year: 2024, Volume and Issue: 129(14)

Published: July 10, 2024

Abstract Shallow convective clouds play a crucial role in Earth's energy budget, as they modulate the radiative transfer atmosphere and participate vertical transport of aerosols, energy, humidity. The parameterizations representing these complex, vital players weather climate models are mostly based on description steady‐state plumes source major uncertainty. Recently, several studies have shown that buoyant thermals inherent atmospheric convection contain toroidal (ring) vortex. This work those vortices growing shallow cumulus (Cu) using high‐resolution (10 m) Large Eddy Simulations resolve much detail. Recent analysis such data showed small‐scale turbulent diffusion is unable to explain large diluted portion cloud. Here we advocate for important Cu vortex (TV) cloud dilution present complex dynamics structure TV. Nevertheless, since dominates cloud's dilution, simplicity emerges when considering lateral mass flux profile. mixing quantified direct calculations Eulerian tracers. In addition, Lagrangian tracers used identify origin entrained air its thermodynamic properties. It shows most by not recycled vortex, yet significantly more humid than environment. We suggest development new describing thermals, together with their vortices, might improve models.

Language: Английский

---

Variable Imaging Projection Cloud Scattering Tomography

IEEE Transactions on Pattern Analysis and Machine Intelligence, Journal Year: 2022, Volume and Issue: unknown, P. 1 - 12

Published: Aug. 2, 2022

Scattering-based computed tomography (CT) recovers a heterogeneous volumetric scattering medium using images taken from multiple directions. It is nonlinear problem. Prior art mainly approached it by explicit physics-based optimization of image-fitting, being slow and difficult to scale. Scale particularly important when the objects constitute large cloud fields, where recovery for climate studies. Besides speed, imaging need be flexible, efficiently handle variable viewing geometries resolutions. These can caused perturbation in camera poses or fusion data different types observational sensors. There fast projection clouds (VIP-CT). We develop learning-based solution, deep-neural network (DNN) which trains on labeled dataset. The DNN parameters are oblivious domain scale, hence work with arbitrarily domains. VIP-CT offers much better quality than state art. inference speed flexibility make effectively real-time context spaceborne observations. paper first demonstrate CT real empirical directly DNN. may offer model solution problems other scientific Our code available online.

Language: Английский

---

Retrieving 3D distributions of atmospheric particles using Atmospheric Tomography with 3D Radiative Transfer – Part 1: Model description and Jacobian calculation

Atmospheric measurement techniques, Journal Year: 2023, Volume and Issue: 16(7), P. 1803 - 1847

Published: April 5, 2023

Abstract. Our global understanding of clouds and aerosols relies on the remote sensing their optical, microphysical, macrophysical properties using, in part, scattered solar radiation. These retrievals assume that form plane-parallel, homogeneous layers utilize 1D radiative transfer (RT) models, limiting detail can be retrieved about 3D variability cloud aerosol fields inducing biases for highly heterogeneous structures such as cumulus smoke plumes. To overcome these limitations, we introduce validate an algorithm retrieving optical or microphysical atmospheric particles using multi-angle, multi-pixel radiances a RT model. The retrieval software, which have made publicly available, is called Atmospheric Tomography with Radiative Transfer (AT3D). It uses iterative, local optimization technique to solve generalized least squares problem thereby find best-fitting state. iterative fast, approximate Jacobian calculation, extended from Levis et al. (2020) accommodate open periodic horizontal boundary conditions (BCs) improved treatment non-black surfaces. We validated accuracy calculation derivatives respect both volume extinction coefficient parameters controlling across media range depths single-scattering it accurate majority over oceanic Relative root mean square errors cloud-like increase 2 % 12 maximum (MODs) medium 0.2 100.0 surfaces Lambertian albedos <0.2. Over 0.7, 20 %. Errors exceed 50 %, unless plane-parallel providing are optically very thin (∼0.1). use theory linear inverse provide insight into physical processes control tomography identify its supported by numerical experiments. show matrix becomes increasing ill-posed size increases forward-scattering peak phase function decreases. This suggests tomographic will become increasingly difficult thicker. Retrievals asymptotically thick likely require other sources information successful. In Loveridge (2023a; hereafter Part 2), examine how varies target synthetic data. do this explore error nature inversion limit affect retrieval. also assess compare them transfer.

Language: Английский

---

Emission Reductions Significantly Reduce the Hemispheric Contrast in Cloud Droplet Number Concentration in Recent Two Decades

Journal of Geophysical Research Atmospheres, Journal Year: 2023, Volume and Issue: 128(2)

Published: Jan. 7, 2023

Abstract Anthropogenic activities have drastically impacted the climate system since Industrial Revolution. However, to what extent anthropogenic emissions influence cloud droplet number concentration ( N d ), critical parameter for understanding aerosol‐cloud interactions, is poorly known on hemispheric scale due considerable retrieval uncertainty. We employed multiple widely used sampling methods evaluate long‐term trend in contrast (Δ d(NH‐SH) ) between Northern Hemisphere (NH) and Southern (SH). Here we show that Δ was halved from 2003 2020 using different channels, even though range of magnitudes large. Such dramatic changes are dominated by significantly decreased over NH (∼20%) emission reductions compared relatively stable pristine nature SH. Aerosol optical depth (AOD) aerosol index (AI) correlate with based trends, they replicate trends. This poor correlation partly contributed stratospheric smoke wildfires Australia had little The northwest Atlantic shows largest contribution, ∼38%, trend, whereas Pacific dominates change AOD AI, contributing more than 60% ∼50% AI NH. Our results imply reduced provide strong observational evidence extensively altered liquid clouds last two decades.

Language: Английский

---

Do Optically Denser Trade‐Wind Cumuli Live Longer?

Geophysical Research Letters, Journal Year: 2023, Volume and Issue: 50(13)

Published: June 29, 2023

Abstract This study investigates the lifetime and temporal evolution of physical properties trade‐wind cumuli based on tracking individual clouds in observations with Advanced Baseline Imager aboard geostationary GOES‐16 satellite during “ElUcidating RolE Cloud–Circulation Coupling ClimAte” (EUREC 4 A) campaign east Barbados winter 2020. A first application our upgraded cloud‐tracking toolbox to measurements high spatio‐temporal resolution (2 × 2 km 1 min) provides probability density functions area that develop as a consequence meso‐to‐synoptic scale motions. By separately considering exist daytime live distinct intervals, we find shallow marine longer when they cover larger surface show higher cloud optical thickness (COT). Besides effect COT, atmospheric motions which interact is also critical their lifetime.

Language: Английский

---

The Temperature Control of Cloud Adiabatic Fraction and Coverage

Geophysical Research Letters, Journal Year: 2023, Volume and Issue: 50(22)

Published: Nov. 17, 2023

Abstract The fraction of cloud water compared to its adiabatic value is defined as the fraction, f ad . accuracy representation in climate models highly sensitive mixing rate, manifested Here, we present first distribution marine boundary layer clouds over global oceans, retrieved by satellite observations. shown decrease exponentially with base temperature (CBT) and depth, agreement increasing evaporation capacity entrained warmer air. Cloud cover decreases CBT, but a much lesser extent than dependence on CBT has little relative humidity or precipitation. relationship between highlights importance core control factor evaporation. simultaneous content can lead positive feedback, resulting greater future warming.

Language: Английский

---

Convective and Turbulent Motions in Nonprecipitating Cu. Part II: LES Simulated Cloud Represented by a Starting Plume

Journal of the Atmospheric Sciences, Journal Year: 2021, Volume and Issue: 79(3), P. 793 - 813

Published: Dec. 22, 2021

Abstract The dynamic structure of a small trade wind cumulus (Cu) is analyzed using novel approach. Cu developing in shear-free environment simulated by 10-m-resolution LES model with spectral bin microphysics. aim to clarify the dynamical nature cloud updraft zone (CUZ) including entrainment and mixing growing Cu. validity concept stating that at state can be represented parcel or jet tested. To investigate CUZ performed motions scales larger than turbulence scales, modeled fields air velocity were filtered wavelet filter separated convective from turbulent ones. Two types objects investigated: volume ascending maximal (point parcel) CUZ. It was found point representing upper part core adiabatic. motion this base determines cloud-top height. top-hat (i.e., averaged) values adiabatic fraction are substantially lower those parcel. Evaluation terms equation typically used 1D models show applied for calculation vertical velocities stage Cu, least up heights inversion layer. Dynamically, resembles starting plume tail nonstationary jet. Both buoyancy acceleration linearly increase height, An important finding lateral (nonturbulent) has little effect on cannot explain changes conservative variables q t θ l . In contrast, entrained lifting inside decreases liquid water content its fraction. Possible reasons these effects discussed. Significance Statement (i) study improves understanding mixing. (ii) shows dominating role convective-scale microphysics dynamics. (iii) comparison results large-eddy simulations simple allows evaluating current schemes parameterization.

Language: Английский

---