Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Cloud phase”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

On the Links Between Ice Nucleation, Cloud Phase, and Climate Sensitivity in CESM2

Ice nucleation in mixed-phase clouds has been identified as a critical factor in projections of future climate. Here we explore how this process influences climate sensitivity using the Community Earth System Model 2 (CESM2). We find that ice nucleation affects simulated cloud feedbacks over most regions and levels of the troposphere, not just extratropical low clouds. However, with present-day global mean cloud phase adjusted to replicate satellite retrievals, similar total cloud feedback is attained whether ice nucleation is simulated as aerosol-sensitive, insensitive, or absent. These model experiments all result in a strongly positive total cloud feedback, as in the default CESM2. A microphysics update from CESM1 to CESM2 had substantially weakened ice nucleation, due partly to a model issue. Our findings indicate that this update reduced global cloud phase bias, with CESM2's high climate sensitivity reflecting improved mixed-phase cloud representation.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Important ice processes are missed by the Community Earth System Model (CESM) in Southern Ocean mixed–phase clouds: Bridging SOCRATES observations to model developments

Global climate models (GCMs) are challenged by difficulties in simulating cloud phase and cloud radiative effect over the Southern Ocean (SO). Some of the new generation GCMs predict too much liquid and too little ice in mixed-phase clouds. This misrepresentation of cloud phase in GCMs results in weaker negative cloud feedback over the SO and a higher climate sensitivity. Based on a model comparison with observational data obtained during the Southern Ocean Cloud Radiation and Aerosol Transport Experimental Study (SOCRATES), this study addresses a key uncertainty in the Community Earth System Model version 2 (CESM2) related to cloud phase, namely ice formation in pristine remote SO clouds. It is found that sea spray organic aerosols (SSOA) are the most important type of ice nucleating particles (INPs) over the SO with concentrations one order of magnitude higher than those of dust INPs based on measurements and CESM2 simulations. Secondary ice production (SIP) which includes riming splintering, rain droplet shattering, and ice-ice collisional fragmentation as implemented in CESM2 is the dominant ice production process in moderately cold clouds with cloud temperatures greater than –20°C. SIP enhances the in-cloud ice number concentrations (Ni) by 1-3 orders of magnitude and predicts more mixed-phase (with percentage occurrence increased from 15% to 21%), in better agreement with the observations. Finally, this study highlights the importance of accurately representing cloud phase over the pristine remote SO by considering the ice nucleation of SSOA and SIP processes, which are currently missing in most GCM cloud microphysics parameterizations.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2018

Distinct Contributions of Ice Nucleation, Large-Scale Environment, and Shallow Cumulus Detrainment to Cloud Phase Partitioning With NCAR CAM5

Mixed-phase clouds containing both liquid droplets and ice particles occur frequently at high latitudes and in the midlatitude storm track regions. Simulations of the cloud phase partitioning between liquid and ice hydrometeors in state-of-the-art global climate models are still associated with large biases. For this study, the phase partitioning in terms of liquid mass phase ratio (MPR liq , defined as the ratio of liquid mass to total condensed water mass) simulated from the NCAR Community Atmosphere Model version 5 (CAM5) is evaluated against the observational data from A-Train satellite remote sensors. Modeled MPR liq is significantly lower than observations on the global scale, especially in the Southern Hemisphere (e.g., Southern Ocean and the Antarctic). Sensitivity tests with CAM5 are conducted to investigate the distinct contributions of heterogeneous ice nucleation, shallow cumulus detrainment, and large-scale environment (e.g., winds, temperature, and water vapor) to the low MPR liq biases. Our results show that an aerosol-aware ice nucleation parameterization increases the MPR liq especially at temperatures colder than -20°C and significantly improves the model agreements with observations in the Polar regions in summer. The decrease of threshold temperature over which all detrained cloud water is liquid from 268 to 253 K enhances the MPR liq and improves the MPR liq mostly over the Southern Ocean. By constraining water vapor in CAM5 toward reanalysis, modeled low biases in many geographical regions are largely reduced through a significant decrease of cloud ice mass mixing ratio.

47 OTHER INSTRUMENTATION↗
DOE OSTI2020

Toward Understanding the Simulated Phase Partitioning of Arctic Single-Layer Mixed-Phase Clouds in E3SM

21 Mixed-phase clouds are widespread in the Arctic and have substantial impacts on the 22 surface energy budget, sea ice extent, and regional climate. Significant changes are found in the 23 modeled phase partitioning of Arctic mixed-phase clouds in the newly developed U.S. 24 Department of Energy (DOE) Energy Exascale Earth System Model (E3SM) Atmosphere Model 25 version 1 (EAMv1) compared to its predecessor (CAM5/EAMv0). The mixed-phase clouds 26 simulated by EAMv1 is overly dominated by super-cool liquid which is often underestimated by 27 CAM5. Compared to CAM5, EAMv1 has made several important changes in its cloud physical 28 parameterizations, including using CLUBB to unify shallow convection, PBL turbulence, and 29 cloud macrophysics, as well as updating the MG two-moment cloud microphysics scheme and 30 using the CNT heterogeneous ice nucleation scheme to replace the Meyers scheme for mixed- 31 phase clouds. To understand which of the newly-implemented physical parameterizations is 32 responsible for the change of modeled mixed-phase cloud phase partitioning in EAMv1, 33 sensitivity tests are conducted by utilizing the short-term hindcast approach and results are 34 compared to the observations from the ARM Mixed-Phase Arctic Cloud Experiment (M-PACE) 35 field campaign. 36 It is found that although each scheme contributes to the decrease of cloud ice water 37 content (IWC) and increase of cloud liquid water content (LWC), only the combined effect of 38 CLUBB and CNT can explain the significant underestimation of IWC. By examining the budgets 39 of cloud microphysical processes, CLUBB is found to play a dominant role due to the fact that 40 CLUBB does not diagnoses cloud ice phase water in its assumed joint PDF method and thus may 41 not provide sufficient cloud ice to initialize ice phase microphysical processes such as the growth 42 of ice crystals through vapor deposition. Although a more effective heterogeneous ice nucleation 43 may alleviate the underestimation of IWC, the ice phase process in CLUBB requires urgent 44 attention in the future model development.

Zhang, Meng↗
DOE OSTI2024

A Theory for the Balance between Warm Rain and Ice Crystal Processes of Precipitation in Mixed-Phase Clouds

Abstract Mixed-phase clouds contain both supercooled cloud liquid and ice crystals. In principle, precipitation may be initiated either by the liquid phase or by the ice phase. Ice crystals may grow by vapor diffusion to become snow (“ice crystal process”), forming “cold” precipitation. Equally, cloud droplets, when large enough, coalesce to form “warm” precipitation by the “warm rain process.” Warm rain could be supercooled and freeze as “warm” graupel. In the present paper, a new simplified theoretical analysis is provided to examine the microphysical system consisting of three species of hydrometeor, namely, cloud liquid, “cold ice” (crystals, snow), and “warm rain” (frozen or supercooled). This is obtained by nondimensionalizing and simplifying the evolution equations for the mass of each species. Analytical formulas are given for equilibria. Feedback analysis shows that the sign of the feedback is linked to the abundance of precipitation, with a neutral surface in the 3D phase space. The system’s precipitation amount explodes while in the initial unstable regime, crossing the neutral surface and approaching the equilibrium point that is a stable attractor. Positive and negative feedbacks are elucidated. In a standard case, the cold ice mass is about 1000 times larger than the warm rain mass. To illustrate the physical behavior of the theory, sensitivity tests are performed with respect to environmental conditions (e.g., aerosol, updraft speed) and microphysical parameters (e.g., riming and sedimentation rates for cold ice). Cold ice prevails, especially in fast ascent, due to its low bulk density, favoring slow sedimentation and a wide cross-sectional area for riming. Significance Statement The theory elucidates how the ice phase can prevail in the precipitation from any mixed-phase clouds with supercooled cloud liquid and crystals. The ice phase radically suppresses cloud liquid by riming when active and “wins” the competition against coalescence. This prevalence of ice is shown to arise from the low bulk density of snow. The cloud is viewed as a system of negative and positive feedbacks that prevail in realms of stability and instability in a 3D phase space.

Meteorology & Atmospheric Sciences↗
DOE OSTI2021

Effects of marine organic aerosols as sources of immersion-mode ice-nucleating particles on high-latitude mixed-phase clouds

Mixed-phase clouds are frequently observed in high-latitude regions and have important impacts on the surface energy budget and regional climate. Marine organic aerosol (MOA), a natural source of aerosol emitted over ~70 % of Earth's surface, may significantly modify the properties and radiative forcing of mixed-phase clouds. However, the relative importance of MOA as a source of ice-nucleating particles (INPs) in comparison to mineral dust, and MOA's effects as cloud condensation nuclei (CCN) and INPs on mixed-phase clouds are still open questions. In this study, we implement MOA as a new aerosol species into the Community Atmosphere Model version 6 (CAM6), the atmosphere component of the Community Earth System Model version 2 (CESM2), and allow the treatment of aerosol–cloud interactions of MOA via droplet activation and ice nucleation. CAM6 reproduces observed seasonal cycles of marine organic matter at Mace Head and Amsterdam Island when the MOA fraction of sea spray aerosol in the model is assumed to depend on sea spray biology but fails when this fraction is assumed to be constant. Model results indicate that marine INPs dominate primary ice nucleation below 400 hPa over the Southern Ocean and Arctic boundary layer, while dust INPs are more abundant elsewhere. By acting as CCN, MOA exerts a shortwave cloud forcing change of -2.78 W m-2 over the Southern Ocean in the austral summer. By acting as INPs, MOA enhances the longwave cloud forcing by 0.35 W m-2 over the Southern Ocean in the austral winter. The annual global mean net cloud forcing changes due to CCN and INPs of MOA are -0.35 and 0.016 W m-2, respectively. These findings highlight the vital importance for Earth system models to consider MOA as an important aerosol species for the interactions of biogeochemistry, hydrological cycle, and climate change.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2016

Impacts of the Manaus pollution plume on the microphysical properties of Amazonian warm-phase clouds in the wet season

The remote atmosphere over the Amazon can be similar to oceanic regions in terms of aerosol conditions and cloud type formations. This is especially true during the wet season. The main aerosol-related disturbances over the Amazon have both natural sources, such as dust transport from Africa, and anthropogenic sources, such as biomass burning or urban pollution. The present work considers the impacts of the latter on the microphysical properties of warm-phase clouds by analyzing observations of the interactions between the Manaus pollution plume and its surroundings, as part of the GoAmazon2014/5 Experiment. The analyzed period corresponds to the wet season (specifically from February to March 2014 and corresponding to the first Intensive Operating Period (IOP1) of GoAmazon2014/5). The droplet size distributions reported are in the range 1 µm ≤ D ≤ 50 µm in order to capture the processes leading up to the precipitation formation. The wet season largely presents a clean background atmosphere characterized by frequent rain showers. As such, the contrast between background clouds and those affected by the Manaus pollution can be observed and detailed. The focus is on the characteristics of the initial microphysical properties in cumulus clouds predominantly at their early stages. The pollution-affected clouds are found to have smaller effective diameters and higher droplet number concentrations. The differences range from 10 to 40 % for the effective diameter and are as high as 1000% for droplet concentration for the same vertical levels. The growth rates of droplets with altitude are slower for pollution-affected clouds (2.90 compared to 5.59 µm km –1 ), as explained by the absence of bigger droplets at the onset of cloud development. Clouds under background conditions have higher concentrations of larger droplets (> 20 µm) near the cloud base, which would contribute significantly to the growth rates through the collision–coalescence process. The overall shape of the droplet size distribution (DSD) does not appear to be predominantly determined by updraught strength, especially beyond the 20 µm range. The aerosol conditions play a major role in that case. However, the updraughts modulate the DSD concentrations and are responsible for the vertical transport of water in the cloud. The larger droplets found in background clouds are associated with weak water vapour competition and a bimodal distribution of droplet sizes in the lower levels of the cloud, which enables an earlier initiation of the collision–coalescence process. This paper shows that the pollution produced by Manaus significantly affects warm-phase microphysical properties of the surrounding clouds by changing the initial DSD formation. The corresponding effects on ice-phase processes and precipitation formation will be the focus of future endeavors.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Mixed‐Phase Clouds Over the Southern Ocean as Observed From Satellite and Surface Based Lidar and Radar

Abstract This study investigates the occurrence of mixed‐phase clouds (MPC, i.e., cloud layers containing both liquid and ice water at sub‐freezing temperatures) over the Southern Ocean (SO) using space‐ and surface‐based lidar and radar observations. The occurrence of supercooled clouds is dominated by geometrically thin (<1 km) layers that rarely contain ice. We diagnose layers that are geometrically thicker than 1 km to contain ice ~65% and ~4% of the time from below by surface remote sensors and from above by orbiting remote sensors, respectively. We examine the discrepancy in MPC occurrence statistics as diagnosed from below and above the cloud layer. From above, we find that MPC occurrence has a gradient associated with the Antarctic Polar Front near 55°S with a rare occurrence of satellite‐derived MPC south of that latitude. In contrast, surface sensors find ice in 33% of supercooled liquid water layers. We infer using observing system simulation experiments and data analysis that space‐based lidar cannot identify the occurrence of MPC except when secondary ice‐forming processes operate in convection that is, sufficiently strong to loft ice crystals to cloud tops. We conclude that the CALIPSO phase statistics of MPC have a severe low bias in MPC occurrence. Based on surface‐based statistics in the SO, we present a parameterization of the frequency of MPC as a function of cloud top temperature that differs substantially from that used in recent climate model simulations.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Hemispheric Asymmetry of Phase Partition in Mixed‐Phase Clouds Based on Near Global‐Scale Airborne Observations

Mixed-phase clouds contribute to substantial uncertainties in global climate models due to their complex microphysical properties. Former model evaluations almost exclusively rely on satellite observations to assess cloud phase distributions globally. This study investigated mixed-phase cloud properties using near global-scale in situ observation data sets from 14 flight campaigns in combination with collocated output from a global climate model. The Southern Hemisphere (SH) shows significantly higher occurrence frequencies and higher mass fractions of supercooled liquid water than Northern Hemisphere (NH) based on observations at 0.2 and 100 km horizontal scales. Such hemispheric asymmetry is not captured by the model. The model also consistently overestimates liquid water content (LWC) in all cloud phases but shows ice water content (IWC) biases that vary with phase. Key processes contributing to model biases in phase partition can be identified through the combination of evaluation of phase frequency, liquid mass fraction, LWC and IWC.

Yang, Ching An [San Jose State University, CA (Uni↗
DOE OSTI2023

Using Downwelling Far- and Thermal-Infrared Hyperspectral Radiance for Cloud Phase Classification in the Antarctic

The cloud phase is one of the most important parameters of clouds. In this paper, we propose a method for cloud phase classification that synergistically utilizes the far- and thermal-infrared bands based on the Atmospheric Emitted Radiance Interferometer (AERI) at the Atmospheric Radiation Measurement West Antarctic Radiation Experiment (AWARE) observatory in 2016. The possible features in the far- and thermal-infrared bands are analyzed based on the differences in the simulated cloud brightness temperature (BT) spectra with different cloud phases. Using the support vector machine (SVM) algorithm, four features are determined to identify the cloud phase, which include the BT at 900 cm -1 , the slope of the fitted function of BT in the 900–1000 cm -1 interval, the BT difference (BTD) between 512 cm -1 and 726 cm -1 , and the BTD between 550 cm -1 and 726 cm -1 . Here, the performance of the proposed method is evaluated with Shupe’s and Turner’s method. The monthly average accuracy of the proposed method, the method without the two far-infrared features, and Turner’s method are about 76%, 36%, and 49%, respectively, which infer the good performance of the proposed method and also indicate that the far-infrared band features can effectively enhance cloud phase classification. It is notable that, compared to Shupe’s method, the accuracy for the proposed method is only 61% during the Antarctic summer, which results from the definitions of cloud phase and radiative effect. In addition, the accuracy is only 44% for Turner’s method in seasons with a low frequency of mixed clouds due to the significant effect of water vapor.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2019

Dust Radiative Effects on Climate by Glaciating Mixed-Phase Clouds

Mineral dust plays an important role in the primary formation of ice crystals in mixed-phase clouds by acting as ice nucleating particles (INPs). It can influence the cloud phase transition and radiative forcing of mixed-phase clouds, both of which are crucial to global energy budget and climate. In this study, we investigate the dust indirect effects on mixed-phase clouds through heterogeneous ice nucleation with the U.S. Department of Energy (DOE) Energy Exascale Earth System Model (E3SM). Dust and INP concentrations simulated from two versions of E3SM with three ice nucleation parameterizations were evaluated against observations in the Northern Hemisphere. Constrained by these observations, E3SM shows that dust INPs induce a global mean net cloud radiative effect of 0.05 to 0.26 W/m 2 with the predominant warming appearing in the Northern Hemisphere midlatitudes. However, a cooling effect is found in the Arctic due to reduced longwave cloud forcing.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2018

Observed aerosol suppression of cloud ice in low-level Arctic mixed-phase clouds

Abstract. The interactions that occur between aerosols and a mixed-phase cloud system, and the subsequent alteration of the microphysical state of such clouds, are a problem that has yet to be well constrained. Advancing our understanding of aerosol–ice processes is necessary to determine the impact of natural and anthropogenic emissions on Earth's climate and to improve our capability to predict future climate states. This paper deals specifically with how aerosols influence ice mass production in low-level Arctic mixed-phase clouds. In this study, a 9-year record of aerosol, cloud and atmospheric state properties is used to quantify aerosol influence on ice production in mixed-phase clouds. It is found that mixed-phase clouds present in a clean aerosol state have higher ice water content (IWC) by a factor of 1.22 to 1.63 at cloud base than do similar clouds in cases with higher aerosol loading. We additionally analyze radar-derived mean Doppler velocities to better understand the drivers behind this relationship, and we conclude that aerosol induced reduction of the ice crystal nucleation rate, together with decreased riming rates in polluted clouds, are likely influences on the observed reductions in IWC.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2017

Mixed-Phase Clouds: Progress and Challenges

Mixed-phase clouds represent a three-phase colloidal system consisting of water vapor, ice particles, and coexisting supercooled liquid droplets. Mixed-phase clouds are ubiquitous in the troposphere, occurring at all latitudes from the polar regions to the tropics. Because of their widespread nature, mixed-phase processes play critical roles in the life cycle of clouds, precipitation formation, cloud electrification, and the radiative energy balance on both regional and global scales.Yet, in spite of many decades of observations and theoretical studies, our knowledge and understanding of mixed-phase cloud processes remains incomplete.Mixed-phase clouds are notoriously difficult to represent in numerical weather prediction and climatemodels, and their description in theoretical cloud physics still presents complicated challenges. In this chapter, the current status of our knowledge on mixed-phase clouds, obtained from theoretical studies and observations, is reviewed. Recent progress, along with a discussion of problems and gaps in understanding the mixed-phase environment is summarized. Specific steps to improve our knowledge of mixed-phase clouds and their role in the climate and weather system are proposed.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2016

Improving Mixed-phase Cloud Parameterization in Climate Model with the ACRF Measurements

Mixed-phase cloud microphysical and dynamical processes are still poorly understood, and their representation in GCMs is a major source of uncertainties in overall cloud feedback in GCMs. Thus improving mixed-phase cloud parameterizations in climate models is critical to reducing the climate forecast uncertainties. This study aims at providing improved knowledge of mixed-phase cloud properties from the long-term ACRF observations and improving mixed-phase clouds simulations in the NCAR Community Atmosphere Model version 5 (CAM5). The key accomplishments are: 1) An improved retrieval algorithm was developed to provide liquid droplet concentration for drizzling or mixed-phase stratiform clouds. 2) A new ice concentration retrieval algorithm for stratiform mixed-phase clouds was developed. 3) A strong seasonal aerosol impact on ice generation in Arctic mixed-phase clouds was identified, which is mainly attributed to the high dust occurrence during the spring season. 4) A suite of multi-senor algorithms was applied to long-term ARM observations at the Barrow site to provide a complete dataset (LWC and effective radius profile for liquid phase, and IWC, Dge profiles and ice concentration for ice phase) to characterize Arctic stratiform mixed-phase clouds. This multi-year stratiform mixed-phase cloud dataset provides necessary information to study related processes, evaluate model stratiform mixed-phase cloud simulations, and improve model stratiform mixed-phase cloud parameterization. 5). A new in situ data analysis method was developed to quantify liquid mass partition in convective mixed-phase clouds. For the first time, we reliably compared liquid mass partitions in stratiform and convective mixed-phase clouds. Due to the different dynamics in stratiform and convective mixed-phase clouds, the temperature dependencies of liquid mass partitions are significantly different due to much higher ice concentrations in convective mixed phase clouds. 6) Systematic evaluations of mixed-phase cloud simulations by CAM5 were performed. Measurement results indicate that ice concentrations control stratiform mixed-phase cloud properties. The improvement of ice concentration parameterization in the CAM5 was done in close collaboration with Dr. Xiaohong Liu, PNNL (now at University of Wyoming).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Advancing the Understanding of Cloud Microphysical Processes and Aerosol Indirect Effects in High-Latitude Mixed-Phase Clouds by Linking ARM Measurements with Climate Model Simulations (Final Report)

The key objectives of this project were to advance our understanding of cloud microphysical characteristics and aerosol indirect effects on mixed-phase clouds in high latitudes. To improve the representation of ice and mixed-phase clouds in Earth System Models (ESMs), we propose an integrated observation and modeling study of cloud macro- and microphysical properties, including spatial heterogeneities, mass partitioning between ice crystals and supercooled liquid water, effects of ice nucleating particles (INPs), and efficiency of secondary ice production (SIP), etc. Specifically, we took four main approaches in this project: (1) examining macro- and microphysical properties of ice and mixed-phase clouds based on in-situ and ground-based observations from multiple field campaigns funded by the U.S. Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) program, including the Mixed-Phase Arctic Cloud Experiment (M-PACE), Indirect and Semi-Direct Aerosol Campaign (ISDAC), Ice Nucleating Particle Sources at Oliktok Point (INPOP), ARM West Antarctic Radiation Experiment (AWARE), Measurements of Aerosols, Radiation, and Clouds over the Southern Ocean (MARCUS), and Macquarie Island Cloud and Radiation Experiment (MICRE); (2) evaluating the DOE Energy Exascale Earth System Model (E3SM) simulations based on observations, particularly for ice and mixed-phase cloud microphysical properties; (3) examining the impacts of INPs on ice and mixed-phase clouds. Specifically, a series of comparisons were conducted using observations over the Arctic, Southern Ocean, and Antarctica, including comparisons between the lower and higher southern latitudes as well as comparisons between the northern and southern hemispheres. In addition, aerosol indirect effects from distinct sources of dust particles were examined; and (4) investigating the impacts of SIP. Ultimately, these results helped to improve cloud microphysics and aerosol-cloud interaction parameterizations in the E3SM model. Overall, the project provided improved understanding regarding various factors, including thermodynamic, dynamic, and aerosol conditions, on the micro- and macrophysical properties of ice and mixed-phase clouds in the high latitudes. Resulting analysis helped to provide an improved physical basis for refining the current cloud microphysics parameterizations related to ice and mixed-phase clouds in E3SM.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Classifying thermodynamic cloud phase using machine learning models

Vertically resolved thermodynamic cloud-phase classifications are essential for studies of atmospheric cloud and precipitation processes. The Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) Thermodynamic Cloud Phase (THERMOCLDPHASE) value-added product (VAP) uses a multi-sensor approach to classify the thermodynamic cloud phase by combining lidar backscatter and depolarization, radar reflectivity, Doppler velocity, spectral width, microwave-radiometer-derived liquid water path, and radiosonde temperature measurements. The measured pixels are classified as ice, snow, mixed phase, liquid (cloud water), drizzle, rain, and liq_driz (liquid+drizzle). We use this product as the ground truth to train three machine learning (ML) models to predict the thermodynamic cloud phase from multi-sensor remote sensing measurements taken at the ARM North Slope of Alaska (NSA) observatory: a random forest (RF), a multi-layer perceptron (MLP), and a convolutional neural network (CNN) with a U-Net architecture. Evaluations against the outputs of the THERMOCLDPHASE VAP with 1 year of data show that the CNN outperforms the other two models, achieving the highest test accuracy, F1 score, and mean intersection over union (IOU). Analysis of ML confidence scores shows that ice, rain, and snow have higher confidence scores, followed by liquid, while mixed, drizzle, and liq_driz have lower scores. Feature importance analysis reveals that the mean Doppler velocity and vertically resolved temperature are the most influential data streams for ML thermodynamic cloud-phase predictions. Lidar measurements exhibit lower feature importance due to rapid signal attenuation caused by the frequent presence of persistent low-level clouds at the NSA site. The ML models' generalization capacity is further evaluated by applying them at another Arctic ARM site in Norway using data taken during the ARM Cold-Air Outbreaks in the Marine Boundary Layer Experiment (COMBLE) field campaign. The models demonstrated similar performance to that observed at the NSA site. Finally, we evaluate the ML models' response to simulated instrument outages and signal degradation and show that a CNN U-Net model trained with input channel dropouts performs better when input fields are missing.

ARM Aerial Facility↗
DOE OSTI2025

Thermodynamic Cloud Phase Classifications Using Machine Learning at NSA and ANX

Vertically resolved thermodynamic cloud phase classifications are essential for studies of atmospheric cloud and precipitation processes. The Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) THERMOCLDPHASE Value-Added Product (VAP) uses a multi-sensor approach to classify thermodynamic cloud phase by combining lidar backscatter and depolarization, radar reflectivity, Doppler velocity, spectral width, microwave radiometer-derived liquid water path, and radiosonde temperature measurements. The measured voxels are classified as ice, snow, mixed-phase, liquid (cloud water), drizzle, rain, and liq_driz (liquid+drizzle). We use this product as the ground truth to train three machine learning (ML) models to predict the thermodynamic cloud phase from multi-sensor remote sensing measurements taken at the ARM North Slope of Alaska (NSA) observatory: a random forest (RF), a multilayer perceptron (MLP), and a convolutional neural network (CNN) with a U-Net architecture. Evaluations against the outputs of the THERMOCLDPHASE VAP with one year of data show that the CNN outperforms the other two models, achieving the highest test accuracy, F1-score, and mean Intersection over Union (IOU). Analysis of ML confidence scores shows ice, rain, and snow have higher confidence scores, followed by liquid, while mixed, drizzle, and liq_driz have lower scores. Feature importance analysis reveals that the mean Doppler velocity and vertically resolved temperature are the most influential datastreams for ML thermodynamic cloud phase predictions. The ML models’ generalization capacity is further evaluated by applying them at another Arctic ARM site in Norway using data taken during the ARM Cold-Air Outbreaks in the Marine Boundary Layer Experiment (COMBLE) field campaign. Finally, we evaluate the ML models’ response to simulated instrument outages and signal degradation.

54 ENVIRONMENTAL SCIENCES↗