Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “cloud type classification”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Synoptic Weather Regime Classifications for September, October and November, from 2000 to 2025

The synoptic weather regime classification has become a highly demanded product for the ARM site in recent years. This type of regime classification has shown applications in various studies and topics, including aerosol-cloud interactions, land-atmosphere interactions, and cloud radiative effects. The VAP employs an unsupervised machine learning method, Self-organizing map (SOM), to classify weather regimes for each day of the AMF campaigns and fixed sites, using ERA5 data. This idea is mainly based on our published study for TRACER in Wang et al. (2022, JGR-A).

node_som_ml↗
DOE OSTI2019

Long-Term Retrievals of Cloud Type and Fair-Weather Shallow Cumulus Events at the ARM SGP Site

Abstract A long-term climatology of classified cloud types has been generated for 13 years (1997–2009) over the Atmospheric Radiation Measurement (ARM) Southern Great Plains (SGP) site for seven cloud categories: low clouds, congestus, deep convection, altocumulus, altostratus, cirrostratus/anvil, and cirrus. The classification was based on the cloud macrophysical quantities of cloud top, cloud base, and physical thickness of cloud layers, as measured by active sensors such as the millimeter-wavelength cloud radar (MMCR) and micropulse lidar (MPL). Climate variability of cloud characteristics has been examined using the 13-yr cloud-type retrieval. Low clouds and cirrus showed distinct diurnal and seasonal cycles. Total cloud occurrence followed the variation of low clouds, with a diurnal peak in early afternoon and a seasonal maximum in late winter. Additionally, further work has been done to identify fair-weather shallow cumulus (FWSC) events for 9 years (2000–08). Periods containing FWSC, a subcategory of clouds classified as low clouds, were produced using cloud fraction information from a total-sky imager and ceilometer. The identified FWSC periods in our study show good agreement with manually identified FWSC, missing only 6 cases out of 70 possible events during the spring to summer seasons (May–August).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Photometric Properties of Jupiter Trojans Detected by the Dark Energy Survey

The Jupiter Trojans are a large group of asteroids that are coorbiting with Jupiter near its L4 and L5 Lagrange points. The study of Jupiter Trojans is crucial for testing different models of planet formation that are directly related to our understanding of solar system evolution. In this work, we select known Jupiter Trojans listed by the Minor Planet Center from the full six years data set (Y6) of the Dark Energy Survey (DES) to analyze their photometric properties. The DES data allow us to study Jupiter Trojans with a fainter magnitude limit than previous studies in a homogeneous survey with griz band measurements. We extract a final catalog of 573 unique Jupiter Trojans. Our sample include 547 asteroids belonging to L5. This is one of the largest analyzed samples for this group. By comparing with the data reported by other surveys we found that the color distribution of L5 Trojans is similar to that of L4 Trojans. We find that L5 Trojans’ g − i and g − r colors become less red with fainter absolute magnitudes, a trend also seen in L4 Trojans. Both the L4 and L5 clouds consistently show such a color–size correlation over an absolute magnitude range 11 < H < 18. We also use DES colors to perform taxonomic classifications. C- and P-type asteroids outnumber D-type asteroids in the L5 Trojans DES sample, which have diameters in the 5–20 km range. This is consistent with the color–size correlation.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2019

SPRUCE Microtopography of Experimental Plots Derived from Terrestrial Laser Scans Beginning in 2016

This dataset consists of maps and summaries of very high resolution (1 cm2) spatial data derived from terrestrial laser scans (TLS) characterizing the microtopography (hummocks and hollows) in 12 SPRUCE experimental plots from 2016-2018. The SPRUCE experiment is located in the S1-Bog of the Marcell Experimental Forest (MEF) near Grand Rapids, Minnesota. Additional data will be added through Summer 2021.The TLS sampling plan follows the SPRUCE experimental design, with scanning of the ten enclosed treatment plots with five temperature treatments (0, 2.25, 4.5, 6.75, and 9°C) with and without elevated CO2 (~900 ppm) and the two monitored ambient plots without enclosures. Enclosures are octagonal in shape and are 12.8 m in diameter (114.8 m2 in area), 8 m tall, with a belowground corral that penetrates through the peat to glacial material below, which averages 2.5 m deep across the site. In the spring of each year, four TLS scans were taken per plot/enclosure and registered together to produce a single point cloud for each plot (Figure 2A). Digital surface models (DSMs) were generated from the point clouds and clipped to the inside boundary of the boardwalk in each plot resulting in an approximately 9 x 9 m study area. . All other characterization data types were derived from plot DSMs (elevation), including: slope (Figure 3), concavity (Figure 4), index used for microform classification (Hollow Index) (Figure 6), and microforms (Figure 2C; hummocks and hollows) classified using three methodologies.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Classification of Cloud Particle Imagery from Aircraft Platforms Using Convolutional Neural Networks

Abstract A vast amount of ice crystal imagery exists from a variety of field campaign initiatives that can be utilized for cloud microphysical research. Here, nine convolutional neural networks are used to classify particles into nine regimes on over 10 million images from the Cloud Particle Imager probe, including liquid and frozen states and particles with evidence of riming. A transfer learning approach proves that the Visual Geometry Group (VGG-16) network best classifies imagery with respect to multiple performance metrics. Classification accuracies on a validation dataset reach 97% and surpass traditional automated classification. Furthermore, after initial model training and preprocessing, 10 000 images can be classified in approximately 35 s using 20 central processing unit cores and two graphics processing units, which reaches real-time classification capabilities. Statistical analysis of the classified images indicates that a large portion (57%) of the dataset is unusable, meaning the images are too blurry or represent indistinguishable small fragments. In addition, 19% of the dataset is classified as liquid drops. After removal of fragments, blurry images, and cloud drops, 38% of the remaining ice particles are largely intersecting the image border (≥10% cutoff) and therefore are considered unusable because of the inability to properly classify and dimensionalize. After this filtering, an unprecedented database of 1 560 364 images across all campaigns is available for parameter extraction and bulk statistics on specific particle types in a wide variety of storm systems, which can act to improve the current state of microphysical parameterizations.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Graph Neural Networks for low-energy event classification & reconstruction in IceCube

IceCube, a cubic-kilometer array of optical sensors built to detect atmospheric and astrophysical neutrinos between 1 GeV and 1 PeV, is deployed 1.45 km to 2.45 km below the surface of the ice sheet at the South Pole. The classification and reconstruction of events from the in-ice detectors play a central role in the analysis of data from IceCube. Reconstructing and classifying events is a challenge due to the irregular detector geometry, inhomogeneous scattering and absorption of light in the ice and, below 100 GeV, the relatively low number of signal photons produced per event. To address this challenge, it is possible to represent IceCube events as point cloud graphs and use a Graph Neural Network (GNN) as the classification and reconstruction method. The GNN is capable of distinguishing neutrino events from cosmic-ray backgrounds, classifying different neutrino event types, and reconstructing the deposited energy, direction and interaction vertex. Based on simulation, we provide a comparison in the 1 GeV–100 GeV energy range to the current state-of-the-art maximum likelihood techniques used in current IceCube analyses, including the effects of known systematic uncertainties. For neutrino event classification, the GNN increases the signal efficiency by 18% at a fixed background rate, compared to current IceCube methods. Alternatively, the GNN offers a reduction of the background (i.e. false positive) rate by over a factor 8 (to below half a percent) at a fixed signal efficiency. For the reconstruction of energy, direction, and interaction vertex, the resolution improves by an average of 13%–20% compared to current maximum likelihood techniques in the energy range of 1 GeV–30 GeV. The GNN, when run on a GPU, is capable of processing IceCube events at a rate nearly double of the median IceCube trigger rate of 2.7 kHz, which opens the possibility of using low energy neutrinos in online searches for transient events.

47 OTHER INSTRUMENTATION↗
DOE OSTI2018

Classifying Precipitation Types in Tropical Cyclones Using the NRL 37 GHz Color Product

The tropical cyclone (TC) webpage developed by the U.S. Naval Research Lab (NRL) has been widely used by the community for real-time TC monitoring and forecasting. Using 14 years of Tropical Rainfall Measuring Mission Microwave Imager and Precipitation Radar data in TCs, precipitation/convection types are quantitatively classified based on vertical radar profiles corresponding to seven different color regions in the NRL's 37 GHz color composite product. The bright cyan and all pink (including pure pink, bright cyan/pink, and weak cyan/pink, and green/pink) regions are identified as nearly 100% precipitative regions. The bright cyan and pure pink regions are shallow and deep convective precipitation, respectively, while the green/pink, weak cyan/pink, and bright cyan/pink regions are mainly stratiform rain. Eighty-four percent of pixels in the weak cyan region are precipitating, including either shallow convective, stratiform, or anvil precipitation. Seventy-six percent of pixels in the green region are nonprecipitating pixels, and 24% of them are very lightly precipitating pixels, which could be contaminated from clouds with high liquid water content. The ability to perform TC precipitation-type analysis of this detail (convective versus stratiform and deep versus shallow) was previously only possible using radar-based classification algorithms. In conclusion, using the 37 GHz precipitation-type analysis, a significant increase of coverage of stratiform precipitation and shallow convection in the inner core is found between 3 and 21 hr before the onset of TC rapid intensification.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2018

The Green Ocean: Precipitation Insights from the GoAmazon2014/5 experiment

This study summarizes the precipitation properties collected during the GoAmazon2014/5 campaign near Manaus in central Amazonia, Brazil. Precipitation breakdowns, summary radar rainfall relationships and self-consistency concepts from a coupled disdrometer and radar wind profiler measurements are presented. The properties of Amazon cumulus and associated stratiform precipitation are discussed, including segregations according to seasonal (Wet/Dry regime) variability, 5 cloud echo-top height and possible aerosol influences on the apparent oceanic characteristics of the precipitation drop size distributions. Overall, we observe that the Amazon precipitation straddles behaviors found during previous U.S. Department of Energy Atmospheric Radiation Measurements program (ARM) tropical deployments, with distributions favoring higher concentrations of smaller drops than ARM continental examples. Oceanic type precipitation characteristics are predominantly observed during the Amazon Wet seasons. An exploration of the controls on Wet season precipitation properties reveals that 10 wind direction, as compared with other standard radiosonde thermodynamic parameters or aerosol count/regime classifications performed at the ARM site, provides a good indicator for those Wet season Amazon events having an oceanic character for their precipitation drop size distributions.

aerosol influences↗
DOE OSTI2025

The NASA ACTIVATE Mission

The NASA Aerosol Cloud Meteorology Interactions over the Western Atlantic Experiment (ACTIVATE) conducted 162 joint flights with two aircraft over the northwest Atlantic to study aerosol–cloud interactions (ACIs), which represent the largest uncertainty in estimating total anthropogenic radiative forcing. The combination of a high-flying King Air and low-flying HU-25 Falcon, equipped with remote sensing and in situ instruments, characterized trace gases, aerosol particles, clouds, and meteorological variables with data collected nearly simultaneously below, within, and above marine boundary layer (MBL) clouds. Flights spanning warm and cold seasons across 3 years (2020–22) provided a broad range of conditions associated with aerosol particles, cloud properties (including particle size and phase), and meteorology, ideally suited for robust ACI calculations and assessing how well models simulate a wide range of MBL clouds from stratiform to cumulus. ACTIVATE data suggest that drivers of cloud droplet number concentration N d , including aerosol particles and MBL dynamics, vary between winter and summer months with a stronger potential to convert aerosol particles into cloud droplets in winter. Models of varying complexity not only highlight some skills in simulating winter and summer cloud types but also identify challenges that still need to be addressed such as treatment of turbulence, wet scavenging, and mesoscale organization. Remote sensing advances range from new retrieval methods for N d , cloud phase classification, vertically resolved aerosol and cloud condensation nuclei number concentration, and ocean surface wind speed. This work describes these scientific and technological advances along with efforts in outreach and open data science.

aerosol indirect effect↗
DOE OSTI2020

Modeling Spatial Heterogeneity in Surface Turbulent Heat Flux in the U.S. Southern Great Plains

Advances in numerical modeling of cloud dynamics are driving a need for improved land model prediction at convective storm scales. Here, satellite and ground-based vegetation remote sensing data were combined with land model experiments to more accurately characterize land surface spatial heterogeneity in the Community Land Model (CLM4.0). The new subgrid classification of plant functional types (PFT) and leaf area index (LAI) enables consistent comparison between models and ground-based flux measurements in the U.S. southern Great Plains. Errors in vegetation data sets (inferred from comparison between 250 m satellite and ground-based LAI), while large, had less impact on the simulated characteristics of spatial heterogeneity compared to errors in model representation of surface energy partitioning (between latent and sensible heat flux) and its relationship to LAI. Predicted spatial heterogeneity in surface energy partitioning was enhanced after replacing soil and stomatal resistance parameters with a new set that better predicts the observed relationship to LAI. These modifications increase the number of smaller (mesoscale) dry land patches having higher sensible heat flux. The parameter experiments suggest that vegetation state and processes (transpiration) act to broaden the size spectrum of surface heat flux heterogeneity, which can influence clouds and convective initiation. Moreover, improvements in vegetation input data and model parameters had partially compensating effects on surface flux heterogeneity, indicating the importance of evaluating input data and parameterizations together to improve prediction at higher spatial resolutions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

MindSynchro

This report presents the developments and results of MindSynchro project as part of DOE OE FOA 1861. DOE and Pacific Northwest National Laboratory (PNNL) have made available to FOA awardees datasets containing years of real historical data recorded from various phasor measurement units (PMUs) which are installed in three large US interconnections: Texas (IC A), Western (IC B), and Eastern (IC C). The main goal of the project, which was successfully achieved, was to develop methods for detection and identification of events which are relevant for power grid operation. Tasks performed for achieving the project goals included data exploration and pre-processing, the development and application of physics-based features, data analysis and labeling based on unsupervised learning approaches, training and testing of DSSL models for classification of events which are relevant for power grid operation, and deployment of solutions to cloud environments. The methods developed in the project can potentially provide relevant benefits to power grid asset owners/operators in general in terms of situational awareness. Two main types of outcomes can be provided by these tools: Identification of specific relevant power grid event types: Semi-supervised ML methods developed in the project can adequately employ not only the relatively scarce labeled data but also the large amount of available unlabeled data to train models for detection of specific event types. Such methods enable the application of trained models for the detection of events in a population of PMUs much larger than that associated to the labeled events. Support in data labeling / label validation: Labels are critical for training of models for identification of specific types of events. However, labeling large amounts of data is a manual and tedious process. This means that such process is error prone and is not scalable. Methods developed in the project, based on ensembles of clustering models, have been successfully employed for turning manual labeling into a scalable process. Accurate identification of specific relevant events can provide the operators with immediate situational awareness that could otherwise require hours or days of analysis from domain experts. We envision that such methods could be initially employed in support of post-mortem analysis of events and, as confidence is gained, they could be employed for online/real-time support, providing, among other benefits, insights for avoiding major events which could happen due to a combination of smaller ones. On the longer term, related methods could potentially be employed to improve protection and control.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2021

Concentric Spherical GNN for 3D Representation Learning

Learning 3D representations that generalize well to arbitrarily oriented inputs is a challenge of practical importance in applications varying from computer vision to physics and chemistry. We propose a novel multi-resolution convolutional architecture for learning over concentric spherical feature maps, of which the single sphere representation is a special case. Our hierarchical architecture is based on alternatively learning to incorporate both intra-sphere and inter-sphere information. We show the applicability of our method for two different types of 3D inputs, mesh objects, which can be regularly sampled, and point clouds, which are irregularly distributed. We also propose an efficient mapping of point clouds to concentric spherical images, thereby bridging spherical convolutions on grids with general point clouds. We demonstrate the effectiveness of our approach in improving state-of-the-art performance on 3D classification tasks with rotated data.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Long-term measurements of ice nucleating particles at Atmospheric Radiation Measurement (ARM) sites worldwide

Ice nucleating particles (INPs) play a critical role in cloud microphysics and precipitation formation, yet long-term, spatially extensive observational datasets remain limited. Here, we present one of the most comprehensive publicly available datasets of immersion-mode INP concentrations using a single analytical method, generated through the U.S. Department of Energy's (DOE) Atmospheric Radiation Measurement (ARM) user facility. INP filter samples have been collected across a broad range of environments – including agricultural plains, Arctic coastlines, high-elevation mountain sites, marine regions, and urban areas – via fixed observatories, mobile facility deployments, and vertically-resolved tethered balloon system operations. We describe the standardized processing and quality assurance pipeline, from filter collection and processing using the Ice Nucleation Spectrometer to final data products archived on the ARM Data Discovery portal. The dataset includes both total INP concentrations and selectively treated samples, allowing for classification of biological, organic, and inorganic INP types. It features a continuous 5-year record of INP measurements from a central U.S. site, with data collection still ongoing. Seasonal and site-specific differences in INP concentrations are illustrated through intercomparisons at −10 and −20 °C, revealing distinct regional sources and atmospheric drivers. We also outline mechanisms for researchers to access existing data, request additional sample analyses, and propose future field campaigns involving ARM INP measurements. This dataset supports a wide range of scientific applications, from observational and mechanistic studies to model development, and provides critical constraints on aerosol-cloud interactions across diverse atmospheric regimes (Creamean et al., 2024, 2020b; https://doi.org/10.5439/1770816).

Creamean, Jessie M. [Colorado State Univ., Fort Co↗
DOE OSTI2016

POPULATION PROPERTIES OF BROWN DWARF ANALOGS TO EXOPLANETS

We present a kinematic analysis of 152 low surface gravity M7-L8 dwarfs by adding 18 new parallaxes (including 10 for comparative field objects), 38 new radial velocities, and 19 new proper motions. We also add low- or moderate-resolution near-infrared spectra for 43 sources confirming their low surface gravity features. Among the full sample, we find 39 objects to be high-likelihood or new bona fide members of nearby moving groups, 92 objects to be ambiguous members and 21 objects that are non-members. Using this age-calibrated sample, we investigate trends in gravity classification, photometric color, absolute magnitude, color–magnitude, luminosity, and effective temperature. We find that gravity classification and photometric color clearly separate 5–130 Myr sources from >3 Gyr field objects, but they do not correlate one to one with the narrower 5–130 Myr age range. Sources with the same spectral subtype in the same group have systematically redder colors, but they are distributed between 1 and 4 σ from the field sequences and the most extreme outlier switches between intermediate- and low-gravity sources either confirmed in a group or not. The absolute magnitudes of low-gravity sources from the J band through W 3 show a flux redistribution when compared to equivalently typed field brown dwarfs that is correlated with spectral subtype. Low-gravity, late-type L dwarfs are fainter at J than the field sequence but brighter by W 3. Low-gravity M dwarfs are >1 mag brighter than field dwarfs in all bands from J through W 3. Clouds, which are a far more dominant opacity source for L dwarfs, are the likely cause. On color–magnitude diagrams, the latest-type, low-gravity L dwarfs drive the elbow of the L/T transition up to 1 mag redder and 1 mag fainter than field dwarfs at M{sub J} but are consistent with or brighter than the elbow at M{sub W1} and M{sub W2}. We conclude that low-gravity dwarfs carry an extreme version of the cloud conditions of field objects to lower temperatures, which logically extends into the lowest-mass, directly imaged exoplanets. Furthermore, there is an indication on color-magnitude diagrams (CMDs; such as M{sub J} versus (J – W 2)) of increasingly redder sequences separated by gravity classification, although it is not consistent across all CMD combinations. Examining bolometric luminosities for planets and low-gravity objects, we confirm that (in general) young M dwarfs are overluminous while young L dwarfs are normal compared to the field. Using model extracted radii, this translates into normal to slightly warmer M dwarf temperatures compared to the field sequence and lower temperatures for L dwarfs with no obvious correlation with the assigned moving group.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

AICCA: AI-Driven Cloud Classification Atlas

Clouds play an important role in the Earth’s energy budget, and their behavior is one of the largest uncertainties in future climate projections. Satellite observations should help in understanding cloud responses, but decades and petabytes of multispectral cloud imagery have to date received only limited use. This study describes a new analysis approach that reduces the dimensionality of satellite cloud observations by grouping them via a novel automated, unsupervised cloud classification technique based on a convolutional autoencoder, an artificial intelligence (AI) method good at identifying patterns in spatial data. Our technique combines a rotation-invariant autoencoder and hierarchical agglomerative clustering to generate cloud clusters that capture meaningful distinctions among cloud textures, using only raw multispectral imagery as input. Cloud classes are therefore defined based on spectral properties and spatial textures without reliance on location, time/season, derived physical properties, or pre-designated class definitions. We use this approach to generate a unique new cloud dataset, the AI-driven cloud classification atlas (AICCA), which clusters 22 years of ocean images from the Moderate Resolution Imaging Spectroradiometer (MODIS) on NASA’s Aqua and Terra instruments—198 million patches, each roughly 100 km × 100 km (128 × 128 pixels)—into 42 AI-generated cloud classes, a number determined via a newly-developed stability protocol that we use to maximize richness of information while ensuring stable groupings of patches. AICCA thereby translates 801 TB of satellite images into 54.2 GB of class labels and cloud top and optical properties, a reduction by a factor of 15,000. The 42 AICCA classes produce meaningful spatio-temporal and physical distinctions and capture a greater variety of cloud types than do the nine International Satellite Cloud Climatology Project (ISCCP) categories—for example, multiple textures in the stratocumulus decks along the West coasts of North and South America. We conclude that our methodology has explanatory power, capturing regionally unique cloud classes and providing rich but tractable information for global analysis. AICCA delivers the information from multi-spectral images in a compact form, enables data-driven diagnosis of patterns of cloud organization, provides insight into cloud evolution on timescales of hours to decades, and helps democratize climate research by facilitating access to core data.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

THE RUNAWAYS AND ISOLATED O-TYPE STAR SPECTROSCOPIC SURVEY OF THE SMC (RIOTS4)

We present the Runaways and Isolated O-Type Star Spectroscopic Survey of the SMC (RIOTS4), a spatially complete survey of uniformly selected field OB stars that covers the entire star-forming body of the Small Magellanic Cloud (SMC). Using the IMACS (Inamori-Magellan Areal Camera and Spectrograph) multislit spectrograph and MIKE (Magellan Inamori Kyocera Echelle) echelle spectrograph on the Magellan telescopes, we obtained spectra of 374 early-type field stars that are at least 28 pc from any other OB candidates. We also obtained spectra of an additional 23 field stars in the SMC bar identified from slightly different photometric criteria. Here, we present the observational catalog of stars in the RIOTS4 survey, including spectral classifications and radial velocities. For three multi-slit fields covering 8% of our sample, we carried out monitoring observations over 9–16 epochs to study binarity, finding a spectroscopic, massive binary frequency of at least ∼60% in this subsample. Classical Oe/Be stars represent a large fraction of RIOTS4 (42%), occurring at much higher frequency than in the Galaxy, consistent with expectation at low metallicity. RIOTS4 confirmed a steep upper initial mass function in the field, apparently caused by the inability of the most massive stars to form in the smallest clusters. Our survey also yields evidence for in situ field OB star formation, and properties of field emission-line star populations, including sgB[e] stars and classical Oe/Be stars. We also discuss the radial velocity distribution and its relation to SMC kinematics and runaway stars. RIOTS4 presents a first quantitative characterization of field OB stars in an external galaxy, including the contributions of sparse, but normal, star formation; runaway stars; and candidate isolated star formation.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

Leveraging machine learning to enhance aerosol classification using Single-Particle Mass Spectrometry

Advancing automated classification of atmospheric aerosols from Single-Particle Mass Spectrometry (SPMS) data remains challenging due to overlapping ion signatures, compositional diversity, and limited labeled data. This study evaluates supervised and semi-supervised learning frameworks to enhance aerosol identification by jointly leveraging labeled and unlabeled spectra. Four models were compared: a supervised Support Vector Machine (SVM), a self-training SVM, a stacked autoencoder classifier, and a stacked autoencoder trained using a temporal-ensembling Mean Teacher approach. All models achieved high and stable accuracies (90.0 %–91.1 %), surpassing previous results on the same dataset (87 %) and matching the performance of state-of-the-art deep learning methods. Despite small global metric differences (≤ 1 %), semi-supervised variants yielded up to 5 %–10 % improvements for compositionally rare particle types – such as soot (0.77 % of spectra, F1-score: 0.93–0.97) and hazelnut pollen (0.98 % of spectra, F1-score: 0.97–1.00) – equating to roughly ∼ 187 additional correctly classified spectra. These gains are scientifically significant, as such rare particles exert disproportionate influence on radiative absorption and ice nucleation processes; their improved detection reduces modeled uncertainties in aerosol absorption optical depth and mixed-phase cloud ice nucleation rates. The models' residual misclassifications (≈ 9 %) largely arise from true spectral overlap among chemically adjacent species (e.g., Na- vs. K-feldspar, coated vs. uncoated feldspars), reflecting physical compositional continuity rather than algorithmic error. Collectively, these findings demonstrate that leveraging unlabeled data to learn robust spectral representations and refine classification enhances both fidelity and interpretability, bridging data-driven analysis with aerosol–climate process understanding.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

HAIMOS Ensemble Forecasts for Intra-day and Day- Ahead GHI, DNI and Ramps

The objective of this research is to develop a hybrid physics-based/data-driven forecast model to improve direct normal and global horizontal irradiance (DNI and GHI) prediction for horizons ranging from 1 to 72 hours. Project objectives also address key gaps in state-of-the-art solar forecasting: accurate probabilistic solar forecasts and the forecasting of large irradiance ramps (ramp onset and magnitude). The proposed model ensembles Numerical Weather Prediction (NWP) forecasts, determinist physics-based algorithms, and new-generation cloud cover products (high-resolution rapid refresh satellite images and Large Eddy Simulations). The result is the Hybrid Adaptive Input Model Objective Selection (HAIMOS) ensemble model. HAIMOS blends state of the art machine learning methodologies with physics-based models for cloud cover and cloud optical depth forecasts. The technical activities followed a two-pronged strategy. First, the preprocessing of data, the selection of inputs to the nonlinear approximators, the type of approximator and objective functions, and post-processing ensembling techniques included in HAIMOS were all optimized adaptively to find the best model for a specific goal (reduce DNI/GHI forecast error, improve the prediction of ramp onset, etc.). Second, a large effort was put in improving cloud identification and the forecast of cloud cover and cloud optical depth. To this end, new-generation cloud parametrization products were developed in this work. These include improved algorithms to assist in cloud identification, cloud classification and cloud parametrization from satellite images – three key factors in the accuracy of 1 to 6-hours irradiance forecasts and prediction of ramp onset. Furthermore, we also included cloud information extracted from high resolution rapid refresh satellite images (GOES-16) and Large Eddy Simulations (LES). LES was used to model the atmosphere in detail over locations of interest and produce cloud optical depth forecasts. Once these data streams were validated, they were used as input data to the HAIMOS forecast. The model was developed using data from several climatologically distinct locations with potential for high solar penetration. In the last year of the project, we conducted a validation campaign according to the guidelines stipulated by the Topic Area 1 project as described in the FOA. This effort brings, for the first time, proven machine-learning methodologies for generating state-of-the-art solar forecasts interweaved with detailed physics-based models for cloud detection, and cloud optical depth forecasts. HAIMOS will generate accurate irradiance probabilistic forecast to assist in reducing solar generation prediction error. Globally optimized solar forecast models are more likely to impact solar energy stakeholders. The goal of this project was to increase the state-of-the-art forecast skill from their present values of 10 to 35%. At the end of the project, we achieved between 30% and 50% forecast skill across a wide range of horizons for both GHI and DNI.

14 SOLAR ENERGY↗