skip to main content


Title: The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning
Abstract. Mixed-phase Southern Ocean clouds are challenging to simulate, and theirrepresentation in climate models is an important control on climatesensitivity. In particular, the amount of supercooled water and frozen massthat they contain in the present climate is a predictor of their planetaryfeedback in a warming climate. The recent Southern Ocean Clouds, Radiation, Aerosol Transport Experimental Study (SOCRATES) vastly increased theamount of in situ data available from mixed-phase Southern Ocean clouds usefulfor model evaluation. Bulk measurements distinguishing liquid and ice watercontent are not available from SOCRATES, so single-particle phaseclassifications from the Two-Dimensional Stereo (2D-S) probe are invaluablefor quantifying mixed-phase cloud properties. Motivated by the presence oflarge biases in existing phase discrimination algorithms, we develop a noveltechnique for single-particle phase classification of binary 2D-S images usinga random forest algorithm, which we refer to as the University of WashingtonIce–Liquid Discriminator (UWILD). UWILD uses 14 parameters computed frombinary image data, as well as particle inter-arrival time, to predict phase.We use liquid-only and ice-dominated time periods within the SOCRATES datasetas training and testing data. This novel approach to model training avoidsmajor pitfalls associated with using manually labeled data, including reducedmodel generalizability and high labor costs. We find that UWILD is wellcalibrated and has an overall accuracy of 95 % compared to72 % and 79 % for two existing phase classificationalgorithms that we compare it with. UWILD improves classifications of smallice crystals and large liquid drops in particular and has more flexibilitythan the other algorithms to identify both liquid-dominated and ice-dominatedregions within the SOCRATES dataset. UWILD misclassifies a small percentageof large liquid drops as ice. Such misclassified particles are typicallyassociated with model confidence below 75 % and can easily befiltered out of the dataset. UWILD phase classifications show that particleswith area-equivalent diameter (Deq)  < 0.17 mm are mostlyliquid at all temperatures sampled, down to −40 ∘C. Largerparticles (Deq>0.17 mm) are predominantly frozen at alltemperatures below 0 ∘C. Between 0 and 5 ∘C,there are roughly equal numbers of frozen and liquid mid-sized particles (0.170.33 mm) are mostly frozen. We also use UWILD's phaseclassifications to estimate sub-1 Hz phase heterogeneity, and we showexamples of meter-scale cloud phase heterogeneity in the SOCRATES dataset.  more » « less
Award ID(s):
1744965
NSF-PAR ID:
10318517
Author(s) / Creator(s):
; ; ; ; ; ;
Date Published:
Journal Name:
Atmospheric Measurement Techniques
Volume:
14
Issue:
11
ISSN:
1867-8548
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract

    Climate and numerical weather prediction models struggle to accurately predict radiative forcing over the Southern Ocean (SO), as the amount of clouds and their phases are poorly represented in such models due to a lack of observations upon which to base parameterizations. To address this, a novel particle identification (PID) scheme, based upon airborne radar and lidar data, was applied to data collected during the Southern Ocean Clouds, Aerosol, Radiation Transport Experimental Study (SOCRATES) to assess the vertical structure of SO boundary layer clouds. A comparison between the PID scheme and in situ phase data from SOCRATES showed relatively good agreement between the two data types. The convectivity of the clouds sampled during SOCRATES was determined using the novel Echo Classification from COnvectivity for Vertically pointing radars product. The PID and in situ data were then used synergistically to identify the following features of cloud vertical structure: (a) Supercooled liquid water was very common (Probability,P∼ 80%) at cloud top for convective and stratiform clouds; (b) Supercooled large drops with maximum dimensions >95 μm frequently appear within a hundred meters below cloud top, particularly within convective clouds (MaxP35%–45%), but also within stratiform clouds (MaxP20%–30%); (c) Ice production was associated with convective activity, withP∼ 20% at cloud top, increasing to 50%–70% 200 m below top, compared toP < 30% everywhere in stratiform clouds; (d) Convective clouds were found to be more vertically heterogeneous than stratiform clouds.

     
    more » « less
  2. Abstract

    Climate models struggle to accurately represent the highly reflective boundary layer clouds overlying the remote and stormy Southern Ocean. We use in situ aircraft observations from the Southern Ocean Clouds, Radiation and Aerosol Transport Experimental Study (SOCRATES) to evaluate Southern Ocean clouds in a cloud‐resolving large‐eddy simulation (LES) and two coarse resolution global atmospheric models, the CESM Community Atmosphere Model (CAM6) and the GFDL Atmosphere Model (AM4), run in a nudged hindcast framework. We develop six case studies from SOCRATES data which span the range of observed cloud and boundary layer properties. For each case, the LES is run once forced purely using reanalysis data (fifth generation European Centre for Medium‐Range Weather Forecasts atmospheric reanalysis, “ERA5 based”) and once strongly nudged to an aircraft profile(“Obs based”). The ERA5‐based LES can be compared with the global models, which are also nudged to reanalysis data and are better for simulating cumulus. The Obs‐based LES closely matches an observed cloud profile and is useful for microphysical comparisons and sensitivity tests and simulating multilayer stratiform clouds. We use two‐moment Morrison microphysics in the LES and find that it simulates too few frozen particles in clouds occurring within the Hallett‐Mossop temperature range. We tweak the Hallett‐Mossop parameterization so that it activates within boundary layer clouds, and we achieve better agreement between observed and simulated microphysics. The nudged global climate models (GCMs) simulate liquid‐dominated mixed‐phase clouds in the stratiform cases but excessively glaciate cumulus clouds. Both GCMs struggle to represent two‐layer clouds, and CAM6 has low droplet concentrations in all cases and underpredicts stratiform cloud‐driven turbulence.

     
    more » « less
  3. Abstract

    Supercooled liquid water (SLW) and mixed phase clouds containing SLW and ice over the Southern Ocean (SO) are poorly represented in global climate and numerical weather prediction models. Observed SLW exists at lower temperatures than threshold values used to characterize its detrainment from convection in model parameterizations, and processes controlling its formation and removal are poorly understood. High‐resolution observations are needed to better characterize SLW over the SO. This study characterizes the frequency and spatial distribution of different cloud phases (liquid, ice, and mixed) using in situ observations acquired during the Southern Ocean Clouds, Radiation, Aerosol Transport Experiment Study. Cloud particle phase is identified using multiple cloud probes. Results show occurrence frequencies of liquid phase samples up to 70% between −20°C and 0°C and of ice phase samples up to 10% between −5°C and 0°C. Cloud phase spatial heterogeneity is determined by relating the total number of 1 s samples from a given cloud to the number of segments whose neighboring samples are the same phase. Mixed phase conditions are the most spatially heterogeneous from −20°C to 0°C, whereas liquid phase conditions from −10°C to 0°C and ice phase conditions from −20°C to −10°C are the least spatially heterogeneous. Greater spatial heterogeneity is associated with broader distributions of vertical velocity. Decreasing droplet concentrations and increasing number‐weighted mean liquid diameters occur within mixed phase clouds as the liquid water fraction decreases, possibly suggesting preferential evaporation of smaller drops during the Wegener‐Bergeron‐Findeisen process.

     
    more » « less
  4. Abstract

    This paper describes a technique for identifying hydrometeor particle types using airborne HIAPER Cloud Radar (HCR) and High Spectral Resolution Lidar (HSRL) observations. HCR operates at a frequency of 94 GHz (3 mm wavelength), while HSRL is an eye‐safe lidar system operating at a wavelength of 532 nm. Both instruments are deployed on the NSF‐NCAR HIAPER aircraft. HCR is designed to fly in an underwing pod and HSRL is situated in the cabin. The HCR and HSRL data used in this study were collected during the Southern Ocean Clouds, Radiation, Aerosol Transport Experimental Study (SOCRATES). Comprehensive observations of the vertical distributions of liquid and mixed‐phase clouds were obtained using in situ probes on the aircraft and remote sensing instruments. Hydrometeor particle types were retrieved from HCR, HSRL, and temperature fields with a newly developed fuzzy logic particle identification (PID) algorithm. The PID results were validated with in situ measurements collected onboard the HIAPER aircraft by a 2D‐Stereo (2D‐S) cloud probe. Particle phases derived from the PID results compare well with those obtained from the 2D‐S observations and agree in over 70% of cases. Size distributions are also consistent between the two methods of observation. Knowledge of the particle type distribution gained from the PID results can be used to constrain microphysical parameterizations and improve the representation of cloud radiative effects in weather and climate models.

     
    more » « less
  5. Abstract

    Recent studies have suggested a correct representation of cloud phase in the Southern Ocean region is important in climate models for an accurate representation of the energy balance. Satellite retrievals indicate many of the clouds are predominantly liquid, despite their low temperatures. However, clouds containing high numbers of ice crystals have sometimes been observed in this region and implicated the secondary ice production process called rime splintering. This study re‐examines rime splintering in Southern Ocean cumuli using both a new data set and high‐resolution numerical modeling. Measurements acquired during the Southern Ocean Clouds Radiation Aerosol Transport Experimental Study (SOCRATES) provide an evaluation of the amount of ice in shallow cumuli sampled over two days in this region. The measurements sometimes exhibit seven orders of magnitude or more ice particles compared to amounts expected from measurements of ice‐nucleating particles (INP) on the same days. Cumuli containing multiple updrafts had the greatest tendency to contain high ice concentrations and meet the expected conditions for rime splintering. Idealized numerical modeling, constrained by the observations, suggests that the multiple updrafts produce more frozen raindrops/graupel, and allow them to travel through the rime‐splintering zone over an extended period of time, increasing the number of ice particles by many orders of magnitude. The extremely low number of INP in the Southern Ocean thus appears to require special conditions like multiple updrafts to help glaciate the cumuli in this region, potentially explaining the predominance of supercooled cumuli observed there.

     
    more » « less