NSF NCAR Mesoscale & Microscale Meteorology Laboratory

Abstract WRF-Solar is a specific configuration and augmentation of the Weather Research and Forecasting (WRF) Model designed for solar energy applications. Recent upgrades to the WRF Model contribute to making the model appropriate for solar power forecasting and comprise 1) developments to diagnose internally relevant atmospheric parameters required by the solar industry, 2) improved representation of aerosol–radiation feedback, 3) incorporation of cloud–aerosol interactions, and 4) improved cloud–radiation feedback. The WRF-Solar developments are presented together with a comprehensive characterization of the model performance for forecasting during clear skies. Performance is evaluated with numerical experiments using a range of different external and internal treatment of the atmospheric aerosols, including both a model-derived climatology of aerosol optical depth and temporally evolving aerosol optical properties from reanalysis products. The necessity of incorporating the influence of atmospheric aerosols to obtain accurate estimations of the surface shortwave irradiance components in clear-sky conditions is evident. Improvements of up to 58%, 76%, and 83% are found in global horizontal irradiance, direct normal irradiance, and diffuse irradiance, respectively, compared to a standard version of the WRF Model. Results demonstrate that the WRF-Solar model is an improved numerical tool for research and applications in support of solar energy.

Midlatitude cyclones with characteristics similar to tropical cyclones (also known as Tropical‐Like Cyclones, TLCs, or medicanes) are sometimes observed in the Mediterranean region. The Wind Induced Surface Heat Exchange (WISHE) mechanism has been considered responsible for their development, in analogy with tropical‐cyclone theory. However, some recent papers have proposed a different explanation, suggesting that the deep warm core in the TLC is mainly an effect of the seclusion of warm air in the cyclone core. To investigate the latter hypothesis, two case‐studies of Mediterranean TLCs are analysed here by means of high‐resolution numerical experiments. The evolution of the near‐surface equivalent potential temperature is followed along back‐trajectories around the cyclone centre, showing for both cases a strong heating when the parcel moves from the outer part of the cyclone to its inner, warmer core. Sensitivity experiments clarify the mechanism of cyclone intensification and the way the warm‐core structure is generated, showing that sea‐surface fluxes and/or condensation latent heating are fundamental to explain the intensification of the cyclones. However, the importance of air–sea interaction processes is case dependent. For the first cyclone, the intense sea‐surface fluxes, associated with tramontane and cierzo winds over the western Mediterranean Sea, transfer a large amount of energy from the ocean to the atmosphere in the area where the cyclone developed, so that the vortex is able to sustain itself in a barotropic environment and reach a tropical‐like structure at a later stage in its lifetime. For the second cyclone, the cyclone never develops a fully tropical‐like structure, evolving in the baroclinic environment associated with the potential vorticity streamer in which the cyclone formed. Based on the distinction emerging in this and other articles, a classification of medicanes in three different categories is proposed.

Abstract Nocturnal convection initiation (NCI) is more difficult to anticipate and forecast than daytime convection initiation (CI). A major component of the Plains Elevated Convection at Night (PECAN) field campaign in the U.S. Great Plains was to intensively sample NCI and its near environment. In this article, we summarize NCI types observed during PECAN: 1 June–16 July 2015. These NCI types, classified using PECAN radar composites, are associated with 1) frontal overrunning, 2) the low-level jet (LLJ), 3) a preexisting mesoscale convective system (MCS), 4) a bore or density current, and 5) a nocturnal atmosphere lacking a clearly observed forcing mechanism (pristine). An example and description of each of these different types of PECAN NCI events are presented. The University of Oklahoma real-time 4-km Weather Research and Forecasting (WRF) Model ensemble forecast runs illustrate that the above categories having larger-scale organization (e.g., NCI associated with frontal overrunning and NCI near a preexisting MCS) were better forecasted than pristine. Based on current knowledge and data from PECAN, conceptual models summarizing key environmental features are presented and physical processes underlying the development of each of these different types of NCI events are discussed.

An analysis is presented here of intense convection affecting the Friuli Venezia Giulia region (FVG, northeastern Italy) during the Intensive Observation Period 2b (IOP2b) in the first Special Observation Period (SOP1) of HyMeX (HYdrological cycle in Mediterranean EXperiment). The present study focuses on the first of three severe‐convection episodes that affected FVG on the morning of 12 September 2012. In the first episode, a supercell, which produced hail and severe damage to trees and buildings, was observed on the plain of FVG. The available observations are analysed together with a high‐resolution mesoscale model, in order to identify the relevant mechanisms for the formation and development of the cell. Six different simulations were performed starting at three different initial times, using respectively two different analysis/forecasts as initial/boundary conditions. A large spread in forecast precipitation is found among the six simulations. Only a few of the simulations were able to reproduce intense rainfall on the plain of FVG during the morning, although with significant differences in the rainfall distribution among them. One of the six simulations well reproduces the observed elongated distribution of the intense rainfall maximum; the characteristics of the cell responsible for this distribution are consistent with those expected for a supercell and its simulated evolution near the Adriatic coast agrees well with the other observations. Some instability parameters over the FVG plain and offshore (over the northern Adriatic Sea) are analysed every 5 min, showing that during this event the potential instability varies significantly over small space and time intervals and among the simulations. The best simulations have the best match to the observed potential instability calculated using the mean characteristics of the lowest 500 m layer.

Abstract. The Model for Prediction Across Scales (MPAS) is a novel set of Earth system simulation components and consists of an atmospheric model, an ocean model and a land-ice model. Its distinct features are the use of unstructured Voronoi meshes and C-grid discretisation to address shortcomings of global models on regular grids and the use of limited area models nested in a forcing data set, with respect to parallel scalability, numerical accuracy and physical consistency. This concept allows one to include the feedback of regional land use information on weather and climate at local and global scales in a consistent way, which is impossible to achieve with traditional limited area modelling approaches. Here, we present an in-depth evaluation of MPAS with regards to technical aspects of performing model runs and scalability for three medium-size meshes on four different high-performance computing (HPC) sites with different architectures and compilers. We uncover model limitations and identify new aspects for the model optimisation that are introduced by the use of unstructured Voronoi meshes. We further demonstrate the model performance of MPAS in terms of its capability to reproduce the dynamics of the West African monsoon (WAM) and its associated precipitation in a pilot study. Constrained by available computational resources, we compare 11-month runs for two meshes with observations and a reference simulation from the Weather Research and Forecasting (WRF) model. We show that MPAS can reproduce the atmospheric dynamics on global and local scales in this experiment, but identify a precipitation excess for the West African region. Finally, we conduct extreme scaling tests on a global 3 km mesh with more than 65 million horizontal grid cells on up to half a million cores. We discuss necessary modifications of the model code to improve its parallel performance in general and specific to the HPC environment. We confirm good scaling (70 % parallel efficiency or better) of the MPAS model and provide numbers on the computational requirements for experiments with the 3 km mesh. In doing so, we show that global, convection-resolving atmospheric simulations with MPAS are within reach of current and next generations of high-end computing facilities.

Abstract Urban land surface processes need to be represented to inform future urban climate and building energy projections. Here, the single layer urban canopy model Town Energy Balance (TEB) is coupled to the Weather Research and Forecasting (WRF) model to create WRF‐TEB. The coupling method is described generically, implemented into software, and the code and data are released with a Singularity image to address issues of scientific reproducibility. The coupling is implemented modularly and verified by an integration test. Results show no detectable errors in the coupling. Separately, a meteorological evaluation is undertaken using observations from Toulouse, France. The latter evaluation, during an urban canopy layer heat island episode, shows reasonable ability to estimate turbulent heat flux densities and other meteorological quantities. We conclude that new model couplings should make use of integration tests as meteorological evaluations by themselves are insufficient, given that errors are difficult to attribute because of the interplay between observational errors and multiple parameterization schemes (e.g., radiation, microphysics, and boundary layer).

Abstract. Several initiatives are currently emerging to observe the exchange of energy and matter between the earth's surface and atmosphere standardized over larger space and time domains. For example, the National Ecological Observatory Network (NEON) and the Integrated Carbon Observing System (ICOS) are set to provide the ability of unbiased ecological inference across ecoclimatic zones and decades by deploying highly scalable and robust instruments and data processing. In the construction of these observatories, enclosed infrared gas analyzers are widely employed for eddy covariance applications. While these sensors represent a substantial improvement compared to their open- and closed-path predecessors, remaining high-frequency attenuation varies with site properties and gas sampling systems, and requires correction. Here, we show that components of the gas sampling system can substantially contribute to such high-frequency attenuation, but their effects can be significantly reduced by careful system design. From laboratory tests we determine the frequency at which signal attenuation reaches 50 % for individual parts of the gas sampling system. For different models of rain caps and particulate filters, this frequency falls into ranges of 2.5–16.5 Hz for CO2, 2.4–14.3 Hz for H2O, and 8.3–21.8 Hz for CO2, 1.4–19.9 Hz for H2O, respectively. A short and thin stainless steel intake tube was found to not limit frequency response, with 50 % attenuation occurring at frequencies well above 10 Hz for both H2O and CO2. From field tests we found that heating the intake tube and particulate filter continuously with 4 W was effective, and reduced the occurrence of problematic relative humidity levels (RH > 60 %) by 50 % in the infrared gas analyzer cell. No further improvement of H2O frequency response was found for heating in excess of 4 W. These laboratory and field tests were reconciled using resistor–capacitor theory, and NEON's final gas sampling system was developed on this basis. The design consists of the stainless steel intake tube, a pleated mesh particulate filter and a low-volume rain cap in combination with 4 W of heating and insulation. In comparison to the original design, this reduced the high-frequency attenuation for H2O by ≈ 3∕4, and the remaining cospectral correction did not exceed 3 %, even at high relative humidity (95 %). The standardized design can be used across a wide range of ecoclimates and site layouts, and maximizes practicability due to minimal flow resistance and maintenance needs. Furthermore, due to minimal high-frequency spectral loss, it supports the routine application of adaptive correction procedures, and enables largely automated data processing across sites.

Abstract One of the objectives of the Deep Convective Clouds and Chemistry (DC3) field experiment was to determine the scavenging of soluble trace gases by thunderstorms. We present an analysis of scavenging of hydrogen peroxide (H 2 O 2 ) and methyl hydrogen peroxide (CH 3 OOH) from six DC3 cases that occurred in Oklahoma and northeast Colorado. Estimates of H 2 O 2 scavenging efficiencies are comparable to previous studies ranging from 79 to 97% with relative uncertainties of 5–25%. CH 3 OOH scavenging efficiencies ranged from 12 to 84% with relative uncertainties of 18–558%. The wide range of CH 3 OOH scavenging efficiencies is surprising, as previous studies suggested that CH 3 OOH scavenging efficiencies would be <10%. Cloud chemistry model simulations of one DC3 storm produced CH 3 OOH scavenging efficiencies of 26–61% depending on the ice retention factor of CH 3 OOH during cloud drop freezing, suggesting ice physics impacts CH 3 OOH scavenging. The highest CH 3 OOH scavenging efficiencies occurred in two severe thunderstorms, but there is no obvious correlation between the CH 3 OOH scavenging efficiency and the storm thermodynamic environment. We found a moderate correlation between the estimated entrainment rates and CH 3 OOH scavenging efficiencies. Changes in gas‐phase chemistry due to lightning production of nitric oxide and aqueous‐phase chemistry have little effect on CH 3 OOH scavenging efficiencies. To determine why CH 3 OOH can be substantially removed from storms, future studies should examine effects of entrainment rate, retention of CH 3 OOH in frozen cloud particles during drop freezing, and lightning‐NO x production.

Abstract Tropopause‐penetrating convection is capable of rapidly transporting air from the lower troposphere to the upper troposphere and lower stratosphere (UTLS), where it can have important impacts on chemistry, the radiative budget, and climate. However, obtaining in situ measurements of convection and convective transport is difficult and such observations are historically rare. Modeling studies, on the other hand, offer the advantage of providing output related to the physical, dynamical, and chemical characteristics of storms and their environments at fine spatial and temporal scales. Since these characteristics of simulated convection depend on the chosen model design, we examine the sensitivity of simulated convective transport to the choice of physical (bulk microphysics or BMP and planetary boundary layer or PBL) and chemical parameterizations in the Weather Research and Forecasting model coupled with Chemistry (WRF‐Chem). In particular, we simulate multiple cases where in situ observations are available from the recent (2012) Deep Convective Clouds and Chemistry (DC3) experiment. Model output is evaluated using ground‐based radar observations of each storm and in situ trace gas observations from two aircraft operated during the DC3 experiment. Model results show measurable sensitivity of the physical characteristics of a storm and the transport of water vapor and additional trace gases into the UTLS to the choice of BMP. The physical characteristics of the storm and transport of insoluble trace gases are largely insensitive to the choice of PBL scheme and chemical mechanism, though several soluble trace gases (e.g., SO 2 , CH 2 O, and HNO 3 ) exhibit some measurable sensitivity.

Abstract State-of-the-art remote sensing techniques applicable to the investigation of ice formation and evolution are described. Ground-based and spaceborne measurements with lidar, radar, and radiometric techniques are discussed together with a global view on past and ongoing remote sensing measurement campaigns concerned with the study of ice formation and evolution. This chapter has the intention of a literature study and should illustrate the major efforts that are currently taken in the field of remote sensing of atmospheric ice. Since other chapters of this monograph mainly focus on aircraft in situ measurements, special emphasis is put on active remote sensing instruments and synergies between aircraft in situ measurements and passive remote sensing methods. The chapter concentrates on homogeneous and heterogeneous ice formation in the troposphere because this is a major topic of this monograph. Furthermore, methods that deliver direct, process-level information about ice formation are elaborated with a special emphasis on active remote sensing methods. Passive remote sensing methods are also dealt with but only in the context of synergy with aircraft in situ measurements.

Abstract A series of webinars and panel discussions were conducted on the topic of the evolving role of humans in weather prediction and communication, in recognition of the 100th anniversary of the founding of the AMS. One main theme that arose was the inevitability that new tools using artificial intelligence will improve data analysis, forecasting, and communication. We discussed what tools are being created, how they are being created, and how the tools will potentially affect various duties for operational meteorologists in multiple sectors of the profession. Even as artificial intelligence increases automation, humans will remain a vital part of the forecast process as that process changes over time. Additionally, both university training and professional development must be revised to accommodate the evolving forecasting process, including addressing the need for computing and data skills (including artificial intelligence and visualization), probabilistic and ensemble forecasting, decision support, and communication skills. These changing skill sets necessitate that both the U.S. Government’s Meteorologist General Schedule 1340 requirements and the AMS standards for a bachelor’s degree need to be revised. Seven recommendations are presented for student and forecaster preparation and career planning, highlighting the need for students and operational meteorologists to be flexible lifelong learners, acquire new skills, and be engaged in the changes to forecast technology in order to best serve the user community throughout their careers. The article closes with our vision for the ways that humans can maintain an essential role in weather prediction and communication, highlighting the interdependent relationship between computers and humans.

Abstract The 2010 Deepwater Horizon oil spill in the Gulf of Mexico (GoM) was an environmental disaster, which highlighted the urgent need to predict the transport and dispersion of hydrocarbon. Although the variability of the atmospheric forcing plays a major role in the upper ocean circulation and transport of the pollutants, the air‐sea interaction on various time scales is not well understood. This study provides a comprehensive overview of the atmospheric forcing and upper ocean response in the GoM from seasonal to diurnal time scales, using climatologies derived from long‐term observations, in situ observations from two field campaigns, and a coupled model. The atmospheric forcing in the GoM is characterized by striking seasonality. In the summer, the time‐average large‐scale forcing is weak, despite occasional extreme winds associated with hurricanes. In the winter, the atmospheric forcing is much stronger, and dominated by synoptic variability on time scales of 3–7 days associated with winter storms and cold air outbreaks. The diurnal cycle is more pronounced during the summer, when sea breeze circulations affect the coastal regions and nighttime wind maxima occur over the offshore waters. Realtime predictions from a high‐resolution atmosphere‐wave‐ocean coupled model were evaluated for both summer and winter conditions during the Grand LAgrangian Deployment (GLAD) in July–August 2012 and the Surfzone Coastal Oil Pathways Experiment (SCOPE) in November–December 2013. The model generally captured the variability of atmospheric forcing on all scales, but suffered from some systematic errors.

Abstract The control variable transform (CVT) is a keystone of variational data assimilation. In publications using such a technique, the background term of the transformed cost function is defined as a canonical inner product of the transformed control variable with itself. However, it is shown in this paper that this practical definition of the cost function is not correct if the CVT uses a square root of the background error covariance matrix that is not square. Fortunately, it is then shown that there is a manifold of the control space for which this flaw has no impact, and that most minimizers used in practice precisely work in this manifold. It is also shown that both correct and practical transformed cost functions have the same minimum. This explains more rigorously why the CVT is working in practice. The case of a singular is finally detailed, showing that the practical cost function still reaches the best linear unbiased estimate (BLUE).

Abstract This study presents a systematic analysis of convective parameterizations performance with interactive radiation, microphysics, and surface on an idealized day with shallow convection. To this end, we analyze a suite of mesoscale numerical experiments (i.e., with parameterized turbulence). In the first set, two different convection schemes represent shallow convection at a 9-km resolution. These experiments are then compared with model results omitting convective parameterizations at 9- and 3-km horizontal resolution (gray zone). Relevant in our approach is to compare the results against two simulations by different large-eddy simulation (LES) models. Results show that the mesoscale experiments, including the 3-km resolution, are unable to adequately represent the timing, intensity, height, and extension of the shallow cumulus field. The main differences with LES experiments are the following: a too late onset, too high cloud base, and a too early transport of moisture too high, overestimating the second cloud layer. Related to this, both convective parameterizations produce warm and dry biases of up to 2 K and 2 g kg−1, respectively, in the cloud layer. This misrepresentation of the cloud dynamics leads to overestimated shortwave radiation variability, both spacewise and timewise. Domain-averaged shortwave radiation at the surface, however, compares satisfactorily with LES. The shortwave direct and diffuse partition is misrepresented by the convective parameterizations with an underestimation (overestimation) of diffuse (direct) radiation both locally and, by a relative 40% (10%), of the domain average.

ABSTRACT Climate change caused by green house gas emissions is now following the trend of rapid warming consistent with a RCP8.5 forcing. Climate models are still unable to represent the mesoscale convective processes that occur at resolutions ∼ O (3 km) and are not capable of resolving precipitation patterns in time and space with sufficient accuracy to represent convection. In this article, the UK Met Office precipitation observations are compared with the simulations for the period 1990–1995 followed by a simulation of a near‐future period 2031–2036 for a regional nested weather model. The convection‐permitting model, resolution ∼ O (3 km), provides a good correspondence to the observational precipitation data and demonstrates the importance of explicit convection for future summer precipitation estimates. The UK summer precipitation is reduced slightly (∼10%) for 2031–2036 and there is no evidence of an increase in the peak maximum hourly precipitation magnitude. A similar pattern is observed over the whole European inner model domain. The results using the Kain–Fritsch convective parameterization scheme at a resolution ∼ O (12 km) in the outer domain increase summer precipitation by ∼10% for the UK. The average precipitation rate per event increases, dry periods extend and wet periods shorten. As part of the change, 10‐m winds of <3 m s −1 become more common – a scenario that would impact on power generation from wind turbines through calmer conditions and cause more frequent pollution episodes.

Prediction of the potentially devastating impact of landfalling tropical cyclones (TCs) relies substantially on numerical prediction systems. Due to the limited predictability of TCs and the need to express forecast confidence and possible scenarios, it is vital to exploit the benefits of dynamic ensemble forecasts in operational TC forecasts and warnings. RSMCs, TCWCs, and other forecast centers value probabilistic guidance for TCs, but the International Workshop on Tropical Cyclones (IWTC-9) found that the “pull-through” of probabilistic information to operational warnings using those forecasts is slow. IWTC-9 recommendations led to the formation of the WMO/WWRP Tropical Cyclone-Probabilistic Forecast Products (TC-PFP) project, which is also endorsed as a WMO Seamless GDPFS Pilot Project. The main goal of TC-PFP is to coordinate across forecast centers to help identify best practice guidance for probabilistic TC forecasts. TC-PFP is being implemented in 3 phases: Phase 1 (TC formation and position); Phase 2 (TC intensity and structure); and Phase 3 (TC related rainfall and storm surge). This article provides a summary of Phase 1 and reviews the current state of the science of probabilistic forecasting of TC formation and position. There is considerable variability in the nature and interpretation of forecast products based on ensemble information, making it challenging to transfer knowledge of best practices across forecast centers. Communication among forecast centers regarding the effectiveness of different approaches would be helpful for conveying best practices. Close collaboration with experts experienced in communicating complex probabilistic TC information and sharing of best practices between centers would help to ensure effective decisions can be made based on TC forecasts. Finally, forecast centers need timely access to ensemble information that has consistent, user-friendly ensemble information. Greater consistency across forecast centers in data accessibility, probabilistic forecast products, and warnings and their communication to users will produce more reliable information and support improved outcomes.

Abstract This study investigates the effects of lakes in mountainous terrain on the evolution of supercell thunderstorms. With a newly developed radar-based, mesocyclone-detection algorithm, a recent study has characterized the occurrence and evolution of supercell thunderstorms in the Swiss Alpine region. That study highlights the influence of orography on both storm intensity and occurrence frequency. To disentangle the different influential factors, an idealized modeling framework is established here using the mesoscale model CM1. The modeling scenarios are based on a high-CAPE environment with unidirectional shear, where a warm bubble serves to initiate the convection. Mimicking the environment of the southern Prealps in central Europe, scenarios with a high mountain ridge, valleys, and lakes are explored. The effect on the supercells of the slopes, high-altitude terrain, and moisture sources emphasizes the highly localized nature of terrain effects, leading to a heterogeneous intensity life cycle with transitory enhancement and weakening of the supercell. The dynamic and thermodynamic impact of mountain valleys with lakes increases the range of atmospheric conditions that supports supercellular development through horizontal vorticity production, increased storm relative helicity, and higher moisture content. This influence results in a systematic location dependence of the frequency, intensity, and lifetime of supercells, as also found in observations.

Abstract The contribution of heterogeneous ice nucleation to the formation of cirrus cloud ice crystals is still not well quantified. This results in large uncertainties when predicting cirrus radiative effects and their role in Earth’s climate system. The goal of this case study is to simulate the composition, and thus activation conditions, of ice nucleating particles (INPs) to evaluate their contribution to heterogeneous cirrus ice formation in relation to homogeneous ice nucleation. For this, the regional model COSMO—Aerosols and Reactive Trace Gases (COSMO-ART) was used to simulate a synoptic cirrus cloud over Texas on 13 April 2011. The simulated INP composition was then compared to measured ice residual particle (IRP) composition from the actual event obtained during the NASA Midlatitude Airborne Cirrus Properties Experiment (MACPEX) aircraft campaign. These IRP measurements indicated that the dominance of heterogeneous ice nucleation was mainly driven by mineral dust with contributions from a variety of other particle types. Applying realistic activation thresholds and concentrations of airborne transported mineral dust and biomass-burning particles, the model implementing the heterogeneous ice nucleation parameterization scheme of Ullrich et al. is able to reproduce the overall dominating ice formation mechanism in contrast to the model simulation with the scheme of Phillips et al. However, the model showed flaws in reproducing the IRP composition.

The Global Navigation Satellite System (GNSS) airborne radio occultation (ARO) technique is used to retrieve profiles of the atmosphere during reconnaissance missions for atmospheric rivers (ARs) on the west coast of the United States. The measurements are a horizontal integral of refractive index over long ray-paths extending between a spaceborne transmitter and a receiver onboard an aircraft. A specialized forward operator is required to allow assimilation of ARO observations into numerical weather prediction models to support forecasting of ARs. A two-dimensional (2D) bending angle operator is proposed to enable capturing key atmospheric features associated with strong ARs. Comparison to a one-dimensional (1D) forward model supports the evidence of large bending angle departures within 3-7 km impact heights for observations collected in a region characterized by the integrated water vapor transport (IVT) magnitude above 500 kg m-1 s-1. The assessment of the 2D forward model for ARO retrievals is based on a sequence of six flights leading up to a significant AR precipitation event in January 2021. Since the observations often sampled regions outside the AR where moisture is low, the significance of horizontal variations is obscured in the average statistics. However, examples from an individual flight preferentially sampling the cross-section of an AR further support the need for the 2D forward model for targeted ARO observations. Additional simulation experiments are performed to quantify forward modeling errors due to tangent point drift and horizontal gradients suggesting contributions on the order of 5 % and 20 %, respectively.

Abstract. HOLODEC, an airborne cloud particle imager, captures holographic images of a fixed volume of cloud to characterize the types and sizes of cloud particles, such as water droplets and ice crystals. Cloud particle properties include position, diameter, and shape. In this work we evaluate the potential for processing HOLODEC data by leveraging a combination of GPU hardware and machine learning with the eventual goal of improving HOLODEC processing speed and performance. We present a hologram processing algorithm, HolodecML, that utilizes a neural network segmentation model and computational parallelization to achieve these goals. HolodecML is trained using synthetically generated holograms based on a model of the instrument, and predicts masks around particles found within reconstructed images. From these masks, the position and size of the detected particles can be characterized in three dimensions. In order to successfully process real holograms, we find we must apply a series of image corrupting transformations and noise to the synthetic images used in training. In this evaluation, HolodecML had comparable position and size estimations performance to the standard processing method, but improved particle detection by nearly 20 % on several thousand manually labeled HOLODEC images. However, the particle detection improvement only occurred when image corruption was performed on the simulated images during training, thereby mimicking non-ideal conditions in the actual probe. The trained model also learned to differentiate artifacts and other impurities in the HOLODEC images from the particles, even though no such objects were present in the training data set. By contrast, the standard processing method struggled to separate particles from artifacts. HolodecML also leverages GPUs and parallel computing that enables large processing speed gains over serial and CPU-only based evaluation. Our results demonstrate that the machine-learning based framework may be a possible path to both improving and accelerating hologram processing. The novelty of the training approach, which leveraged noise as a means for parameterizing non-ideal aspects of the HOLODEC detector, could be applied in other domains where the theoretical model is incapable of fully describing the real-world operation of the instrument and accurate truth data required for supervised learning cannot be obtained from real-world observations.