Workshop on
particle-based modeling of cloud microphysics

Particle-based modeling of cloud microphysics is becoming popular during the last 10 years. This workshop provides an opportunity to meet researchers in related fields, and exchange ideas to explore the possibilities of the methodology.

Date: 19-20 November 2018

Venue: 1F seminar room, RIKEN Center for Computational Science(R-CCS), Kobe, Japan. (


Registration (free):

Deadlines: On-site registration available (14 Oct 17 Sep (speakers), 19 Nov 12 Nov (other attendees))


  • Gustavo Abade (University of Warsaw)
  • Sylwester Arabas (Jagiellonian University)
  • Daniel Cardoso Cordeiro (Osaka University)
  • Piotr Dziekan (University of Warsaw)
  • Wojciech W. Grabowski (NCAR)
  • Fabian Hoffmann (NOAA/CIRES)
  • Toshiki Matsushima (R-CCS)
  • Seiya Nishizawa (R-CCS)
  • Yign Noh (Yonsei University)
  • Ong Chia Rui (University of Tokyo)
  • Izumi Saito (Nagoya Institute of Technology)
  • Kento Sakai (University of Hyogo)
  • Shin-ichiro Shima (University of Hyogo/R-CCS)
  • Hirofumi Tomita (R-CCS)


Mon, 19 Nov
13:00-13:05 Opening remark
Chair: Sylwester Arabas
13:05-13:45 Impact of turbulence on diffusional growth of cloud droplets: from DNS to LES and beyond,
Dr. Wojciech W. Grabowski (NCAR)
13:45-14:25 Growth of cloud droplets in turbulence in cloud microphysics simulator
Prof. Izumi Saito (Nagoya Institute of Technology)
14:25-15:05 The Linear Eddy Model as a subgrid-scale model for Lagrangian cloud models,
Dr. Fabian Hoffmann (NOAA/CIRES)
Coffee break
Chair: Yign Noh
15:30-16:10 Broadening of droplet spectra and stochastic activation in turbulent clouds,
Prof. Gustavo Abade (University of Warsaw)
16:10-16:50 Spectral broadening of cloud droplets in cumulus congestus cloud obtained by large eddy simulation with super droplet method,
Dr. Toshiki Matsushima (R-CCS)
16:50-17:30 University of Warsaw Lagrangian Cloud Model: model formulation and validation,
Dr. Piotr Dziekan (University of Warsaw)
18:30- Buffet dinner at Kobe Portopia Hotel "GOCOCU"
Tue, 20 Nov
Chair: Gustavo Abade
09:30-10:10 A Cloud Microphysics Parameterization for Shallow Cumulus Clouds Based on Lagrangian Cloud Model Simulations,
Prof. Yign Noh (Yonsei University)
10:10-10:30 Performance comparison among three Monte Carlo schemes for collision-coalescence: O'Rourke method, No-time counter method, and Super-droplet method,
Prof. Shin-ichiro Shima (University of Hyogo/R-CCS)
10:30-10:50 Application of the Super-Droplet Method to Mixed-Phase Clouds Based on the Porous Spheroid Approximation of Ice Particles,
Prof. Shin-ichiro Shima (University of Hyogo/R-CCS)
10:50-11:10 Box model evaluation of the particle-based mixed-phase cloud microphysics model,
Mr. Kento Sakai (University of Hyogo)
11:10-11:30 Large Eddy Simulation of oil blowout in water,
Mr. Daniel Cardoso Cordeiro (Osaka University)
Lunch break
Chair: Fabian Hoffmann
13:00-13:20 The improved immersed boundary method and water droplet simulations,
Mr. Ong Chia Rui (University of Tokyo)
13:20-13:40 Effect of sub-meso scale topography on precipitation,
Dr. Seiya Nishizawa (R-CCS)
13:40-14:20 Modeling of Cloud Microphysics: Can We Do Better?
Dr. Wojciech W. Grabowski (NCAR)
14:20-15:00 Particle-Based Cloud Microphysics COST Action Proposal,
Dr. Sylwester Arabas (Jagiellonian University)
15:00-15:15 Coffee break
15:15-15:45 Discussion
15:45-15:50 Closing remark


Organizers: Shin-ichiro Shima (University of Hyogo/R-CCS), Hirofumi Tomita (R-CCS)

Sponsorship: Graduate School of Simulation Studies, University of Hyogo

Co-sponsorship: RIKEN Center for Computational Science (R-CCS)

Supported by: Joint research program of the Institute for Space-Earth Environmental Research, Nagoya University; MEXT KAKENHI (Grant Number 18H04448);  JST CREST (Grant Number JPMJCR1312)

Contact: Shin-ichiro Shima (島伸一郎), e-mail:


Mon, 19 Nov

Impact of turbulence on diffusional growth of cloud droplets: from DNS to LES and beyond
Wojciech W. Grabowski
National Center for Atmospheric Research, Boulder, USA
This presentation will discuss spectral broadening of the droplet size distribution in turbulent clouds. At small-scales, this has been investigated using the direct numerical simulation over the last two decades. Those studies show that the spectral broadening is relatively small and slowly increases with the computational domain size (i.e., with the scale of the largest eddies) because of the increasing supersaturation fluctuations. At larger scales (say, meters and up) supersaturation fluctuations are limited by the quasi-equilibrium supersaturation. Spectral broadening at those scales can take place through a mechanism referred to as eddy hopping. The key idea, suggested in late 1980ies by Al Cooper, is that droplets arriving at a given location within a turbulent cloud follow different trajectories and thus have different growth histories, and that this leads to a significant spectral broadening. This presentation will discuss recent progress in this area and argue that a Lagrangian Cloud Model with a stochastic subgrid-scale scheme shows a lot of promise for studying spectral broadening in natural clouds.

Growth of cloud droplets in turbulence in cloud microphysics simulator
Izumi Saito, Toshiyuki Gotoh, Tatsuya Yasuda, Takeshi Watanabe
Nagoya Institute of Technology, Nagoya, Japan
Direct numerical simulation of turbulent air flow and the detailed computation of Lagrangian dynamics of cloud droplets with cloud microphysical processes are very useful and powerful to obtain new knowledge about the fundamental cloud physics. We have developed a DNS model "cloud microphysics simulator", and successfully computed the continuous growth of cloud droplets to rain drops from the microscopic view points. In this talk, I will provide an overview of the recent results of cloud microphysics simulator regarding the formation of the steady-state size distribution of droplets due to collision-coalescence and condensation-evaporation in turbulence.

The Linear Eddy Model as a subgrid-scale model for Lagrangian cloud models
Fabian Hoffmann
NOAA ESRL Chemical Sciences Devision; CIRES, University of Colorado Boulder
To fully understand the microphysical composition of clouds, their radiative properties, and their ability to precipitate, lengthscales of multiple orders of magnitude need to be considered. During entrainment, for example, the switch from predominantly inhomogeneous to homogenous mixing (and their distinct effects on cloud microphysics) takes place at the centimeter-scale, but standard large-eddy simulations (LES) with grid spacings on the order of decameters exert erroneous homogeneous mixing over the entire subgrid-scale. On the other hand, ultra-high resolution direct numerical simulation (DNS) captures the physics of small-scale mixing correctly, but does not represent the large-scale dynamics of the cloud responsible for entrainment. The challenge is to represent this range of scales with a single model.

In this talk, I present a novel modeling approach in which the LES subgrid-scale is represented by the ‘linear eddy model’, an economical, one-dimensional model that resolves turbulent compression, folding, and molecular diffusion in each grid box of the LES explicitly. This approach is applied to test cases of shallow cumuli and stratocumuli, and first applications for mixed-phase clouds will be discussed. Generally, clouds susceptible to inhomogeneous mixing show a reduction in the droplet number concentration and stronger droplet growth, in agreement with theory. Stratocumulus  entrainment rates tend to be lower in the new approach compared to simulations without it.

All in all, the simulations presented can be seen as a first step to bridge the gap between DNS and LES, allowing an appropriate representation of small-scale mixing processes, but also the consideration of the large-scale cloud system.

Broadening of droplet spectra and stochastic activation in turbulent clouds
Gustavo C. Abade (1), Wojciech W. Grabowski (2), Hanna Pawlowska (1)

1. Institute of Geophysics, Faculty of Physics, University of Warsaw, Warsaw, Poland
2. National Center for Atmospheric Research, Boulder, Colorado
This talk reports on the simulations of droplet-size distributions under the influence of cloud turbulence, entrainment, and activation of entrained cloud condensation nuclei (CCN). The microphysics is forced by using two idealized frameworks: (a) an entraining cloud parcel, and (b) a synthetic, turbulent-like flow that mimics a stratocumulus topped boundary layer (STBL). Cloud droplets and unactivated (CCN) are described by Lagrangian particles (superdroplets). Collisions and coalescence of droplets are not considered. The processes occurring in the range of unresolved scales (e.g., the transport of cloud particles, supersaturation fluctuations, turbulent mixing, and the resulting stochastic droplet activation and growth by condensation) are modeled using a Monte Carlo scheme. In this stochastic microphysics, supersaturation fluctuations drive a “random walk” of the droplet radii in the Köhler potential landscape set by the local resolved supersaturation. It is shown that the unresolved turbulence plays a key role in broadening the droplet-size distribution towards larger sizes. Also, the feedback on vapor of stochastically activated droplets buffers the variations of the mean supersaturation driven by updrafts and downdrafts. This extends the distance over which entrained CNN are activated inside the cloud and produces multimodal droplet-size distributions.

Spectral broadening of cloud droplets in cumulus congestus cloud obtained by large eddy simulation with super droplet
methodroadening of droplet spectra and stochastic activation in turbulent clouds
T. Matsushima1, S. Nishizawa1, S. Shima2,1 and H. Tomita1
1. RIKEN Center for Computational Science (R-CCS), Kobe, Japan
2. Graduate School of Simulation Studies, University of Hyogo, Kobe, Japan

Tue, 20 Nov

A Cloud Microphysics Parameterization for Shallow Cumulus Clouds Based on Lagrangian Cloud Model Simulation
Yign Noh (1), Donggun Oh (1), Fabian Hoffmann (2), and Siegfried Raasch (2)
1. Department of Atmospheric Sciences, Yonsei University, Seoul, South Korea
2. Institute of Meteorology and Climatology, Leibniz Universität Hannover, Hannover, Germany
Cloud microphysics parameterizations for shallow cumulus clouds are analyzed based on Lagrangian cloud model (LCM) data, focusing on autoconversion and accretion. The autoconversion and accretion rates, \(A\) and \(C\), are calculated directly by capturing the moment of the conversion of individual Lagrangian droplets from cloud droplets to raindrops, and it results in the reproduction of the formulae of \(A\) and \(C\) for the first time. Comparison with various parameterizations reveals the closest agreement with Tripoli and Cotton (1980) (TC80), such as \(A=\alpha N_c^{-1/3}q_c^{7/3}H(R-R_T)\) and \(C=\beta q_c q_r\), where  \(q_c\) and  \(N_c\) are the mixing ratio and the number concentration of cloud droplets, \(q_r\) is the mixing ratio of raindrops, \(R_T\) is the threshold volume radius, \(H\) is the Heaviside function. Furthermore, it is found that \(\alpha\) increases linearly with the dissipation rate \(\epsilon\) and the standard deviation of radius \(\sigma\), and that \(R_T\) decreases rapidly with \(\sigma\), while disappearing at  \(\sigma> 3.5\,\mathrm{\mu m}\). The LCM also reveals that \(\sigma\) and \(\epsilon\) increase with time during the period of autoconversion, which helps to suppress the early precipitation by reducing \(A\) with smaller \(\alpha\) and larger \(R_T\) in the initial stage. Finally, \(\beta\) is found to be affected by the accumulated collisional growth, which determines the drop size distribution.

Performance comparison among three Monte Carlo schemes for collision-coalescence: O'Rourke method, No-Time Counter method, and Super-Droplet method
Shin-ichiro Shima1, Hiroshi Yamaguchi2, Hitoshi Hongou3, Hideaki Yokohata3
1. Graduate School of Simulation Studies, University of Hyogo, Kobe, Japan
2. IDAJ Co., LTD., Yokohama, Japan
3. Mazda Motor Corporation, Hiroshima, Japan
There are several Monte Carlo schemes for the stochastic collision-coalescence of particles. We compared the computational performance of O'Rourke method, no-time counter method, and super-droplet method through a box-model test. We confirmed that the super-droplet method outperforms the other two at least under the condition we investigated.

Application of the Super-Droplet Method to Mixed-Phase Clouds Based on the Porous Spheroid Approximation of Ice Particles
Shin-ichiro Shima1,2, Yousuke Sato3,2, Akihiro Hashimoto4, and Ryohei Misumi5
1. Graduated School of Simulation Study, University of Hyogo, Kobe, Japan
2. RIKEN Center for Computational Science, Kobe, Japan
3. Department of Applied Energy, Graduate School of Engineering, Nagoya University, Nagoya, Japan
4. Meteorological Research Institute, Japan Meteorological Agency,Tsukuba, Japan
5. National Research Institute for Earth Science and Disaster Resilience, Tsukuba, Japan
The super-droplet method (SDM) is a particle-based and probabilistic numerical scheme, which enables accurate simulation of cloud microphysics with less demand on computation. In the SDM, the time evolution of aerosol/cloud/precipitation particles is calculated explicitly by solving the fundamental governing equations of cloud microphysics.

For the purpose of performing large eddy simulations of mixed-phase clouds, we applied the SDM to ice phase cloud microphysics. Our strategy is to translate the multicomponent bin model of Chen and Lamb (1994) into the SDM framework, in which the ice particles are represented by porous spheroids. We also introduced some refinements and updates of the theory.

In the talk, the basic equations employed for our mixed-phase SDM, and preliminary results of a cumulonimbus simulation will be presented.

Box model evaluation of the particle-based mixed-phase cloud microphysics model
Kento Sakai and Shin-ichiro Shima
Graduate School of Simulation Studies, University of Hyogo, Kobe, Japan
The cloud microphysics of mixed phase clouds is complicated. A model that approximates an ice particle with a porous spheroid is known as one of the most detailed mathematical models.  The super-droplet method enables the use of the porous spheroid model for the numerical simulation of the entire cloud. However, our understanding of mixed-phase cloud microphysics is not sufficient. Therefore, we compared our numerical results with wind tunnel measurements to verify and refine the porous spheroid model. This time, we have tested depositional growth and riming growth. We made one small ice particle and calculated its growth by deposition and riming. The model successfully reproduced the depositional growth, but tends to overestimate the growth by riming.

Large Eddy Simulation of oil blowout in water
Daniel Cardoso Cordeiro, Atsushi Sekimoto and Yasunori Okano
Department of Materials Engineering Science, Osaka University, Osaka, Japan
In 2010, one of the largest offshore oil spills of all time happened in the Gulf of Mexico. For the first time, the sub-sea injection of chemical dispersants was used to treat deepwater oil spills. However, with only few studies prior to its application, the overall effectiveness of this method is still questioned as appropriate measures of the oil droplets were not performed in situ.

Regarding deepwater oil blowouts, real-scale experiments are unpractical. Therefore, this study uses computational simulations to investigate the fate of the blowout oil treated with chemical dispersants in the area proximal to the oil release. An Euler-Euler model with a Large Eddy Simulation turbulent model was used to investigate the turbulent breakup of the oil jet into the water and the effects of surface tension, inlet diameter and velocity. The modeled droplet size distribution was compared to experimental results. The results agree fairly well for cases with a higher Reynolds number.

The improved immersed boundary method and water droplet simulations
Ong Chia Rui and Hiroaki Miura
Department of Earth and Planetary Science, Graduate School of Science, The University of Tokyo, Tokyo, Japan
A good estimation of the physical characteristics of a free-falling water droplet such as the terminal velocity, shape, oscillation amplitude, and collision efficiency under different environments are required in cloud microphysics and radar observation. There have been many previous experimental works devoted to the estimation of those characteristics; however, they are only valid for a limited range pressure and temperature. In this study, numerically, the immersed boundary method is improved for stable 2D and axisymmetric two-phase flow simulations of water droplets as a new tool to study water droplet motion. We show some test case results to demonstrate its robustness. Freely oscillating and free-fall water droplet simulations are performed. In particular, the terminal velocity and drag coefficient are compared to some existing experimental data and empirical formulae. Preliminary results of the collision and merging of two free-falling water droplets are also presented. Based on the successful simulations of 2D water droplets, we are in the process of developing a 3D version.

(CANCELED) Necessity of theoretical cloud-turbulence scheme for simulation of large-scale cloud organization
Hirofumi Tomita, Kenta Sueki, Toshiki Matsushima, Seiya Nishizawa
RIKEN Center for Computational Science (R-CCS)
We introduce our on-going research for the statistics of clouds through their organization. As a first step, using an LES with a simple microphysics scheme, deep convections in a Madden-Julian Oscillation is investigated from the viewpoint of numerical convergence. The simulation will reveal several problems, e.g., the energy spectrum in the dense cloud area, by thus conventional schemes. Based on this knowledge, we will discuss the expectation of state-of-the-art microphysics and turbulence scheme such as Lagrange-based microphysics with turbulence scheme accompanied with condensation.

Effect of sub-meso scale topography on precipitation
Seiya Nishizawa
RIKEN Center for Computational Science (R-CCS)
It is known that the terrain has large impacts on cumulus convection. Large scale topography forces horizontal convergence or upward flow
that causes the cumulus convection. However, the influence of smaller scale topography than convection is not known. We found that sub-meso scale topography can weaken local summer-time precipitation like heat lightning through the LES experiment. This is a preliminary result, but we speculate that atmospheric turbulence may play an important role in the weakening.

Modeling of cloud microphysics: Can we do better?
Wojciech W. Grabowski1, Hugh Morrison1, Shin-ichiro Shima2, Gustavo C. Abade3, Piotr Dziekan3, Hanna Pawlowska3, and Fabian Hoffmann4
1Mesoscale and Microscale Meteorology Laboratory, NCAR, Boulder, USA
2University of Hyogo and RIKEN Center for Computational Science, Kobe, Japan
3Institute of Geophysics, Faculty of Physics, University of Warsaw, Warsaw, Poland
4CIRES University of Colorado and NOAA ESRL Chemical Sciences Division, Boulder, Colorado, USA
Representation of cloud microphysics is a key aspect of simulating clouds. From the early days of cloud modeling, numerical models have relied on an Eulerian approach for all cloud variables, not only for temperature and water vapor, but also for cloud condensate and precipitation. Over time the sophistication of microphysics schemes has steadily increased, from simple single- moment bulk warm-rain schemes, through double- and triple-moment bulk warm-rain and ice schemes, to complex bin (spectral) schemes that predict the evolution of cloud and precipitation particle size distributions. As computational resources grow, there is a clear trend toward wider use of bin schemes, including their use as benchmarks to develop and test simplified bulk schemes. We argue that continuing on this path brings fundamental challenges difficult to overcome. This is because of the complexity of processes involved (especially for ice), the multiscale nature of cloud-scale flows that Eulerian approaches are not able to cope with, conceptual issues with the Smoluchowski equation that is solved by bin schemes to predict evolution of the particle size distributions, and numerical problems when applying bin schemes in multidimensional cloud simulations. The Lagrangian particle-based probabilistic approach is a practical alternative in which the myriad of cloud and precipitation particles present in a natural cloud is represented by a judiciously selected ensemble of point particles called super-droplets or super-particles. Advantages of the Lagrangian particle-based approach when compared to the Eulerian bin methodology will be discussed and illustrated with computational examples. Prospects of applying the Lagrangian particle-based approach to more comprehensive simulations involving clouds, for instance targeting deep convection or frontal cloud systems, will be discussed.

Particle-Based Cloud Microphysics COST Action Proposal,
Sylwester Arabas
Division of Computational Mathematics, Jagiellonian University, Cracow, Poland
COST stands for European Cooperation in Science and Technology ( and is the longest-running European framework for supporting scientific and technological collaboration. Currently funded through the European Commission’s H2020, COST collects proposals for creation of research collaboration networks called COST Actions. COST funding is aimed at supporting collaboration between institutions from COST Member Countries in Europe and Mediterranean, as well as from so-called International Partner Countries including Japan, Korea and US. On the occasion of the workshop in Kobe, I will put forward a plan for a Particle-Based Aerosol-Cloud Microphysics COST Action to be proposed in the next COST open call in September 2019.