slides, in English - Faculty of Physics

Aerosol processing by drizzling stratocumulus:
a modelling study using a novel particle-based approach
Sylwester Arabas
Anna Jaruga
Hanna Pawłowska
University of Warsaw
Faculty of Physics
Institute of Geophysics
8 th International Conference on Cloud and Precipitation
Leipzig, August 1 st 2012

aerosol processing: concepts
• interactions: aerosol —⊲ cloud & precipitation —⊲ aerosol
• processed CCN formed by evaporation of
• collisionally-grown drops
• drops within which irreversible oxidation occurred
• CCN spectrum modification by wet deposition
• adequate cloud µ-physics representations?

aerosol processing: concepts
• interactions: aerosol —⊲ cloud & precipitation —⊲ aerosol
• processed CCN formed by evaporation of
• collisionally-grown drops
• drops within which irreversible oxidation occurred
• CCN spectrum modification by wet deposition
• adequate cloud µ-physics representations?

aerosol processing: concepts
• interactions: aerosol —⊲ cloud & precipitation —⊲ aerosol
• processed CCN formed by evaporation of
• collisionally-grown drops
• drops within which irreversible oxidation occurred
• CCN spectrum modification by wet deposition
• adequate cloud µ-physics representations?

aerosol processing: concepts
• interactions: aerosol —⊲ cloud & precipitation —⊲ aerosol
• processed CCN formed by evaporation of
• collisionally-grown drops
• drops within which irreversible oxidation occurred
• CCN spectrum modification by wet deposition
• adequate cloud µ-physics representations?

aerosol processing: concepts
• interactions: aerosol —⊲ cloud & precipitation —⊲ aerosol
• processed CCN formed by evaporation of
• collisionally-grown drops
• drops within which irreversible oxidation occurred
• CCN spectrum modification by wet deposition
• adequate cloud µ-physics representations?

aerosol processing: adequate µ-physics?
key features of the Lagrangian (in size) approach:
• diffusive error-free particle growth schemes
(aka ”moving sectional”)
• scales better than ND-bin
with growing number of particle attributes
• . . .
coupled with Lagrangian-in-space super-droplet approach

aerosol processing: adequate µ-physics?
single-moment bulk:
no influence of aerosol on cloud
key features of the Lagrangian (in size) approach:
• diffusive error-free particle growth schemes
(aka ”moving sectional”)
• scales better than ND-bin
with growing number of particle attributes
• . . .
coupled with Lagrangian-in-space super-droplet approach

aerosol processing: adequate µ-physics?
single-moment bulk:
no influence of aerosol on cloud
multi-moment bulk,
1D-bin:
aerosol may influence cloud
no "memory" of drop nuclei
key features of the Lagrangian (in size) approach:
• diffusive error-free particle growth schemes
(aka ”moving sectional”)
• scales better than ND-bin
with growing number of particle attributes
• . . .
coupled with Lagrangian-in-space super-droplet approach

aerosol processing: adequate µ-physics?
"dry radius" (nucleus)
"wet radius" (solution drop)
single-moment bulk:
no influence of aerosol on cloud
multi-moment bulk,
1D-bin:
aerosol may influence cloud
no "memory" of drop nuclei
key features of the Lagrangian (in size) approach:
• diffusive error-free particle growth schemes
(aka ”moving sectional”)
• scales better than ND-bin
with growing number of particle attributes
• . . .
coupled with Lagrangian-in-space super-droplet approach

aerosol processing: adequate µ-physics?
"dry radius" (nucleus)
2D-bin
Eulerian
"wet radius" (solution drop)
single-moment bulk:
no influence of aerosol on cloud
multi-moment bulk,
1D-bin:
aerosol may influence cloud
no "memory" of drop nuclei
key features of the Lagrangian (in size) approach:
• diffusive error-free particle growth schemes
(aka ”moving sectional”)
• scales better than ND-bin
with growing number of particle attributes
• . . .
coupled with Lagrangian-in-space super-droplet approach

aerosol processing: adequate µ-physics?
"dry radius" (nucleus)
2D-bin
Eulerian
Lagrangian
"wet radius" (solution drop)
single-moment bulk:
no influence of aerosol on cloud
multi-moment bulk,
1D-bin:
aerosol may influence cloud
no "memory" of drop nuclei
key features of the Lagrangian (in size) approach:
• diffusive error-free particle growth schemes
(aka ”moving sectional”)
• scales better than ND-bin
with growing number of particle attributes
• . . .
coupled with Lagrangian-in-space super-droplet approach

aerosol processing: adequate µ-physics?
"dry radius" (nucleus)
2D-bin
Eulerian
Lagrangian
"wet radius" (solution drop)
single-moment bulk:
no influence of aerosol on cloud
multi-moment bulk,
1D-bin:
aerosol may influence cloud
no "memory" of drop nuclei
key features of the Lagrangian (in size) approach:
• diffusive error-free particle growth schemes
(aka ”moving sectional”)
• scales better than ND-bin
with growing number of particle attributes
• . . .
coupled with Lagrangian-in-space super-droplet approach

aerosol processing: adequate µ-physics?
"dry radius" (nucleus)
2D-bin
Eulerian
Lagrangian
"wet radius" (solution drop)
single-moment bulk:
no influence of aerosol on cloud
multi-moment bulk,
1D-bin:
aerosol may influence cloud
no "memory" of drop nuclei
key features of the Lagrangian (in size) approach:
• diffusive error-free particle growth schemes
(aka ”moving sectional”)
• scales better than ND-bin
with growing number of particle attributes
• . . .
coupled with Lagrangian-in-space super-droplet approach

Lagrangian µ-physics: key elements
• each particle (aka super-droplet) many ”similar” real-world particles
• attributes: multiplicity, dry radius, wet radius, nucleus type, . . .
• aerosol, cloud, precip. particles not distinguished, subject to same processes
Eulerian / PDE
advection of heat
advection of moisture
Lagrangian / ODE
particle transport by the flow
condensational growth
collisional growth
sedimentation
∑
∂ t (ρ d r) + ∇(⃗vρ d r) = ρ d ṙ ṙ =
. . .
particles ∈ ∆V
∑
∂ t (ρ d θ) + ∇(⃗vρ d θ) = ρ d ˙θ ˙θ = . . .
particles ∈ ∆V
advection of trace gases
in-particle aqueous chemistry
. . . . . .
• recent examples in context of precipitating clouds:
• Shima et al. 2009, QJ
• Andrejczuk et al. 2010, JGR
• Riechelmann et al. 2012, NJP

Lagrangian µ-physics: key elements
• each particle (aka super-droplet) many ”similar” real-world particles
• attributes: multiplicity, dry radius, wet radius, nucleus type, . . .
• aerosol, cloud, precip. particles not distinguished, subject to same processes
Eulerian / PDE
advection of heat
advection of moisture
Lagrangian / ODE
particle transport by the flow
condensational growth
collisional growth
sedimentation
∑
∂ t (ρ d r) + ∇(⃗vρ d r) = ρ d ṙ ṙ =
. . .
particles ∈ ∆V
∑
∂ t (ρ d θ) + ∇(⃗vρ d θ) = ρ d ˙θ ˙θ = . . .
particles ∈ ∆V
advection of trace gases
in-particle aqueous chemistry
. . . . . .
• recent examples in context of precipitating clouds:
• Shima et al. 2009, QJ
• Andrejczuk et al. 2010, JGR
• Riechelmann et al. 2012, NJP

Lagrangian µ-physics: key elements
• each particle (aka super-droplet) many ”similar” real-world particles
• attributes: multiplicity, dry radius, wet radius, nucleus type, . . .
• aerosol, cloud, precip. particles not distinguished, subject to same processes
Eulerian / PDE
advection of heat
advection of moisture
Lagrangian / ODE
particle transport by the flow
condensational growth
collisional growth
sedimentation
∑
∂ t (ρ d r) + ∇(⃗vρ d r) = ρ d ṙ ṙ =
. . .
particles ∈ ∆V
∑
∂ t (ρ d θ) + ∇(⃗vρ d θ) = ρ d ˙θ ˙θ = . . .
particles ∈ ∆V
advection of trace gases
in-particle aqueous chemistry
. . . . . .
• recent examples in context of precipitating clouds:
• Shima et al. 2009, QJ
• Andrejczuk et al. 2010, JGR
• Riechelmann et al. 2012, NJP

Lagrangian µ-physics: key elements
• each particle (aka super-droplet) many ”similar” real-world particles
• attributes: multiplicity, dry radius, wet radius, nucleus type, . . .
• aerosol, cloud, precip. particles not distinguished, subject to same processes
Eulerian / PDE
advection of heat
advection of moisture
Lagrangian / ODE
particle transport by the flow
condensational growth
collisional growth
sedimentation
∑
∂ t (ρ d r) + ∇(⃗vρ d r) = ρ d ṙ ṙ =
. . .
particles ∈ ∆V
∑
∂ t (ρ d θ) + ∇(⃗vρ d θ) = ρ d ˙θ ˙θ = . . .
particles ∈ ∆V
advection of trace gases
in-particle aqueous chemistry
. . . . . .
• recent examples in context of precipitating clouds:
• Shima et al. 2009, QJ
• Andrejczuk et al. 2010, JGR
• Riechelmann et al. 2012, NJP

Lagrangian µ-physics: key elements
• each particle (aka super-droplet) many ”similar” real-world particles
• attributes: multiplicity, dry radius, wet radius, nucleus type, . . .
• aerosol, cloud, precip. particles not distinguished, subject to same processes
Eulerian / PDE
advection of heat
advection of moisture
Lagrangian / ODE
particle transport by the flow
condensational growth
collisional growth
sedimentation
∑
∂ t (ρ d r) + ∇(⃗vρ d r) = ρ d ṙ ṙ =
. . .
particles ∈ ∆V
∑
∂ t (ρ d θ) + ∇(⃗vρ d θ) = ρ d ˙θ ˙θ = . . .
particles ∈ ∆V
advection of trace gases
in-particle aqueous chemistry
. . . . . .
• recent examples in context of precipitating clouds:
• Shima et al. 2009, QJ
• Andrejczuk et al. 2010, JGR
• Riechelmann et al. 2012, NJP

Lagrangian µ-physics: key elements
• each particle (aka super-droplet) many ”similar” real-world particles
• attributes: multiplicity, dry radius, wet radius, nucleus type, . . .
• aerosol, cloud, precip. particles not distinguished, subject to same processes
Eulerian / PDE
advection of heat
advection of moisture
Lagrangian / ODE
particle transport by the flow
condensational growth
collisional growth
sedimentation
∑
∂ t (ρ d r) + ∇(⃗vρ d r) = ρ d ṙ ṙ =
. . .
particles ∈ ∆V
∑
∂ t (ρ d θ) + ∇(⃗vρ d θ) = ρ d ˙θ ˙θ = . . .
particles ∈ ∆V
advection of trace gases
in-particle aqueous chemistry
. . . . . .
• recent examples in context of precipitating clouds:
• Shima et al. 2009, QJ
• Andrejczuk et al. 2010, JGR
• Riechelmann et al. 2012, NJP

Lagrangian µ-physics: key elements
• each particle (aka super-droplet) many ”similar” real-world particles
• attributes: multiplicity, dry radius, wet radius, nucleus type, . . .
• aerosol, cloud, precip. particles not distinguished, subject to same processes
Eulerian / PDE
advection of heat
advection of moisture
Lagrangian / ODE
particle transport by the flow
condensational growth
collisional growth
sedimentation
∑
∂ t (ρ d r) + ∇(⃗vρ d r) = ρ d ṙ ṙ =
. . .
particles ∈ ∆V
∑
∂ t (ρ d θ) + ∇(⃗vρ d θ) = ρ d ˙θ ˙θ = . . .
particles ∈ ∆V
advection of trace gases
in-particle aqueous chemistry
. . . . . .
• recent examples in context of precipitating clouds:
• Shima et al. 2009, QJ
• Andrejczuk et al. 2010, JGR
• Riechelmann et al. 2012, NJP

Int. Cloud Modelling Workshop 2012 ”drizzling Sc case”
(Wojciech Grabowski & Zach Lebo)
• VOCALS-inspired
• 2D prescribed-flow (single eddy)
• bi-modal initial dry aerosol spectrum
• details:
http://rap.ucar.edu/˜ gthompsn/workshop2012/case1/

Int. Cloud Modelling Workshop 2012 ”drizzling Sc case”
(Wojciech Grabowski & Zach Lebo)
• VOCALS-inspired
• 2D prescribed-flow (single eddy)
• bi-modal initial dry aerosol spectrum
• details:
http://rap.ucar.edu/˜ gthompsn/workshop2012/case1/

Int. Cloud Modelling Workshop 2012 ”drizzling Sc case”
(Wojciech Grabowski & Zach Lebo)
• VOCALS-inspired
• 2D prescribed-flow (single eddy)
• bi-modal initial dry aerosol spectrum
• details:
http://rap.ucar.edu/˜ gthompsn/workshop2012/case1/
super-droplet conc. [1/dx/dy/dz]
cloud droplet conc. [1/cm 3 ]
aerosol concentration [1/cm 3 ]
1.4
140
1.4
140
1.4
140
1.2
120
1.2
120
1.2
120
1
100
1
100
1
100
Y [km]
0.8
0.6
80
60
Y [km]
0.8
0.6
80
60
Y [km]
0.8
0.6
80
60
0.4
40
0.4
40
0.4
40
0.2
20
0.2
20
0.2
20
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0

icicle
http://icicle.igf.fuw.edu.pl/
• icicle – a new open-source advection eq. systems solver
• Eulerian advection: MPDATA (Smolarkiewicz 1983, . . . )
• . . .
• icicle’s Lagrangian µ-physics module:
• coalescence: Super-Droplet Monte-Carlo Scheme (Shima et al. 2009)
• aerosol hygroscopicity: κ-Köhler (Petters & Kreidenweis, 2007)
• gravitational sedimentation: Khvorostyanov & Curry, 2002
• . . .
• implementation: C++ / Thrust (GPU-ready!)

icicle
http://icicle.igf.fuw.edu.pl/
• icicle – a new open-source advection eq. systems solver
• Eulerian advection: MPDATA (Smolarkiewicz 1983, . . . )
• . . .
• icicle’s Lagrangian µ-physics module:
• coalescence: Super-Droplet Monte-Carlo Scheme (Shima et al. 2009)
• aerosol hygroscopicity: κ-Köhler (Petters & Kreidenweis, 2007)
• gravitational sedimentation: Khvorostyanov & Curry, 2002
• . . .
• implementation: C++ / Thrust (GPU-ready!)

icicle
http://icicle.igf.fuw.edu.pl/
• icicle – a new open-source advection eq. systems solver
• Eulerian advection: MPDATA (Smolarkiewicz 1983, . . . )
• . . .
• icicle’s Lagrangian µ-physics module:
• coalescence: Super-Droplet Monte-Carlo Scheme (Shima et al. 2009)
• aerosol hygroscopicity: κ-Köhler (Petters & Kreidenweis, 2007)
• gravitational sedimentation: Khvorostyanov & Curry, 2002
• . . .
• implementation: C++ / Thrust (GPU-ready!)

Monte-Carlo coalescence scheme (Shima et al. 2009)
• for all n super-droplets in a grid box of volume ∆V in timestep ∆t
• each representing ξ real particles (aerosol/cloud/drizzle/rain)
• the probability of coalescence of i-th and j-th super-droplets is:
P ij = max(ξ i , ξ j ) · E(r i , r j ) · π(r i + r j ) 2 · |v i − v j | · ∆t
} {{ } ∆V · n·(n−1)
2
/ [ ]
n
2
coalescence kernel
where r – drop radii, E(r i , r j ) – collection efficiency, v – drop velocities
• coalescence takes place once in a number of timesteps (def. by P ij )
• all min(ξ i ,ξ j ) droplets coalesce
there’s always a ”bin” of the right size to store the collided particles
• collisions triggered by comparing a uniform random number with P ij
• extensive parameters summed ( conserved), intensive averaged
• [n/2] random non-overlapping (i,j) pairs examined instead of all (i,j) pairs
cost: O ( n 2) O ( n ) , probability upscaled by n·(n−1)
2
/ [ ]
n
2

Monte-Carlo coalescence scheme (Shima et al. 2009)
• for all n super-droplets in a grid box of volume ∆V in timestep ∆t
• each representing ξ real particles (aerosol/cloud/drizzle/rain)
• the probability of coalescence of i-th and j-th super-droplets is:
P ij = max(ξ i , ξ j ) · E(r i , r j ) · π(r i + r j ) 2 · |v i − v j | · ∆t
} {{ } ∆V · n·(n−1)
2
/ [ ]
n
2
coalescence kernel
where r – drop radii, E(r i , r j ) – collection efficiency, v – drop velocities
• coalescence takes place once in a number of timesteps (def. by P ij )
• all min(ξ i ,ξ j ) droplets coalesce
there’s always a ”bin” of the right size to store the collided particles
• collisions triggered by comparing a uniform random number with P ij
• extensive parameters summed ( conserved), intensive averaged
• [n/2] random non-overlapping (i,j) pairs examined instead of all (i,j) pairs
cost: O ( n 2) O ( n ) , probability upscaled by n·(n−1)
2
/ [ ]
n
2

Monte-Carlo coalescence scheme (Shima et al. 2009)
• for all n super-droplets in a grid box of volume ∆V in timestep ∆t
• each representing ξ real particles (aerosol/cloud/drizzle/rain)
• the probability of coalescence of i-th and j-th super-droplets is:
P ij = max(ξ i , ξ j ) · E(r i , r j ) · π(r i + r j ) 2 · |v i − v j | · ∆t
} {{ } ∆V · n·(n−1)
2
/ [ ]
n
2
coalescence kernel
where r – drop radii, E(r i , r j ) – collection efficiency, v – drop velocities
• coalescence takes place once in a number of timesteps (def. by P ij )
• all min(ξ i ,ξ j ) droplets coalesce
there’s always a ”bin” of the right size to store the collided particles
• collisions triggered by comparing a uniform random number with P ij
• extensive parameters summed ( conserved), intensive averaged
• [n/2] random non-overlapping (i,j) pairs examined instead of all (i,j) pairs
cost: O ( n 2) O ( n ) , probability upscaled by n·(n−1)
2
/ [ ]
n
2

Monte-Carlo coalescence scheme (Shima et al. 2009)
• for all n super-droplets in a grid box of volume ∆V in timestep ∆t
• each representing ξ real particles (aerosol/cloud/drizzle/rain)
• the probability of coalescence of i-th and j-th super-droplets is:
P ij = max(ξ i , ξ j ) · E(r i , r j ) · π(r i + r j ) 2 · |v i − v j | · ∆t
} {{ } ∆V · n·(n−1)
2
/ [ ]
n
2
coalescence kernel
where r – drop radii, E(r i , r j ) – collection efficiency, v – drop velocities
• coalescence takes place once in a number of timesteps (def. by P ij )
• all min(ξ i ,ξ j ) droplets coalesce
there’s always a ”bin” of the right size to store the collided particles
• collisions triggered by comparing a uniform random number with P ij
• extensive parameters summed ( conserved), intensive averaged
• [n/2] random non-overlapping (i,j) pairs examined instead of all (i,j) pairs
cost: O ( n 2) O ( n ) , probability upscaled by n·(n−1)
2
/ [ ]
n
2

Monte-Carlo coalescence scheme (Shima et al. 2009)
• for all n super-droplets in a grid box of volume ∆V in timestep ∆t
• each representing ξ real particles (aerosol/cloud/drizzle/rain)
• the probability of coalescence of i-th and j-th super-droplets is:
P ij = max(ξ i , ξ j ) · E(r i , r j ) · π(r i + r j ) 2 · |v i − v j | · ∆t
} {{ } ∆V · n·(n−1)
2
/ [ ]
n
2
coalescence kernel
where r – drop radii, E(r i , r j ) – collection efficiency, v – drop velocities
• coalescence takes place once in a number of timesteps (def. by P ij )
• all min(ξ i ,ξ j ) droplets coalesce
there’s always a ”bin” of the right size to store the collided particles
• collisions triggered by comparing a uniform random number with P ij
• extensive parameters summed ( conserved), intensive averaged
• [n/2] random non-overlapping (i,j) pairs examined instead of all (i,j) pairs
cost: O ( n 2) O ( n ) , probability upscaled by n·(n−1)
2
/ [ ]
n
2

Monte-Carlo coalescence scheme (Shima et al. 2009)
• for all n super-droplets in a grid box of volume ∆V in timestep ∆t
• each representing ξ real particles (aerosol/cloud/drizzle/rain)
• the probability of coalescence of i-th and j-th super-droplets is:
P ij = max(ξ i , ξ j ) · E(r i , r j ) · π(r i + r j ) 2 · |v i − v j | · ∆t
} {{ } ∆V · n·(n−1)
2
/ [ ]
n
2
coalescence kernel
where r – drop radii, E(r i , r j ) – collection efficiency, v – drop velocities
• coalescence takes place once in a number of timesteps (def. by P ij )
• all min(ξ i ,ξ j ) droplets coalesce
there’s always a ”bin” of the right size to store the collided particles
• collisions triggered by comparing a uniform random number with P ij
• extensive parameters summed ( conserved), intensive averaged
• [n/2] random non-overlapping (i,j) pairs examined instead of all (i,j) pairs
cost: O ( n 2) O ( n ) , probability upscaled by n·(n−1)
2
/ [ ]
n
2

Monte-Carlo coalescence scheme (Shima et al. 2009)
• for all n super-droplets in a grid box of volume ∆V in timestep ∆t
• each representing ξ real particles (aerosol/cloud/drizzle/rain)
• the probability of coalescence of i-th and j-th super-droplets is:
P ij = max(ξ i , ξ j ) · E(r i , r j ) · π(r i + r j ) 2 · |v i − v j | · ∆t
} {{ } ∆V · n·(n−1)
2
/ [ ]
n
2
coalescence kernel
where r – drop radii, E(r i , r j ) – collection efficiency, v – drop velocities
• coalescence takes place once in a number of timesteps (def. by P ij )
• all min(ξ i ,ξ j ) droplets coalesce
there’s always a ”bin” of the right size to store the collided particles
• collisions triggered by comparing a uniform random number with P ij
• extensive parameters summed ( conserved), intensive averaged
• [n/2] random non-overlapping (i,j) pairs examined instead of all (i,j) pairs
cost: O ( n 2) O ( n ) , probability upscaled by n·(n−1)
2
/ [ ]
n
2

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 12 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 24 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 36 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 48 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 60 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 72 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 84 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 96 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 108 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 120 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 132 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 144 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 156 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 168 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 180 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 192 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 204 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 216 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 228 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 240 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 252 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 264 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 276 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 288 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 300 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 312 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 324 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

proof-of-concept simulation with super-droplets using icicle
1.4
water vapour mixing ratio [g/kg]
8
t = 336 s
1.4
potential temperature [K]
293
Y [km]
1.2
1
0.8
0.6
7.5
7
Y [km]
1.2
1
0.8
0.6
292
291
290
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
6.5
6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
289
288
64 SD / grid cell
( low res!)
1.4
particle (> 1 um) concentration [1/cm 3 ]
140
1.4
effective radius [um] (particles > 1 um)
E(r i , r j ) = 10
( unphysical!)
1.2
1
120
100
1.2
1
100
Y [km]
0.8
0.6
0.4
80
60
40
Y [km]
0.8
0.6
0.4
10
0.2
20
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
X [km]
1

0.4
3D LES with super-droplets (Arabas & Shima 2012)
height [m]
1.2
sdm-middle-32
1.0 • 24h LES using the ”RICO” set-up (van Zanten et al. 2011)
0.8 • Nagoya Univ. CReSS model (Tsuboki 2008)
0.6
• comparison with aircraft measurements (OAP-2DS, Fast-FSSP)
0.4
sdm-middle-128
1.2
height [m]
1.0
0.8
0.6
0.4
-1 0 1 2
w [m/s]
-1 0 1
S = RH - 1 [%]
20 60 100
CDNC [cm-3]
5 10 15 20
r eff [µm]
0 0.5 1
LWC [g/m 3 ]
0.6 0.8 1
k [1]
0 2 4 6
σ r [µm]
More:
• ICCP poster no. P.8.16
• arXiv:1205.3313

Thanks for your attention!
Acknowledgements:
Shin-ichiro Shima (Hyogo Univ.)
Piotr Smolarkiewicz & Wojciech Grabowski (NCAR)
Implementation of the super-droplet µ-physics in icicle is supported by
Polish National Science Centre grant no. DEC-2011/01/N/ST10/01483
Thanks are due authors of open-source software used in icicle, incl.:
Blitz++, Thrust, Boost.units, gnuplot-iostream, . . .

Super-Droplet concentration
number of ”bins” (exchanged among ”parcels”)
number of ”parcels” (each carrying a single ”bin”)
16 SD / cell
32 SD / cell
64 SD / cell
256 SD / cell
t = 336 s
t = 336 s
t = 336 s
t = 336 s
particle (> 1 um) concentration [1/cm 3 ]
particle (> 1 um) concentration [1/cm 3 ]
particle (> 1 um) concentration [1/cm 3 ]
particle (> 1 um) concentration [1/cm 3 ]
1.4
140
1.4
140
1.4
140
1.4
140
1.2
120
1.2
120
1.2
120
1.2
120
1
100
1
100
1
100
1
100
Y [km]
0.8
0.6
80
60
Y [km]
0.8
0.6
80
60
Y [km]
0.8
0.6
80
60
Y [km]
0.8
0.6
80
60
0.4
40
0.4
40
0.4
40
0.4
40
0.2
20
0.2
20
0.2
20
0.2
20
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
0
X [km]
X [km]
X [km]
X [km]
effective radius [um] (particles > 1 um)
effective radius [um] (particles > 1 um)
effective radius [um] (particles > 1 um)
effective radius [um] (particles > 1 um)
1.4
1.4
1.4
1.4
1.2
1
100
1.2
1
100
1.2
1
100
1.2
1
100
Y [km]
0.8
0.6
10
Y [km]
0.8
0.6
10
Y [km]
0.8
0.6
10
Y [km]
0.8
0.6
10
0.4
0.4
0.4
0.4
0.2
0.2
0.2
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
1
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
1
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
1
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
1
X [km]
X [km]
X [km]
X [km]
”multiple collisions” needed for low SD conc. (cf. Shima et al. 2009)
not implemented yet in icicle!

aerosol processing: minimal modelling framework
single
parcel?
precipitation
needed
• kinematic (prescribed-flow)
perhaps still enough? (focus on µ-physics)
• decoupled from cloud dynamics
computationally cheap

aerosol processing: minimal modelling framework
single
parcel?
precipitation
needed
single
column?
up- & downdraft
needed
• kinematic (prescribed-flow)
perhaps still enough? (focus on µ-physics)
• decoupled from cloud dynamics
computationally cheap

aerosol processing: minimal modelling framework
single
parcel?
precipitation
needed
single
column?
up- & downdraft
needed
2D
enough?
• kinematic (prescribed-flow)
perhaps still enough? (focus on µ-physics)
• decoupled from cloud dynamics
computationally cheap

aerosol processing: minimal modelling framework
...
single
parcel?
precipitation
needed
single
column?
2D other...
up- & downdraft
needed
enough?
• kinematic (prescribed-flow)
perhaps still enough? (focus on µ-physics)
• decoupled from cloud dynamics
computationally cheap

aerosol processing: minimal modelling framework
...
single
parcel?
precipitation
needed
single
column?
2D other...
up- & downdraft
needed
enough?
• kinematic (prescribed-flow)
perhaps still enough? (focus on µ-physics)
• decoupled from cloud dynamics
computationally cheap

aerosol processing: minimal modelling framework
...
single
parcel?
precipitation
needed
single
column?
2D other...
up- & downdraft
needed
enough?
• kinematic (prescribed-flow)
perhaps still enough? (focus on µ-physics)
• decoupled from cloud dynamics
computationally cheap