Parameterization of precipitation in boundary layer clouds at the cloud system scale

choppedspleenΜηχανική

21 Φεβ 2014 (πριν από 3 χρόνια και 3 μήνες)

74 εμφανίσεις


Olivier Geoffroy




Parameterization of precipitation in boundary
layer clouds at the cloud system scale

Pier Siebesma, Roel Neggers

RK science lunch, 05/10/2010

Microphysical processes

w

CCN :

D ~ 0.01
-
10 µm

D

40 µm

n(D)

Cloud droplets :

~ 1 µm < D < ~ 40 µm

CCN
Activation

Microphysical processes

Condensation

w

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Cloud droplets :

~ 1 µm < D < ~ 40 µm

CCN
Activation

Microphysical processes

Condensation

w

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Cloud droplets :

~ 1 µm < D < ~ 40 µm

CCN
Activation

Mixing

D

40 µm

n(D)

Microphysical processes

Condensation

w

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Cloud droplets :

~ 1 µm < D < ~ 40 µm

CCN
Activation

Cloud droplet
sedimentation

Mixing

D

40 µm

n(D)

Microphysical processes

Cloud droplets :

~ 1 µm < D < ~ 40 µm

Condensation

w

precipitation embryo
D ~ 40 µm


Cloud droplet
sedimentation

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Collection

D

40 µm

n(D)

ECS

CCN
Activation

Collection : Efficient for D > 40 µm


(Bartlett, 1970)

Mixing

D

40 µm

n(D)

Microphysical processes

Cloud droplets :

~ 1 µm < D < ~ 40 µm

Condensation

precipitation embryo
D ~ 40 µm


Cloud droplet
sedimentation

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Collection

D

40 µm

n(D)

ECS

CCN
Activation

Collection : Efficient for D > 40 µm


(Bartlett, 1970)

Mixing

D

40 µm

n(D)

Polluted cloud

w


precipitation efficiency
(?)



LWP

(
?

(feedbacks)
)


(2
nd

aerosol indirect effect)

Microphysical processes

Condensation

w

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Cloud droplets :

~ 1 µm < D < ~ 40 µm

CCN
Activation

Cloud droplet
sedimentation

Mixing

D

40 µm

n(D)

Marine cloud

Microphysical processes

Cloud droplets :

~ 1 µm < D < ~ 40 µm

Condensation

w

precipitation embryo
D ~ 40 µm


Cloud droplet
sedimentation

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Collection

D

40 µm

n(D)

ECS

CCN
Activation

Collection : Efficient for D > 40 µm


(Bartlett, 1970)

Mixing

D

40 µm

n(D)

Microphysical processes

Condensation

w

Cloud droplet
sedimentation

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Collection

D

40 µm

n(D)

ECS

Rain drops
~ 40 µm <D < 100
-
500 µm

Cloud droplets :

~ 1 µm < D < ~ 40 µm

precipitation embryo
D ~ 40 µm


CCN
Activation

Mixing

D

40 µm

n(D)

Growth depends on the
available amount of water
i.e. H or LWP

Microphysical processes

Condensation

w

Rain drops
~ 40 µm <D < 100
-
500 µm

Cloud droplet
sedimentation

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Mixing

D

40 µm

n(D)

Collection

D

40 µm

n(D)

ECS

Cloud droplets :

~ 1 µm < D < ~ 40 µm

precipitation embryo
D ~ 40 µm


CCN
Activation

Rain
sedimentation


Microphysical processes

Condensation

Rain drops
~ 40 µm <D < 100
-
500 µm

Cloud droplet
sedimentation

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Mixing

D

40 µm

n(D)

Collection

D

40 µm

n(D)

ECS

Cloud droplets :

~ 1 µm < D < ~ 40 µm

precipitation embryo
D ~ 40 µm


CCN
Activation

Rain
sedimentation


size sorting


Rain
evaporation

Microphysical processes

Condensation

Rain drops
~ 40 µm <D < 100
-
500 µm

Cloud droplet
sedimentation

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Mixing

D

40 µm

n(D)

Collection

D

40 µm

n(D)

ECS

Cloud droplets :

~ 1 µm < D < ~ 40 µm

precipitation embryo
D ~ 40 µm


CCN
Activation

Rain
sedimentation


size sorting


Rain
evaporation

Microphysical processes

Condensation

Rain drops
~ 40 µm <D < 100
-
500 µm

Cloud droplet
sedimentation

CCN :

D ~ 0.01
-
10 µm

D

40 µm

D

40 µm

n(D)

n(D)

D
dt
dD
1

Mixing

D

40 µm

n(D)

Rain
sedimentation


size sorting


Collection

D

40 µm

n(D)

ECS

Cloud droplets :

~ 1 µm < D < ~ 40 µm

precipitation embryo
D ~ 40 µm


CCN
Activation

Rain
evaporation

Objective

-

Development of a precipitation scheme for boundary layer clouds at the
GCM scale


Precipitation in a key process in BLC evolution.





Low cloud regimes and transitions between regimes


Earth radiation budget, general circulation, hydrological cycle.




Quantification of the aerosol indirect effect.

explicit or bin

D

n(D)

Bulk

Cloud

rain

D

D
0


n(D)

Autoconversion

Accretion

Self
-
collection

Self
-
collection

2 bins


4 collection

processes

Collection processes :

Stochastic Collection Equation (SCE)

LES collection schemes

Cloud

rain

Cloud

:

q
c

(g kg
-
1
)


N
c

(cm
-
3
)


Rain

:

q
r

(g kg
-
1
)

N
r

(cm
-
3
)

Measured spectra

D
0


~ 40
-

100 µm

D
0


Cloud

rain

D (
μ
m)

Autoconversion

Cloud

:

q
c

(g kg
-
1
)


N
c

(cm
-
3
)


Rain

:

q
r

(g kg
-
1
)

N
r

(cm
-
3
)

Precipitation formation, autoconversion rate

Treshold
N
q
f
k
t
q
c
c
auto
auto
r










)
(
)
(
Khairoutdinov and

Kogan (2000)

2.47

-
1.79

Beheng (1994)

4.7

-
3.3

Seifert and Beheng

(2001)

4

-
2

Tripoli and Cotton

(2000)

2.33

-
0.33

α

β

Kessler (1969)

1

Treshold

H(q
c
-
q
treshold
)

Sundqvist (1978)

1

Liu and Daum (2006)

2.33

-
0.33

2
)
exp(
1
treshold
c
q
q

H(r
6
-
r
treshold
)

H(r
v
-
r
vtreshold
)

N
aerosol


Highly non linear

0

0

1

1

1

Auto rate

q
c

Aerosol indirect effect


dependance in N
c

necessary

N
c

N
c

Accretion

r
c
accr
accr
r
q
q
k
t
q



)
(

Depends on local values

Autoconversion / accretion rates

mean profiles (12H) in cumulus

Only in
cloud core

-

Formation of precipitation in
cloud core

-
Accretion =


~ 10 x autoconversion

-
Simulations show larger
accretion rate for w
up
-
v
qr

> 0

(drops go upward)

Only in
cloud core

auto

accr

w
up
-
v
qr

> 0

w
up
-
v
qr

< 0

v
qr

w
up

accr

)
)
exp(
1
(
0015
.
0
)
(
2
treshold
c
c
auto
r
q
q
q
t
q




Autoconversion:

Frac=cste

0 overlap, Sundqvist (1978) scheme

w=w
up
(k
-
1)

q
c
=l
up
(k)

1
)
(
)
(
)
(


















k
auto
r
c
c
up
auto
q
auto
c
w
z
t
q
q
z
q
w
z
F
t
q
c
=cste?

New scheme

b
c
a
c
auto
c
N
kq
t
q





)
(
Autoconversion:

Frac=cste

w=w
up
(k
-
1)

q
c
=l
up
(k)

c
up
qr
q
w
F

New scheme, accretion regime

Accretion:

Frac=cste

r
c
coll
qr
se
r
accr
r
q
q
k
z
F
t
q
t
q











0
)
(
)
(
dim
7
.
0
04
.
0
qr
r
F
q

(From RICO in situ
measurements)

3
.
0
/
1
3
.
0
1
)
)
(
(





k
c
k
F
z
kq
F
auto
F
New scheme,

o
verlap

Frac=cste

k+1

k

)
)/h
z
-
(z
-
(1
o
k
top

k

z
k

h
o

k
k
kclear






1
z
top

3
.
0
/
1
3
.
0
1
)
)
(
(




k
c
k
k
F
z
kq
F

?

Autoconversion formulation

param
r
t
q
)
(


core
auto
r
t
q




)
(
(kg kg
-
1

s
-
1
)

(kg kg
-
1

s
-
1
)

LES simulations (12H)

-

RICO, moister RICO

-

N
c
=
40, 50, 70, 100, 200

cm
-
3

-

Seifert and Beheng (2006)


scheme:



Identification of individual
clouds and average clouds
of same height

core
auto
r
t
q




)
(
core
c
q


power law hypothesis:



c
core
c
core
auto
c
N
q
a
t
q







)
(
0027
.
0
44
.
2
56
.
3




a


& regression

auto
LES

= f(auto
param
)

1/1 line

)
)
1
(
)
(
1
(
~
)
(
2
2
4
2006







c
c
SB
c
N
q
t
q
auto
3
7
.
0
7
.
0
)
1
(
~
)
(





Horizontal mean
values of:

r
c
c
q
q
q



1

SCM results

LWP, rain flux at surface,

N
c
=50, 70, 100, 200 cm
-
3

50 cm
-
3

70 cm
-
3

200 cm
-
3

No rain

100 cm
-
3

LWP:

precipitation
at surface:

60 g m
-
2

= 25 Wm
-
2

~ 0.4 mm h
-
1

~ 10 mm j
-
1


50 cm
-
3

70 cm
-
3

100 cm
-
3

200 cm
-
3

rain flux profiles

N
c
=50, 200 cm
-
3

N
c

= 50 cm
-
3

N
c

= 200 cm
-
3

Autoconversion time scale

w=5ms
-
1

w=1ms
-
1

w=w
up

w
z
t
q
q
z
q
w
z
F
t
q
auto
r
c
c
up
auto
q
auto
c
c

















)
(
)
(
)
(
w=5ms
-
1

w=1ms
-
1

w=w
up

Sensitivity to the overlap

h
0
=1000

h
0
=600

h
0
=300

LWP

N
c

= 50 cm
-
3

h
0
=300

h
0
=600

h
0
=1000

-

Developpment and implementation of a precipitation scheme in ECMWF
SCM model coupled with the DualM scheme.

-

Possibility to take in account the shear effect

-

Possible to take in account interaction between precipitation flux and the
stratiform component of the cloud (for Sc).


-

Dependency in N
CCN



Test of the scheme using the KPT and half a year of precipitation flux and
CCN concentration measurements:










Conclusion and perpective

North
Sea

origins

Dust
episode

CCN concentration at Cabauw during IMPACT (May 2008)

Regional background

Regional

background

Sundqvist no evap

SB, no evap, N
c

70

lwp

lwp

R
surf

R
surf

Sundqvist no evap

SB, no evap, N
c

50

lwp

lwp

R
surf

R
surf

SB, evap2, N
c

70

lwp

R
surf

R
surf

SB, evap2, N
c

200

lwp

R
surf

h
0
=300, 500, 800, 1500, SB, Nc 50, evap2

lwp

lwp

lwp

lwp

h
0
300 m

h
0
500 m

h
0
800 m

h
0
1500 m

SB, Nc 50, original evap

lwp

auto2, Nc 50, original evap

auto2, Nc 50, evap2

SB, evap2, N
c

50

lwp

R
surf

Nc 50 cm
-
3

Nc 70 cm
-
3

Nc 200 cm
-
3

lwp

lwp

lwp

auto 2 (a=2.74, b=
-
1.35)

Param accr up + auto

w
0
=5 ms
-
1

10e
-
6

a
qr
=1.85

b
Nc
=1.17

Param accr up + auto

w
0
=5 ms
-
1

2.5e
-
6

LWP, rain flux at surface,

N
c
=50, 70, 100, 200 cm
-
3

50 cm
-
3

70 cm
-
3

200 cm
-
3

No rain

100 cm
-
3

LWP:

precipitation
at surface:

60 g m
-
2

= 25 Wm
-
2

~ 0.4 mm h
-
1

~ 10 mm j
-
1


50 cm
-
3

70 cm
-
3

100 cm
-
3

200 cm
-
3

5 Wm
-
2

Steady state:

accr
auto
z
qr
w
v
z
F
sed
accr
auto
up
qr
qr












)
(
0
w
up
-
v
qr

< 0

w
up
-
v
qr

> 0