Some thoughts on s12 stratocumulus feedback

earsplittinggoodbeeInternet and Web Development

Nov 3, 2013 (3 years and 5 months ago)

64 views

© Crown copyright Met Office

Some thoughts on s12

stratocumulus feedback


Adrian Lock

EUCLIPSE WP3 meeting, Toulouse, April 2012

© Crown copyright Met Office

© Crown copyright Met Office

© Crown copyright Met Office

© Crown copyright Met Office

© Crown copyright Met Office

Subsidence pdfs:


HadGEM2 vs CGILS

s11

s12

Mean w
unchanged

© Crown copyright Met Office

Why do the LES robustly show a
significant positive feedback at s12 when
subsidence unchanged?

© Crown copyright Met Office

Why the positive feedback at s12 with
subsidence kept constant?


Robust changes with +2K:

1)
Cloud is warmer → greater upward LW flux at cloud top
→ more cloud
-
top cooling → more entrainment → thinner
cloud?


Actually find downward LW flux increases too and
balances almost exactly → cloud
-
top cooling unchanged

© Crown copyright Met Office

Why the positive feedback at s12 with
subsidence kept constant?


Robust changes with +2K:

1)
Cloud is warmer → greater upward LW flux at cloud top → more cloud
-
top cooling?


Actually find downward LW flux increases too and balances almost exactly →
cloud
-
top cooling unchanged

2)
Surface LH flux increases (assuming unchanged RH, U, T
1
-
T
surf
):


LHF = c
h
(q
1
-
q
sat
(T
surf
)) ~ c
h
(RH
1
q
sat
(T
1
)
-
q
sat
(T
surf
))

3)
Δ
q more negative (larger RH in the PBL dominates larger dq
sat
/dT
in warm free atmosphere):



(2)+(3) →
if entrainment rate unchanged,
expect stronger LH
fluxes throughout PBL:


LH flux

© Crown copyright Met Office

Impact of enhanced LH fluxes


Buoyancy flux =
α

w
θ
l

+
β

wq
T


wq
T

term is most important in the cloud layer (where

β

is
much larger)


Larger wq
T

→ larger buoyancy flux → more
turbulent PBL → more entrainment of dry air →
thinner cloud → positive cloud feedback


Do the steps of this argument hold up?

© Crown copyright Met Office

LH flux


entrainment feedback?


Larger wq
T

→ larger buoyancy flux → more turbulent PBL →
more entrainment of dry air → thinner cloud → positive cloud
feedback


“… → more entrainment of dry air…” should also → deeper
PBL (given identical subsidence)


but cloud
-
top drops → less entrainment!


© Crown copyright Met Office

LH flux


entrainment feedback?

MetO LEM fluxes after 3 days


LH flux still larger throughout PBL, despite lower
inversion (larger
Δ
q must dominate reduced w
e
)


Buoyancy flux weaker (consistent with reduced w
e
)


But why reduced w
e
,

to less than in the control?

© Crown copyright Met Office

Initial adjustment




Entrainment
is

larger
initially
(<first 6 hours), driven by larger
buoyancy flux, as expected


→ cloud thins (+ve feedback), as
expected


BUT cloud becomes
optically

thinner
in LW too
:

→ reduced LW cloud
-
top cooling

→ reduced buoyancy flux

→ reduced turbulence

→ reduced entrainment

→ cloud
-
top falls back (relative to
control)


© Crown copyright Met Office

Balanced state


So, despite enhanced LH fluxes at all levels,
SH flux is reduced because of reduced LW
cooling implying reduced buoyancy flux

Total

Turb

Rad

© Crown copyright Met Office

What if?


…the initial s12 cloud layer were deeper?


Same initial response (larger w
e
→ higher cloud
-
top
and thinner cloud layer)?


Noting that cloud depth has a strong control on the
integrated buoyancy flux, thinner cloud layer still
implies
w
e

reduces but not now to less than control?


So equilibrium state would still be a thinner cloud layer
(ie positive feedback) but with a higher cloud
-
top?


Chris Jones has tested this in a MLM…

© Crown copyright Met Office

Initial conditions don’t matter!

© Crown copyright Met Office

What if?


What if the circulation changes?


Eg, the shape of the w pdf changes?


HadGEM2 shows no change in the mean but less ascent and
more weak subsidence


If w>0 associated with small cloud fraction then this might
suggest a negative cloud feedback (but this doesn’t happen in
HadGEM2)


Something to try in time
-
varying CGILS…


Eg, the wind speed changes? (See Webb and Lock, soon!)

s12