Clouds
and their radiative impact
as examples of histogram
(binning) methods
Brian Mapes
Global warming projections
in terms of T
Remember this from class 5?
• Climate heat budget over ocean + atm
– ∫ (ρCp dT/dt ) dV = ∫ (Frad_TOA) dA (+small)
• Units: Watts
• pert: = ∫ (-OLR’) dA + ∫ (ASR’) dA
– outgoing longwave and absorbed solar
Integrate over time (indefinite integral):
• ∫ {∫ (ρCp dT/dt ) dV} dt:
– units Joules
• or YottaJoules (10^22 = Yotta I think)
• Global warming due to increasing ASR (pdf)
Issue 1: integrating over area
• dA = (df) (cosf dl) in ∫ (Frad_TOA) dA
– weight by cos(f) when summing over lon bins
• OR: dA = (dsinf) (dl)
– Rebin latitude to equally spaced sinf bins
– Then ou can just sum them up!
–
• Related to map projection issue (equal area)
but that’s just for “eyeball” integrals
Equal area map projections
Radiative imbalance
• IPCC model ensemble (CMIP3)
Cumulative longwave trapping by
increasing GHGs
(clear sky = broken lines)
Effect is reduced somewhat by
clouds (total sky = solid/shaded)
Trenberth and Fasullo 2009 GRL
“Global warming due to increasing
absorbed solar radiation”
• All-sky mean longwave trapping quits by 2030
as skies clear (‘iris’ effect of clouds?)
2030
All
sky
Trenberth and Fasullo 2009 GRL
Global warming due to increasing
absorbed solar radiation
• From 2030, models warm largely by reduced
albedo (clearing skies/ cloud reductions?)
All
sky
2030
Trenberth and Fasullo 2009 GRL
Cloud cover reductions – where?
Non equal area
Yellow
overemphasized
in perception?
see
colorbrewer.org
Cloud radiative forcing
• “Stuff” (an additive scalar quantity):
– B&W best!
• Color is ambiguous among viewers
– Wm-2 units
• Area integration (or averaging) is what it’s all about
• Can be distributed over “bins”
– area bins matter (use sin(lat))
– but another dimension (like z) is free
2007 Cloud Radiative Effect CRE (aka CRF)
from CloudSat FLXHR product
19Wm-2 -55Wm-2
Ztop
(km)
19 LW global mean -55 SW
(Wm-2)
Caution: Simple average of 0130 and 1330 local
time samples, not true diurnal mean estimate!
Distributions: each ink molecule corresponds to
an equal amount of the Stuff (CRF)
Ztop
(km)
total LW
19
-55 Wm-2
19 Wm-2
-55 total
SW
LW
Decomposing CRE into cloud types
Lowest
possible base,
high top:
“Storms” vs. “layer clouds”
All else: layer
clouds
Decomposing the 19 and -55
8W -25W
11W -30W
Latitude distributions
-25W
Have CRE
impact
everywhere
8W
Impact at
high latitudes
-30W
(and equator) 11W
SW CRE: Storms
-16W out of -30W SW are
poleward of latitude 40 N/S
-30W SW CRE Mostly in local summer
G. Alaska, Kamchatka
-14W in
40S-40N
Cape Horn 56S
Day of year 2007
SW CRE by latitude and size
Summary
• Current clouds (cloudsat echo objects) have a
shortwave effect of -55 Wm-2 and longwave
effect of +19 Wm-2 according the (imperfect!
2xdaily) Cloudsat FLX-HR data set.
• These total impacts can be distributed over
latitude, cloud object size, season etc.
– gray scale: total impact a amount-of-ink-on-page.
Other ways of binning area
on the globe
• methods used also in
Bony et al.
method
• summarized
in wyant et
al.
these must be sin(lat) columns
in order for simple sum (2) to
be a true area x time average
Bony et al. method contd
• 30N-30S
Omega500 maps
collapse into a 1D PDF
scatterplots, and bin averages
of CRF
Stretch bins so that dx on the page
represents dA (area on globe)
• T-Tbar:
• RH:
cloud water
changes when
SST warms:
• good use of
color
• p is the proper
vertical coord
(mass)
• so the vertical
sum is
meaningful:
Decomposing changes into shift of
bins vs. changes in bin means