1
Understanding black carbon radiative forcing uncertainty due to vertical distributions
2
3
4
5
6
7
8
9
10
11
12
13
14
Tentative list:
B.H.Samset, Myhre, G., R.B. Skeie, T.K. Berntsen (CICERO – OsloCTM2)
A. Kirkevåg, T.Iversen, Ø. Seland (met.no - CAM4-Oslo)
S. J. Ghan, X. Liu, R. C. Easter, P. J. Rasch, J.-H. Yoon (PNNL – CAM5.1)
S. Bauer ,K. Tsigaridis (GISS-ModelE)
T. Diehl, M. Chin (GOCART)
G. Lin, J. Penner (UMICH – IMPACT)
Y. Balkanski, M. Schulz (IPSL/met.no –INCA)
N. Bellouin (MetOffice -HadGEM2)
K. Zhang, P. Stier (MPIHAM)
J.-F. Lamarque (NCAR – NCAR-CAM3)
T. Takemura (SPRINTARS)
15
Unlike most atmospheric aerosols, black carbon (BC) absorbs solar radiation. This
16
warming effect of BC has led to suggestions for mitigation of BC emissions to limit
17
global warming both in the scientific community1-3 (+ UNEP rapport) and among policy
18
makers. The uncertainties in the radiative forcing (RF) of the direct aerosol effect of BC
19
are however large4,5. Among the causes of these uncertainties are differences in the
20
vertical profile of BC6. Here we show, based on a large set of global aerosol models, that
21
the diversity in the vertical profile of can contribute up to a third of the variability in
22
modeled BC RF. A significant fraction of this variability comes from high altitudes, as
23
we find that, globally, models simulate more than 40% of their total BC RF above 5km.
24
We isolate variations due to clouds, optical properties, surface albedo and radiative
25
transfer codes and adopt only the three-dimensional aerosol fields from the global
26
models. While BC concentrations quickly decrease with altitude, BC forcing sensitivity
27
increases due to the presence of water vapor, background aerosols and clouds. Cloud
28
fields and regional albedo differences are found to be of equal importance for the
29
variability, but both BC emission regions and areas with transported BC are found to be
30
important. These insights into the importance of the vertical profile of BC will benefit
31
the understanding of the differences among the global aerosol models, and may guide
32
the initiation of observational studies with the aim of improving the models in a focused
33
way.
34
The efficiency with which BC can induce radiative forcing is dependent on external factors
35
such as surface albedo, water vapor, background aerosol distributions and, most notably,
36
clouds (Haywood and Shine, 1997, Zarzycki and Bond, 2010). The total BC forcing is
37
therefore highly sensitive to the full 3D distribution of aerosol concentration (Samset and
38
Myhre, 2012). Anthropogenic BC forcing estimated by the current major global aerosol
39
models varies between 0.19 W/m2 and 0.38 W/m2 (Myhre et al, 2012, subm to Tellus?). This
40
range is little changed from previous estimates (Schulz et al, 2006, others?) Understanding the
41
causes of this variability in BC forcing would be a good step towards reducing the total
42
uncertainty on the total anthropogenic aerosol forcing.
43
Several studies have previously shown that the vertical transport is one area where the models
44
still differ significantly (Koffi et al, 2012, Schwarz et al, 2010). Using input from 11 global
45
aerosol models, we study the fraction of model variability in BC forcing attributable to
46
vertical differences alone. By combining the models’ own concentration profiles with a
47
common efficiency profile (EP) of RF per gram of BC (see Methods), we are able to
48
recalculate the induced RF of the BC direct aerosol effect at various altitudes and regions. We
49
can also separate out the contributions from the cloud field and variations due to regional
50
differences.
51
Figure 1a shows the global mean total BC burden from 11 global aerosol models participating
52
in AeroCom Phase 2, in addition to three models from AeroCom Phase 1. Individual numbers
53
are shown in Supplementary Table 1. We find a mean of 0.19 g/m2, but with a relative
54
standard deviation (RSD) of 32% and a model spread from 0.09 to 0.37 g/m2. Under an
55
assumption of equal forcing per gram, this alone will cause a large diversity in the total BC
56
forcing predicted by the models. AeroCom Phases 1 and 2 found a similar ranges, as shown
57
by the whisker boxes in Figure 1a. Note that the AeroCom P2 results are for BC from fossil
58
fuel burning (BCFF) only. Figure 1b shows the RF for each model, recalculated from the
59
concentration profiles using a common 3D, time dependent RF per gram profile. That the
60
average is stronger than for previous estimates is due to the strength of the profile used (see
61
Methods for discussion). The spread is highly correlated (95%) with the spread in burden,
62
again due to the use of a common BC EP. However, dividing the recalculated RF by the
63
global means gives new global estimates of BC RF per gram, shown in Figure 1c. If all model
64
vertical profiles were identical, this spread would vanish. It therefore carries information on
65
the impact of the vertical profiles of BC concentration on the model spread, separate from the
66
variability due to burden differences.
67
For AeroCom P1 and P2, the RSDs of the burden and RF per gram (quantifying the amount of
68
aerosol and the effect they have, respectively) are of approximately equal value, ranging from
69
32% to 46%. The two estimates are however internally correlated (Schulz 2006), such that the
70
spread in RF is less than if we simply combine the uncertainties. Our RSD in RF per gram is
71
16%, or 40% of that from AeroCom P1 and P2. Vetoing models in the P1 and P2 results that
72
have known issues (see Methods) brings this fraction up to 50%. Temporarily ignoring the
73
correlations, the vertical distributions can therefore be said to cause approximately 25% of the
74
spread in BC RF. Correlations are however mainly important between factors that are
75
independent of the vertical profile, and its importance is therefore likely to be higher than this
76
estimate. QUANTIFY IN METHODS, GIVE CONCLUSION.
77
As also shown in previous model comparisons (Shwarz 2010, more?), there are large
78
differences in the modeled BC vertical concentration profiles. Figure 2 shows the annual
79
mean vertical profiles of BC concentration from all models, globally and for two selected
80
regions (Arctic and China, see Methods for definitions). The Arctic is a region with no
81
indigenous BC emissions, but where the effects of transported aerosols are often discussed
82
(REF?). While all models show a significant BC contribution at high altitudes the model
83
spread in the Arctic is quite large, ranging from virtually constant below 200hPa to double-
84
peaked structures with maxima at 900 hPa and 200hPa. This highlights the differences in
85
transport schemes between the climate models. There is no discernible trend between the
86
AeroCom P1 and P2 models. China has heavy BC emissions at the surface, and also sees
87
contributions from transport at high altitudes, as is evident from Figure 2c.
88
Figure 2d-f shows the corresponding recalculated RF per meter of height, divided by the
89
model global mean burden. The resulting vertical profiles show at what altitudes RF is
90
generated, globally (2d) and for the two regions (2e-f). The interplay of the concentration
91
profiles, declining with altitude, and the RF per gram, increasing with altitude, is evident for
92
all regions. Globally, RF can be seen to be mostly induced in a range around 800hPa for most
93
models, with a secondary peak between 400hPa and 200hPa for some models. The China
94
region follows the global pattern, but with the low altitude peak closer to the ground,
95
reflecting the emissions there. The Arctic region, however, has a strong peak at high altitudes
96
for most models, evidencing the relative importance of high altitude BC RF there.
97
Figure 2g-i illustrate the model spread in forcing at a given altitude due to vertical profiles, by
98
integrating the absolute forcing per model layer from the surface and upwards. Globally, the
99
majority of models reach 50% of their forcing around 500hPa (approximately 5km), however
100
there is a large spread around this value. For the regions, this picture is quite different. In the
101
Arctic all models have a significant forcing component above 5km, while for the industrial
102
regions of China the opposite is true.
103
We find a significant fraction of modeled BC RF from high altitudes, with a notable regional
104
pattern. Figure 3 maps the model mean fraction of induced forcing (3a) and aerosol mass (3b)
105
from model layers of altitude above 500hPa. Firstly, the distinction between emission regions
106
of BC and regions that see mostly transported BC is evident. The former have small mass and
107
RF fractions at high altitudes, typically 10-20%, while transport regions show RF fractions up
108
to 80%. Figure 3c compares the model mean fractions and spreads globally and for three
109
regions (Arctic, China and Europe). Note that the fraction of RF above 5km is systematically
110
higher than the mass fraction, due to the strongly increasing shape of the RF efficiency
111
profiles. Globally, more than 40% of the model simulated RF from BC comes from model
112
layers above 5km, while only 24% of the mass is found in this region. This illustrates the
113
importance of validating model vertical profiles and transport codes not only in industrial
114
regions, but also in transport regions with low indigenous BC emissions.
115
Supplemental figure 1 shows the RF and mass fractions in four altitude bands, globally and
116
for the three regions. Such information is useful for comparing models with observational
117
data, and for constructing best estimates for global BC forcing.
118
We have shown that, even on global mean, a significant fraction of the variability in the BC
119
forcing per gram is due to differences in vertical profiles. It is however instructive to further
120
divide this variability into the components inherent in the efficiency profiles. In Figure 4a we
121
show the zonal mean RF per gram for all models, and investigate the relative importance of
122
the cloud field and of regional differences in albedo for the resulting spread. In Figure 4c we
123
have run the analysis using a global, annual mean efficiency profile instead of the full time
124
dependent 3D profile, and in addition used clear sky conditions. The solid line shows the
125
model mean forcing per gram, and the dashed lines show the maximum and minimum values
126
from individual models. The magnitude can be seen to be virtually independent of latitude,
127
due to the global profile masking the effects of varying surface albedo, and to have a low but
128
non-vanishing spread. As shown in Samset and Myhre 2011, not all vertical sensitivity of BC
129
forcing is due to the aerosols being above or below clouds. The remainder, caused
130
predominantly by Rayleigh scattering, water vapor and the competing effects of other aerosols,
131
is the cause of the variability here. RSD on global mean values is 10% (see Figure 4b). Figure
132
4d shows the same analysis using the full 3D efficiency profile. The RSD increases to 13%,
133
and we now see the latitudinal effects of the high polar albedo. Figure 4e shows the analysis
134
using the global mean efficiency profile again, but under all-sky conditions. Again the RSD
135
increases to 13%, due to the cloud field. The model variability in forcing efficiency that is due
136
to vertical profile differences can therefore be decomposed into three factors: Equal and
137
significant contributions from the cloud field and from and from regional differences, and a
138
major contribution from the underlying sensitivity of BC forcing to altitude even in the
139
absence of clouds and albedo differences. Harmonizing model treatment of clouds and albedo
140
is therefore not sufficient to remove uncertainties in BC forcing due to vertical profiles.
141
Results in the present study are given for total BC aerosol emissions only. However, due to
142
different source regions, the vertical profiles of BCFF could be different from BC. Six models
143
also provided concentration profiles for BCFF. Using the same efficiency profiles we
144
performed the analysis also for BCFF, and found results consistent with what we have
145
presented for BC (not shown). While the absolute forcing numbers differ due to lower total
146
burdens for BCFF, the variability in vertical profiles is very similar to that for total BC. Our
147
conclusions here are therefore also applicable to model comparison results on BCFF only.
148
In conclusion, we have shown that the previously documented large spread in BC aerosol
149
concentration profiles is enhanced for vertical BC RF profiles. Using 11 global aerosol
150
models, we show that most models globally simulate 40% of their BC forcing above 5km, and
151
that regionally the fraction can be above 70%. The spread between models is however quite
152
large, and we have indications that up to a third of the differences in modeled BC RF can be
153
attributed to differences in vertical profiles. To propose efficient mitigation measures for BC,
154
its radiative forcing needs to be well understood both in emission and transport regions.
155
Further model improvements and comparisons with data are needed, and should focus on both
156
of these types of region. It is however clear that while the BC vertical profiles are important,
157
they are not sufficient to explain the remaining differences between global aerosol models.
158
Methods
159
Model simulations
160
The AeroCom initiative asks models to simulate the direct aerosol radiative effect using the
161
same base years (2000 and 1850), with the same meteorology and microphysics. Two sets of
162
comparisons have been performed, here labeled AeroCom Phase 1 (P1 ref) and AeroCom
163
Phase 2 (P2 ref). For the present study, monthly 3D BC concentration fields from 11 global
164
aerosol models participating in AeroCom Phase 2 are used. Supplementary table 1 lists the
165
models, descriptions can be found in Myhre et al 2012 and references therein. For six of the
166
models, concentration fields of BC from fossil fuel burning only (BCFF) are also used. Three
167
of the same models also provided BC concentration profiles to AeroCom Phase 1. These
168
fields are also included here, as separate models since the all three have seen heavy
169
development in the intervening years. The concentration fields and the 3D model resolutions
170
are the only inputs taken from each separate model. The 1850 concentrations are subtracted
171
from the 2000 ones, leaving the contribution due to anthropogenic BC or BCFF emissions.
172
RF recalculation
173
Samset and Myhre 2011 presents a set of vertical profiles of the direct radiative forcing per
174
gram of aerosol for BCFF, i.e. the amount of top-of-atmosphere shortwave radiative forcing
175
seen by a particular model (OsloCTM2) per gram of aerosols at a given altitude. Here we term
176
these efficiency profiles (EP). In the specialized literature, the term normalized radiative
177
forcing is more commonly used. Here we employ the full time dependent 3D efficiency
178
profiles, either for all sky or clear sky conditions as described above. We find the forcing per
179
gram BC for each model grid point and time step, and then calculate the contribution to the
180
shortwave, top-of-atmosphere BC radiative forcing by multiplying in the modeled BC
181
concentration. This gives us intercomparable 3D RF fields, something that is not immediately
182
available from each model. We label this the recalculated RF, to emphasize that it is heavily
183
correlated with the burden and should not be taken directly as an estimate of BC forcing of
184
each model. Since the model that was used to produce the profiles of RF per gram has among
185
the strongest global mean forcings per gram (Myhre et al 2012), most recalculated RFs can be
186
expected to be stronger than their host model would predict. However, by dividing the
187
recalculated RF by the total aerosol burden of the host model, we extract the variability in
188
forcing per gram from each model that is due only to vertical profiles, and can subsequently
189
use the vertical RF profiles to study it.
190
As this method removes contributions to the spread in BC RF due to differences in model
191
specific treatment of clouds, water uptake and microphysics, we are left solely with variations
192
due to the concentration profiles and the total aerosol load. Dividing by the load, we can study
193
the contribution to the variability due to regional differences (high or low surface albedo,
194
indigenous emissions, transport region or both), and due to the presence of clouds. To study
195
the latter effects, we ran parallel analyses using only a global mean profile of RF per gram
196
(labeled GP above), a distinct profile of RF per gram for clear sky conditions (labeled CS), or
197
both (labeled CSGP).
198
Region definitions
199
For the present analysis, “Europe”, defined as the box of longitudes -10 to 30, latitudes 38 to
200
60, “China” which covers longitudes 105 to 135, latitudes 15 to 45, and “Arctic” which
201
covers latitudes 70 to 90. Europe and China represent regions with high industrial BC
202
emissions and significant fractions of the global BCFF emissions, making their total BC
203
forcing sensitive to pure vertical transport and wet scavenging. The Arctic is a region with
204
low indigenous emissions, but important BC contributions at high altitudes transported from
205
other regions. Its total BC forcing is therefore sensitive to model differences in vertical and
206
long range transport.
207
208
References
209
210
211
212
213
214
215
216
217
218
219
220
221
1
222
223
2
3
4
5
6
Grieshop, A. P., Reynolds, C. C. O., Kandlikar, M., and Dowlatabadi, H., A blackcarbon mitigation wedge. Nature Geoscience 2 (8), 533 (2009).
Hansen, J. et al., Global warming in the twenty-first century: An alternative scenario.
Proceedings of The National Academy of Sciences of The United States of America 97
(18), 9875 (2000).
Nature_Editorial, Time for early action. Nature 460 (7251), 12 (2009).
Ramanathan, V. and Carmichael, G., Global and regional climate changes due to black
carbon. Nature Geoscience 1 (4), 221 (2008).
Schulz, M. et al., Radiative forcing by aerosols as derived from the AeroCom presentday and pre-industrial simulations. Atmospheric Chemistry and Physics 6, 5225 (2006).
Zarzycki, C. M. and Bond, T. C., How much can the vertical distribution of black
carbon affect its global direct radiative forcing? Geophysical Research Letters 37,
L20807 (2010).
224
Tables
225
226
227
228
Table 1: Modeled BC burden, RF calculated by use of full 3D efficiency profiles (RF) or by a global mean profile (RFGP),
and forcing per gram (NRF). All numbers shown for global mean and for three selected regions. RF_fraction shows the
fraction of the total BC forcing simulated within the stated region. M>5km and RF>5km show the fractions of aerosol
mass and RF, respectively, simulated above an altitude of 5km (500hPa).
Global
Model
NCAR-CAM3.5
CAM4-Oslo
CAM5.1
GISS-modelE
GOCART-v4
HadGEM2
IMPACT
INCA
MPIHAM
OsloCTM2
SPRINTARS
CAM4-Oslo-P1
IMPACT-P1
OsloCTM2-P1
Mean
Std.dev
Europe
Model
NCAR-CAM3.5
CAM4-Oslo
CAM5.1
GISS-modelE
GOCART-v4
HadGEM2
IMPACT
INCA
MPIHAM
OsloCTM2
SPRINTARS
CAM4-Oslo-P1
IMPACT-P1
OsloCTM2-P1
Mean
Std.dev
229
230
Burden
[mg/m2]
0.14
0.22
0.09
0.21
0.19
0.37
0.17
0.23
0.18
0.20
0.19
0.17
0.19
0.19
0.19
0.06
RF
[W/m2]
0.25
0.54
0.16
0.40
0.38
0.81
0.24
0.41
0.26
0.40
0.37
0.34
0.30
0.27
0.37
0.16
NRF
[W/g]
1801
2445
1655
1905
1982
2185
1450
1833
1492
2003
1959
1955
1546
1477
1835
290
RFGP
[W/g]
0.27
0.57
0.18
0.44
0.41
0.86
0.28
0.45
0.31
0.44
0.41
0.36
0.34
0.33
0.40
0.16
RF fraction
[%]
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
0
Mass>5km
[%]
20.3
46.0
18.1
30.1
27.1
33.6
5.8
28.9
10.8
30.1
30.3
26.4
14.2
11.7
23.8
10.8
RF>5km
[%]
39.3
64.7
36.4
56.4
46.3
50.6
13.1
53.8
25.7
48.3
54.2
46.9
30.7
24.2
42.2
14.5
Burden
[mg/m2]
0.20
0.36
0.21
0.62
0.30
0.58
0.36
0.45
0.39
0.30
0.23
0.36
0.50
0.38
0.37
0.13
RF
[W/m2]
0.27
0.73
0.25
0.79
0.48
1.09
0.37
0.54
0.41
0.47
0.38
0.52
0.54
0.45
0.52
0.22
NRF
[W/g]
1386
2029
1230
1276
1626
1864
1036
1206
1052
1557
1658
1425
1081
1189
1401
308
RFGP
[W/g]
0.32
0.80
0.30
0.91
0.54
1.20
0.45
0.63
0.47
0.54
0.43
0.59
0.63
0.53
0.60
0.24
RF fraction
[%]
1.15
1.41
1.70
2.03
1.34
1.40
1.61
1.37
1.62
1.23
1.08
1.62
1.87
1.71
1.51
0.27
Mass>5km
[%]
16.2
36.2
9.1
16.5
21.6
27.3
4.3
17.4
5.8
19.5
27.9
16.5
8.5
5.9
16.6
9.5
RF>5km
[%]
40.8
60.1
25.0
43.6
44.0
47.4
13.0
49.6
19.4
41.0
58.7
38.5
24.9
16.0
37.3
15.2
Arctic
Model
NCAR-CAM3.5
CAM4-Oslo
CAM5.1
GISS-modelE
GOCART-v4
HadGEM2
IMPACT
INCA
MPIHAM
OsloCTM2
SPRINTARS
CAM4-Oslo-P1
IMPACT-P1
OsloCTM2-P1
Mean
Std.dev
China
Model
NCAR-CAM3.5
CAM4-Oslo
CAM5.1
GISS-modelE
GOCART-v4
HadGEM2
IMPACT
INCA
MPIHAM
OsloCTM2
SPRINTARS
CAM4-Oslo-P1
IMPACT-P1
OsloCTM2-P1
Mean
Std.dev
Burden
[mg/m2]
0.05
0.20
0.02
0.16
0.14
0.34
0.05
0.07
0.03
0.07
0.08
0.12
0.11
0.02
0.10
0.09
RF
[W/m2]
0.20
0.79
0.07
0.64
0.52
1.19
0.16
0.29
0.11
0.27
0.34
0.46
0.35
0.07
0.39
0.31
NRF
[W/g]
3907
3877
4187
3889
3746
3465
3214
4016
4078
3879
4413
3716
3266
3637
3806
335
RFGP
[W/g]
0.16
0.60
0.06
0.49
0.38
0.86
0.11
0.24
0.08
0.21
0.26
0.34
0.26
0.05
0.29
0.23
RF fraction
[%]
2.47
4.41
1.41
4.74
4.15
4.40
1.95
2.12
1.21
2.03
2.77
4.10
3.48
0.74
2.85
1.34
Mass>5km
[%]
63.2
60.9
81.6
65.7
53.2
40.3
35.9
78.8
75.5
71.0
83.1
48.9
39.4
63.0
61.5
15.9
RF>5km
[%]
76.1
71.0
88.5
77.2
64.1
52.1
48.1
89.2
89.2
82.4
89.9
61.1
52.9
80.4
73.0
15.0
Burden
[mg/m2]
0.89
0.84
0.65
1.60
1.12
2.08
0.83
1.06
1.31
1.27
1.33
0.67
0.96
0.84
1.10
0.39
RF
[W/m2]
1.17
1.35
0.74
1.89
1.50
3.29
0.94
1.27
1.39
1.95
1.66
0.89
0.87
1.00
1.42
0.66
NRF
[W/g]
1312
1612
1139
1185
1340
1581
1138
1198
1061
1544
1248
1336
904
1189
1270
202
RFGP
[W/g]
1.42
1.55
0.93
2.43
1.82
3.82
1.21
1.53
1.80
2.26
2.04
1.06
1.20
1.27
1.74
0.75
RF fraction
[%]
8.88
4.66
8.87
8.78
7.46
7.61
7.32
5.70
9.85
9.17
8.30
4.96
5.41
6.82
7.41
1.69
Mass>5km
[%]
9.7
22.7
8.1
10.4
11.3
15.9
3.3
13.7
2.8
15.4
10.2
10.5
5.6
6.6
10.4
5.3
RF>5km
[%]
25.0
48.3
24.4
30.5
28.4
34.2
9.2
40.8
9.1
33.0
28.3
27.4
19.8
18.1
26.9
10.9
231
Figures
232
233
234
235
Figure 1: Modeled BC global mean (a) burden, (b) RF and (c) RF per gram BC. Yellow boxes indicate mean, one standard
deviation and max/min values. Mean values and spreads for AeroCom P1 and P2 (hatched whisker boxes) are taken from
Schulz et al 2006 and Myhre et al 2012 respectively.
236
237
238
239
240
Figure 2: (a-c) BC concentration vertical profiles, global mean and for two selected regions. Overlain is the annual mean
forcing efficiency profile for the selected region. Solid lines show AeroCom P2 submissions, dashed lines show P1. (d-f)
BC RF per height, divided by the modeled global mean BC burden, globally and for three selected regions. (g-i) Vertical
profile of integrated absolute BC RF. Lines indicate the 50% mark and 500hPa altitude.
241
242
243
244
245
246
Figure 3: (a) Fraction of modeled BC RF above 5km. (b) Fraction of modeled BC mass above 5km. (c) Mean fraction of
mass (orange) and RF (yellow) globally and for three selected regions. Boxes indicate one standard deviation on the
model spread, whiskers show maximum and minimum values.
247
248
249
250
251
252
253
254
Figure 4: Breakdown of spread in BC RF per gram. (a) Zonal mean of BC RF per gram using 3D efficiency profiles and all
sky conditions, for all models. (c-e) Model mean (solid line) and maximum/minimum (dashed lines) when instead using
(c) a global, annual mean efficiency profile and clear sky conditions, (d) 3D profile and clear sky conditions, and (e) a
global profile and all sky conditions. (b) shows the global mean values for the four cases. Boxes show one standard
deviation on the model spread, whiskers show maximum/minimum values.
255
256
257
258
Supplementary figure 1: Mass and forcing fractions in four altitude bands, globally and for three regions. Yellow boxes
show mass fractions, orange boxes show RF fractions. (a) Global mean, (b) Arctic, (c) Europe, (d) China.