4
C MIP
in CMIP5
Pierre Friedlingstein,
Alessandro Anav (UExeter),
Chris Jones and Ben Booth (MetOffice)
Turning to the CMIP5 long-term experiments, Fig. 3 shows the set of core experiments
that include AMIP runs, a coupled control run and at least one 20th century experiment
with all forcings (also referred to here as an “historical” run). There are two projection
simulations forced with specified concentrations consistent with a high emissions
scenario (RCP8.5) and a medium mitigation scenario (RCP4.5). For AOGCMs that have
been coupled to a carbon cycle model (subsequently referred to as earth system models
or ESMs), there are control, 20th century simulations, and a future simulation with the
high scenario (RCP8.5) driven by emissions.
CMIP5 Protocol • 
• 
• 
• 
Figure 3: Schematic summary of CMIP5 long-term experiments. Green font indicates
simulations that will be performed only by models with carbon cycle representation.
Historical runs (C-­‐driven) RCPs (C-­‐driven) CMIP1% runs Hist+RCP8.5 (E-­‐driven) ESMs* involved
"   CanESM2
"   HadGEM2-ES
"   IPSL-CM5
"   BCC
"   INMCM4
"   NorESM-1
"   + MIROC ?
* = with carbon cycle
Historical runs evalua?on •  LAI –  annual mean and seasonal cycle –  evaluated against MODIS LAI •  GPP –  annual mean and seasonal cycle –  “evaluated” against gridded FluxNET product •  C uptake –  decadal mean and IAV –  Evaluated against atmospheric inversions Leaf Area Index LAI seasonal cycle Leaf Area Index GLOBE
SOUTHERN HEMISPHERE
e l a t i on Coe f f i c i ent
Co r r
0
ï0.2
0.2
el
Co r r
ï0.2
ï0.4
0.4
a t i on Coe f f i c i en
t
0
0.2
ï0.4
0.4
ï0.6
0.6
0.
0.
5
0.6
4
ï0.6
0.8
ï0.8
0.8
ï0.95
S
1
ï1
0.99
R
ï0.99
M
S
0.99
R
M
ï0.99
0.95
D
0.95
D
ï0.95
0.
0.
2
25
ï0.8
ï1
0
0.2
1
0.4
0
Standard deviation
0.25
0.5
Standard deviation
MODIS
CanESM2
HadGEM2
INMCM4
IPSLïCM5A
NORTHERN HEMISPHERE
TROPICS (30Sï30N)
BCCïCSM1
e l a t i on Coe f f i c i ent
Co r r
0
ï0.2
0.2
el
Co r r
ï0.2
NorESM1
ï0.4
0.4
a t i on Coe f f i c i en
t
0
0.2
ï0.4
0.4
ï0.6
0.6
0.
0.
2
0.6
5
ï0.6
0.8
ï0.8
0.8
0.
0.
1
25
ï0.8
ï0.95
S
1
ï1
ï1
0
Standard deviation
0.99
R
ï0.99
M
S
0.99
R
M
ï0.99
0.95
D
0.95
D
ï0.95
0.25
0.5
1
0
Standard deviation
0.1
0.2
Historical runs Global land and ocean sink 4
IPCC AR4 1990’s average 2
SINK
1.5
Land Carbon (PgC)
Land Carbon (PgC)
3
2
1
0
ï1
0.9 to 4.3 1
0.5
0
ï0.5
SOURCE
ï2
ï1
4
2.5
0
0.005
0.01
0.015
0.02
SINK
3
2.2 ± 0.4 Ocean Carbon (PgC)
Ocean Carbon (PgC)
2
2
1
0
1.5
1
ï1
SOURCE
ï2
1850
1865
1880
1895
1910
1925
YEAR
1940
1955
1970
1985
2000
0.5
2
4
6
8
PDF Frequency
10
12
10
ï3
Par??oning of C sinks GLOBE
4
2
PgC
0
ï2
ï4
ï6
ï8
Inversion
CanESM2
HadGEM2
INMCM4
IPSLïCM5A NorESM1
Inversion
CanESM2
INMCM4
IPSLïCM5A BCCïCSM1
Inversion
CanESM2
INMCM4
IPSLïCM5A BCCïCSM1
Inversion
CanESM2
INMCM4
IPSLïCM5A BCCïCSM1
Inversion
CanESM2
INMCM4
IPSLïCM5A BCCïCSM1
SOUTHERN HEMISPHERE
2
PgC
0
ï2
ï4
Inversion
CanESM2
HadGEM2
INMCM4
IPSLïCM5A NorESM1
NORTHERN HEMISPHERE
4
PgC
2
0
ï2
ï4
Inversion
CanESM2
HadGEM2
INMCM4
IPSLïCM5A NorESM1
TROPICS (30Sï30N)
4
PgC
2
0
ï2
Inversion
CanESM2
HadGEM2
INMCM4
IPSLïCM5A NorESM1
Zero Order Draft
RCPs Chapter 6
IPCC WGI Fifth Assessment Report
Comment [c12]: Che
flux panel is correct – look
v.big model spread??
1
14
15
16
17
18
19
20
21
22
23
fuel emission from 2005 to 2100.
[INSERT FIGURE 6.26 HERE]
Figure 6.26: compatible fossil fuel emissions simulated by the CMIP5 models for the 4 RCP scenarios. Top:
timeseries of instantaneous emission rate. Thick lines represent the historical estimates and RCP scenarios,
think lines show results from CMIP5 ESMs. Bottom: cumulative emissions for the historical period (18602005) and 21st century (defined in CMIP5 as 2005-2100) for historical estimates and RCP scenarios (bars)
and ESMs (symbols).
RCPs 2
3
4
5
6
7
8
Figure 6.25: changes in airborne, land and ocean fraction of fossil fuel carbon emissions. The figure shows 3
axes whose sum is always unity – airborne fraction increases vertically, land fraction from top left to bottom
right, and ocean fraction from top right to bottom left. Solid circles show model simulations for the 1990s,
and the open circle shows the observed estimate based on table 6.2. The coloured lines and symbols denote
the change in uptake fractions under the different RCP scenarios for each model, calculated using the
cumulative change in carbon from 2005-2100.
Change in airborne frac?on 1 9
Comment [Chris
1 Comment [Chris
0 0 0 Land-­‐borne frac?on Comment [Chris
1 CMIP 1% feedback analysis Full BGC impact of climate change Control (stable) Rad 8
9
10
11
12
13
14
15
16
17
18
19
and the differences between models. Whilst ! (especially on land) has been identified as the largest
contributor to model spread in the gain factor, g, " (land) is the largest contributor to model spread in future
CO2 concentration and # the largest contributor to spread in future temperature. Model spread in land
response is greater than for ocean, but no single process or region dominates the total uncertainty with the
most important process depending on the quantity of interest.
CMIP 1% feedback analysis [INSERT FIGURE 6.20 HERE]
Figure 6.20: Comparison of carbon cycle feedback metrics between C4MIP and CMIP5 (could also add
perturbed Hadley ensemble results too).
Comment [c6]: Define b
1 sigma
Anthrop. CO2 Emissions (PgC/yr) Glob. Surt Temp (K) Atm CO2 (ppmv) E-­‐driven simula?ons Historical and RCP8.5 E-­‐driven C-­‐driven E-­‐driven Status of the MIP; What science have we learned so far? Need for cross-­‐MIPs analyses? Plans for an introduc?on paper; Plans for worshops and ar?cula?on with the CMIP5 workshop Status of the MIP; see above What science have we learned so far? Need for cross-­‐MIPs analyses? Plans for an introduc?on paper; Plans for worshops and ar?cula?on with the CMIP5 workshop Status of the MIP; What science have we learned so far? To be honest, not much, but I’m s?ll op?mis?c Need for cross-­‐MIPs analyses? Plans for an introduc?on paper; Plans for worshops and ar?cula?on with the CMIP5 workshop Status of the MIP; What science have we learned so far? Need for cross-­‐MIPs analyses? Yes: -­‐ cross feedback analysis (as in Gregory et al. 2009) -­‐ cross uncertainty analysis -­‐ cross evalua?on -­‐ PMIP + C4MIP = PCMIP Plans for an introduc?on paper; Plans for worshops and ar?cula?on with the CMIP5 workshop Status of the MIP; What science have we learned so far? Need for cross-­‐MIPs analyses? Plans for an introduc?on paper; Yes, special issue planned in Jclim (see hgp://ar5carbon.wikispaces.com/) Introduc?on paper had been discussed Plans for worshops and ar?cula?on with the CMIP5 workshop Status of the MIP; What science have we learned so far? Need for cross-­‐MIPs analyses? Plans for an introduc?on paper; Plans for worshops and ar?cula?on with the CMIP5 workshop No plans for a C4MIP only workshop, but clear inten?on to contribute to the CMIP5 workshop