TSSM-1

TSSM

This supplementary material should be cited as:

Stocker, T.F., D. Qin, G.-K. Plattner, L.V. Alexander, S.K. Allen, N.L. Bindoff, F.-M. Bréon, J.A. Church, U. Cubasch,

S. Emori, P. Forster, P. Friedlingstein, N. Gillett, J.M. Gregory, D.L. Hartmann, E. Jansen, B. Kirtman, R. Knutti, K.

Krishna Kumar, P. Lemke, J. Marotzke, V. Masson-Delmotte, G.A. Meehl, I.I. Mokhov, S. Piao, V. Ramaswamy, D.

Randall, M. Rhein, M. Rojas, C. Sabine, D. Shindell, L.D. Talley, D.G. Vaughan and S.-P. Xie, 2013: Technical Sum-

mary Supplementary Material. In: Climate Change 2013: The Physical Science Basis. Contribution of Working

Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin,

G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Available from

www.climatechange2013.org and www.ipcc.ch.

Coordinating Lead Authors:

Thomas F. Stocker (Switzerland), Qin Dahe (China), Gian-Kasper Plattner (Switzerland)

Lead Authors:

Lisa V. Alexander (Australia), Simon K. Allen (Switzerland/New Zealand), Nathaniel L. Bindoff

(Australia), François-Marie Bréon (France), John A. Church (Australia), Ulrich Cubasch

(Germany), Seita Emori (Japan), Piers Forster (UK), Pierre Friedlingstein (UK/Belgium), Nathan

Gillett (Canada), Jonathan M. Gregory (UK), Dennis L. Hartmann (USA), Eystein Jansen (Norway),

Ben Kirtman (USA), Reto Knutti (Switzerland), Krishna Kumar Kanikicharla (India), Peter Lemke

(Germany), Jochem Marotzke (Germany), Valérie Masson-Delmotte (France), Gerald A. Meehl

(USA), Igor I. Mokhov (Russian Federation), Shilong Piao (China), Venkatachalam Ramaswamy

(USA), David Randall (USA), Monika Rhein (Germany), Maisa Rojas (Chile), Christopher Sabine

(USA), Drew Shindell (USA), Lynne D. Talley (USA), David G. Vaughan (UK), Shang-Ping Xie

(USA)

Contributing Authors:

Myles R. Allen (UK), Olivier Boucher (France), Don Chambers (USA), Jens Hesselbjerg

Christensen (Denmark), Philippe Ciais (France), Peter U. Clark (USA), Matthew Collins (UK),

Joseﬁno C. Comiso (USA), Viviane Vasconcellos de Menezes (Australia/Brazil), Richard A. Feely

(USA), Thierry Fichefet (Belgium), Gregory Flato (Canada), Jesús Fidel González Rouco (Spain),

Ed Hawkins (UK), Paul J. Hezel (Belgium/USA), Gregory C. Johnson (USA), Simon A. Josey (UK),

Georg Kaser (Austria/Italy), Albert M.G. Klein Tank (Netherlands), Janina Körper (Germany),

Gunnar Myhre (Norway), Timothy Osborn (UK), Scott B. Power (Australia), Stephen R. Rintoul

(Australia), Joeri Rogelj (Switzerland/Belgium), Matilde Rusticucci (Argentina), Michael Schulz

(Germany), Jan Sedláček (Switzerland), Peter A. Stott (UK), Rowan Sutton (UK), Peter W. Thorne

(USA/Norway/UK), Donald Wuebbles (USA)

Review Editors:

Sylvie Joussaume (France), Joyce Penner (USA), Fredolin Tangang (Malaysia)

Technical Summary

Supplementary Material

TSSM

TSSM-2

Table of Contents

TS.SM.1 Notes and Technical Details on

Observed Global Surface Temperature

Figures in the Summary for

Policymakers – Figure SPM.1 ................... TS-SM-3

TS.SM.2 Notes and Technical Details on

Observed Change in Precipitation

Over Land Figures in the Summary

for Policymakers – Figure SPM.2 ............ TS-SM-3

TS.SM.3 Notes and Technical Details on Observed

Indicators of a Changing Global

Climate Figures for the Summary

for Policymakers – Figure SPM.3 ............ TS-SM-3

TS.SM.4 Notes and Technical Details on

Observed Changes in the Global

Carbon Cycle Figures in the Summary

for Policymakers – Figure SPM.4 ............ TS-SM-5

TS.SM.5 Notes and Technical Details on

Radiative Forcing Estimates Figure in

the Summary for Policy Makers –

Figure SPM.5 ................................................. TS-SM-6

TS.SM.6 Notes and Technical Details on Comparison

of Observed and Simulated Climate

Change Figures for the Summary

for Policymakers – Figure SPM.6 ............ TS-SM-6

TS.SM.7 Notes and Technical Details on CMIP5

Simulated Time Series Figures in the

Summary for Policymakers –

Figure SPM.7 ................................................. TS-SM-7

TS.SM.8 Notes and Technical Details on Maps

Showing CMIP5 Results in the

Summary for Policymakers –

Figure SPM.8 ............................................... TS-SM-11

TS.SM.9 Notes and Technical Details on the Sea Level

Projection Figure for the Summary

for Policymakers – Figure SPM.9 .......... TS-SM-15

TS.SM.10 Notes and Technical Details on the Summary

for Policymakers Figure Plotting Global

Mean Temperature Increase as a Function

of Cumulative Total Global CO

Emissions – Figure SPM.10 ..................... TS-SM-15

References ......................................................................... TS-SM-17

TSSM

Technical Summary Supplementary Material

TSSM-3

TS.SM.1 Notes and Technical Details on Observed

Global Surface Temperature Figures in

the Summary for Policymakers –

Figure SPM.1

Data and programming code (IDL) used to create Summary for Policy-

makers and Technical Summary ﬁgures originating from Sections 2.4

and 2.5 of Chapter 2 can be obtained from the IPCC WGI AR5 website

www.climatechange2013.org.

TS.SM.1.1 Annual and Decadal Global Surface

Temperature Anomalies – Figure SPM.1a

Global Mean Surface Temperature (GMST) anomalies as provided by

the dataset producers are given normalized relative to a 1961–1990

climatology from the latest version (as at 15 March 2013) of three

combined Land-Surface Air Temperature (LSAT) and Sea Surface Tem-

perature (SST) datasets. These combined datasets and the correspond-

ing colours used in Figure SPM.1a are:

HadCRUT4 (version 4.1.1.0) – black

NASA GISS – blue

NCDC MLOST (version 3.5.2) – orange.

An overview of methodological diversity between these three temper-

ature datasets is provided in Table 2.SM.6 of the Supplementary Mate-

rial to Chapter 2, and full comprehensive details on the construction

process for these datasets are provided in the references cited in this

table. For time-series of LSAT only, and SST only, the reader is referred

to Figure TS.1.

For the decadal anomalies, 90% conﬁdence intervals are shown for the

HadCRUT4 dataset (based on Morice et al., 2012).

TS.SM.1.2 Maps of Observed Changes in Surface

Temperature – Figure SPM.1b

Maps of observed changes in surface temperature are based on trends

calculated from the 3 datasets listed above for the period 1901–2012.

See the Supplementary Material of Chapter 2 for a detailed description

of the methodology used for trend and uncertainty calculations (Sec-

tion 2.SM.3.3). Trends have been calculated only for those grid boxes

with greater than 70% complete records and more than 20% data

availability in the ﬁrst and last 10% of the time period. White areas

indicate incomplete or missing data. Black plus signs (+) indicate grid

boxes where trends are signiﬁcant at the 2-tailed 10% signiﬁcance

level (i.e., a trend of zero lies outside the 90% conﬁdence interval).

The Technical Summary provides maps for all 3 datasets (Figure TS.2),

while the Summary for Policymakers provides a map based on NCDC

MLOST only (Figure SPM.1b).

TS.SM.2 Notes and Technical Details on

Observed Change in Precipitation

Over Land Figures in the Summary

for Policymakers – Figure SPM.2

Data and programming code (IDL) used to create Summary for Policy-

makers and Technical Summary ﬁgures originating from Sections 2.4

and 2.5 of Chapter 2 can be obtained from the IPCC WGI AR5 website

www.climatechange2013.org.

TS.SM.2.1 Map of Observed Changes in Precipitation

Over Land – Figure SPM.2

Maps of observed changes in annual precipitation over land show

trends calculated from 3 datasets:

CRU TS 3.10.01 (updated from Mitchell and Jones, 2005)

GHCN V2 (updated through 2011; Vose et al., 1992)

GPCC V6 (Becker et al., 2013)

Trends in annual precipitation are expressed per decade, and are calcu-

lated for the time periods 1901–2010 and 1951–2010. See the Supple-

mentary Material of Chapter 2 for a detailed description of the meth-

odology used for trend and uncertainty calculations (Section 2.SM.3.3).

Trends have been calculated only for those grid boxes with greater

than 70% complete records and more than 20% data availability in

ﬁrst and last 10% of the time period. White areas indicate incomplete

or missing data. Black plus signs (+) indicate grid boxes where trends

are signiﬁcant at the 2-tailed 10% signiﬁcance level (i.e., a trend of

zero lies outside the 90% conﬁdence interval).

The Technical Summary provides maps for all 3 datasets (TS TFE.1,

Figure 2), while the Summary for Policymakers provides a map based

on GPCC only (Figure SPM.2).

TS.SM.3 Notes and Technical Details on

Observed Indicators of a Changing

Global Climate Figures for the Summary

for Policymakers – Figure SPM.3

This material documents the provenance of the data and plotting

procedures that were used to create Figure SPM.3 in the IPCC WGI

Fifth Assessment Report. This ﬁgure is closely derived from Figure TS.1

and FAQ 2.1, Figure 2 (see Chapter 2 Supplementary Material Section

2.SM.5), but includes fewer observed indicators. In addition, Figure

SPM.3 includes an estimate of uncertainty for those datasets where

this is available and has been assessed, illustrated with shading. Figure

SPM.3 includes datasets and parameters assessed in Chapters 3 (ocean

heat content, sea level), and 4 (snow cover, sea ice).

TSSM

Technical Summary Supplementary Material

TSSM-4

TS.SM.3.1 Northern Hemisphere Spring Snow Cover –

Figure SPM.3a

TS.SM.3.1.1 Datasets

Green: Northern Hemisphere annual March-April average snow-cover

extent based on an updated series from Brown and Robinson (2011),

1922–2012.

Shaded uncertainty estimate indicated by the 95% conﬁdence interval.

TS.SM.3.1.2 Plotting Techniques

Annual values are plotted.

TS.SM.3.2 Arctic Summer Sea Ice Extent – Figure SPM.3b

All datasets provide Arctic annual July-August-September average sea

ice extent.

Green: Updated from Walsh and Chapman (2001). Annual values are

from 1900–1978.

Blue: Hadley Centre Sea Ice and Sea Surface Temperature dataset (Had-

ISST1.2) (Rayner et al., 2003). Annual values are from 1900–1939 and

1953–2012. Values are excluded for the period 1940–1952 because

the available data showed no change. It was a period when in situ data

were very sparse and the gaps were ﬁlled in for completeness with

climatology. For this assessment, this was not considered sufﬁciently

robust and therefore the data during the period were excluded from

the time series.

Red: Bootstrap algorithm (SBA) applied to data from the Scanning Mul-

tichannel Microwave Radiometer (SMMR) (updated from Comiso and

Nishio, 2008). Annual values are from 1979–2012.

Black: NASA Team algorithm (NT1) applied to data from the Special

Sensor Microwave/Imager (SSM/I) (Cavalieri et al., 1984) – updated

in Cavalieri and Parkinson (2012) and Parkinson and Cavalieri (2012).

Annual values are from 1979–2011.

Yellow: Bootstrap algorithm (ABA) applied to data from the Advanced

Microwave Scanning Radiometer - Earth Observing System (AMSR-

E) (updated from Comiso and Nishio, 2008). Annual values are from

2002–2011.

Orange: Revised NASA Team algorithm (NT2) applied to data from the

Advanced Microwave Scanning Radiometer - Earth Observing System

(AMSR-E) (updated from Markus and Cavalieri, 2000). Annual values

are from 2002–2011.

Uncertainty estimates for each data point in the plots have been cal-

culated based on the interannual variability of the ice extents. The sys-

tematic errors are not considered because they are generally unknown

and are expected to be approximately constant from one year to

another and would not change the results of trend analyses signiﬁ-

cantly. The interannual variability of the extent and actual area of the

sea ice cover during the satellite era (since 1979) can be quantiﬁed

accurately because of global coverage at good temporal resolution and

the high contrast in the signature of ice free and ice covered oceans.

The uncertainty (shaded range) that is shown is 1 standard deviation

of the more than 30 years of satellite data, assuming a Gaussian dis-

tribution. The standard deviation is calculated after the data have been

linearly detrended.

For the pre-satellite data (pre 1979), the true interannual variability is

not known because available data are sparse and limited to only a few

locations. Based on the expected quality of the Walsh and Chapman

(2001) data and because of the lack of a better procedure, we use

1.75 standard deviations for the period 1880 to 1952 when data were

sparse and 1.5 standard deviation for the period 1953 to 1978 when

signiﬁcantly more data were available. For the HadISST1.2 data set,

which includes both pre- and post-satellite data (Rayner et al., 2003),

we use 1 standard deviation for the entire period since 1900, calculat-

ed after the data has been linearly detrended.

TS.SM.3.2.2 Plotting Techniques

Annual values are plotted.

TS.SM.3.3 Global Average Upper Ocean Heat Content –

Figure SPM.3c

TS.SM.3.3.1 Datasets

All datasets provide global annual upper-ocean (0 to 700 m depth)

heat content anomalies.

Blue: Updated from Palmer et al. (2007). Annual values are from 1950–

2011.

Green: Updated from Domingues et al. (2008). Annual values, smoothed

with a 3-year running mean, are from 1950–2011.

Yellow: Updated from Ishii and Kimoto (2009). Annual values are from

1950–2011.

Orange: Updated from Smith and Murphy (2007). Annual values are

from 1950–2010.

Black: Updated from Levitus et al. (2012). Annual values are from

1955–2011.

Uncertainty estimates are as reported in the cited publications. These

are one standard error of the mean, except for Levitus et al. (2012)

which provide one standard deviation. No uncertainty estimate is

available for Smith and Murphy (2007).

TS.SM.3.3.2 Plotting Techniques

The published ocean heat content anomaly datasets are relative to dif-

ferent climatological reference periods. Therefore, the datasets have

been aligned in Figure SPM.3c for the period 2006–2010, ﬁve years

that are well measured by Argo, and then plotted relative to the result-

TSSM

Technical Summary Supplementary Material

TSSM-5

ing mean of all curves for 1970, a time when the increasing availability

of annual data from XBTs causes the uncertainty estimates to reduce

considerably. Speciﬁcally the alignment procedure for Figure SPM.3c

involved the following steps:

Obtain all ﬁve upper ocean heat content anomaly time series.

1. Recognize that all the time-series values are annual values, cen-

tered on the middle of calendar years.

2. Find the average value of each time series for the years 2006–2010.

3. Subtract the average 2006–2010 value for each time series from

that speciﬁc time-series.

4. Find the value of each time series for the year 1970.

5. Average these ﬁve values from the year 1970.

6. Subtract this 1970 average value from all of the time-series.

TS.SM.3.4 Global Average Sea Level – Figure SPM.3d

TS.SM.3.4.1 Datasets

Black: Church and White (2011) tide gauge reconstruction. Annual

values are from 1900–2009.

Yellow: Jevrejeva et al. (2008) tide gauge reconstruction. Annual values

are from 1900–2002.

Green: Ray and Douglas (2011) tide gauge reconstruction. Annual

values are from 1900–2007.

Red: Nerem et al. (2010) satellite altimetry. A 1-year moving average

boxcar ﬁlter has been applied to give annual values from 1993–2009.

Shaded uncertainty estimates are one standard error as reported in

the cited publications. The one standard error on the 1-year averaged

altimetry data (Nerem et al., 2010) is estimated at ±1 mm, and thus

considerably smaller than for all other datasets.

TS.SM.3.4.2 Plotting Techniques

The published Global Mean Sea Level (GMSL) datasets use arbitrary

and different reference periods where they start from zero. Further-

more, the altimetry data begins only in 1993. Therefore, the datasets

have been aligned in Figure SPM.3d to a common reference period of

time using the following steps:

1. The longest running record (Church and White, 2011) is taken as the

reference to which all other datasets are aligned.

2. GMSL from Church and White (2011) is calculated relative to the

average for the period 1900–1905, and the resulting value for the

year 1993 (127 mm) is identiﬁed.

3. All other records are then adjusted to give the same value of 127

mm in 1993 (i.e., for each dataset the offset required to give 127

mm in 1993 is applied to all annual values in that dataset).

TS.SM.4 Notes and Technical Details on

Observed Changes in the Global

Carbon Cycle Figures in the Summary

for Policymakers – Figure SPM.4

TS.SM.4.1 Atmospheric Concentrations of Carbon

Dioxide – Figure SPM.4a

The top panel of Figure TS.5, and panel (a) of Figure SPM.4 show time

series of atmospheric concentrations of carbon dioxide (CO

). CO

con-

centrations are expressed as a mole fraction in dry air, micromol/mol,

abbreviated as ppm. Time series are shown for the Mauna Loa Obser-

vatory (red in Figure SPM.4a), and South Pole (black in Figure SPM.4a).

Data were accessed from the following sources (active at the time of

publication):

1. Mauna Loa Observatory

ftp://ftp.cmdl.noaa.gov/ccg/co2/trends/co2_mm_mlo.txt.

Monthly averages are plotted from March 1958 to August 2012. For

further details on the measurements see Keeling et al. (1976a) and

Thoning et al. (1989).

2. South Pole

http://scrippsco2.ucsd.edu/data/ﬂask_co2_and_isotopic/

monthly_co2/monthly_spo.csv

Monthly averages are plotted from June 1957 to February 2012. For

further details on the measurements see Keeling et al. (1976b; 2001).

TS.SM.4.2 Ocean Surface Carbon Dioxide and

In Situ pH – Figure SPM.4b

The top panel of Figure TS.5, and panel (b) of Figure SPM.4 show time

series of observed partial pressure of dissolved CO

(pCO

given in

µatm) at the ocean surface, together with time series of ocean surface

in situ pH (total scale). All ocean time series are plotted as 12-month

running means (6 months before to 6 months after the sample date)

for each 6-month period centered on 1 January and 2 July of each year.

Data for both pCO

and in situ pH were measured at the following

stations and obtained from the following sources (active at the time

of publication):

1. Hawaii Ocean Time-Series program (HOT) from the station ALOHA

(updated from, Dore et al., 2009)

http://hahana.soest.hawaii.edu/hot/products/HOT_surface_CO2.txt

Shown as light green and light blue time series in Figure SPM.4b,

for in situ pH and pCO

respectably. Data were plotted for the period

1988–2011.

Further technical details regarding the data are available from the

readme ﬁle: http://hahana.soest.hawaii.edu/hot/products/HOT_

surface_CO2_readme.pdf.

TSSM

Technical Summary Supplementary Material

TSSM-6

2. Bermuda Atlantic Time-Series Study (BATS):

http://bats.bios.edu/bats_form_bottle.html

Shown as green and blue time series in Figure SPM.4b, for in situ pH

and pCO

, respectively, but not shown in Figure TS.5. Data were plotted

for the period 1991 – 2011.

Measured dissolved inorganic carbon (DIC) and total alkalinity (TA) at

in situ temperature were used to calculate pH on the total scale as well

as pCO

in μatm.

Further technical details are described in Bates (2007).

3. European Station for Time series in the Ocean (ESTOC; see

González-Dávila and Santana-Casiano, 2009):

http://cdiac.ornl.gov/ftp/oceans/ESTOC_data

Shown as dark green and dark blue time series in Figure SPM.4b, for in

situ pH and pCO

, respectively, but not shown in Figure TS.5. Data were

plotted for the period 1996–2009.

Further technical details regarding the data are available from

González-Dávila (2010).

Note that the data for Figure SPM.4 (and Figure TS.5) provided at

the external sources cited above may be subject to revision based on

recalibration, and other quality control procedures conducted over

time by the data providers.

TS.SM.5 Notes and Technical Details on Radiative

Forcing Estimates Figure in the Summary

for Policy Makers – Figure SPM.5

This material documents the underlying traceability for values that

were used to create Figure SPM.5 in the IPCC WG1 Fifth Assessment

Report. This ﬁgure is closely related to Figures TS.6 and TS.7 and Chap-

ter 8, Figures 8.14 to 8.18. The reader is therefore referred to the Sup-

plementary Material of Chapter 8 for detailed information on methods

and sources used to estimate forcing values.

Figure SPM.5 (and Figure TS.7) plots Radiative Forcing (RF) estimates in

2011 relative to 1750 and aggregated uncertainties for the main drivers

of climate change. This ﬁgure is different from similar ﬁgures shown in

previous IPCC report SPMs (though an analogous ﬁgure was shown in

Chapter 2 of AR4) as it evaluates the RF based on the emissions rather

than the concentration changes. An emitted compound changes the

atmospheric concentration of the same substance but may also impact

that of other atmospheric constituents through chemistry processes.

Values are global average RF, partitioned according to the emitted

compounds or processes that result in a combination of drivers. In cal-

culations of RF for well-mixed greenhouse gases and aerosols in this

report, physical variables, except for the ocean and sea ice, are allowed

to respond to perturbations with rapid adjustments. The resulting forc-

ing is called Effective Radiative Forcing (ERF) in the underlying report.

For all drivers other than well-mixed greenhouse gases and aerosols,

rapid adjustments are less well characterized and assumed to be small,

and thus the traditional RF is used.

The ‘level of conﬁdence’ given in Figure SPM.5 is based on Table 8.5.

For the main emitted compounds of CO

, CH

, Halocarbons, N

O, CO,

NMVOC and NO

, the underlying values, their sources, and uncertain-

ties can be found in the Chapter 8 Supplementary Material, Tables

8.SM.6 and 8.SM.7.

The value of –0.27 W m

–2

for aerosols and precursors shown in Figure

SPM.5 results from –0.35 W m

–2

from RFari (Table 8.6) with the addi-

tion of 0.04 W m

–2

from BC-on-snow and the subtraction of the small

nitrate contribution from NO

of –0.04 W m

–2

(Table 8.SM.6).

The value of –0.55 W m

–2

for cloud adjustments due to aerosols given

in Figure SPM.5 results from the combination of ERFaci –0.45 [–1.2 to

0.0] W m

–2

and rapid adjustment of ari –0.1 [–0.3 to +0.1] W m

–2

reported in Figure TS.7. Detailed information can be found in Chapter 8

and the Chapter 8 Supplementary Material, Table 8.SM.6.

The values for albedo changes due to land use and changes in solar

irradiance come from Table 8.6 of Chapter 8.

Total anthropogenic RF relative to 1750 is based on values given in

Table 8.6 (for 2011) and Figure 8.18 (values for 1950 and 1980 given

in the caption).

TS.SM.6 Notes and Technical Details on

Comparison of Observed and Simulated

Climate Change Figures for the Summary

for Policymakers – Figure SPM.6

Figure SPM.6 and the related Figure TS.12 are reduced versions of

Figure 10.21 in Chapter 10. The reader is therefore referred to the

detailed description of the main components of Figure 10.21 for data-

sets and methods used (see the Chapter 10 Supplementary Material,

Section 10.SM.1). Here, mainly the differences of Figure SPM.6 and

TS.12 from Figure 10.21 are listed.

Figures SPM.6 and TS.12 show time series of decadal average, plotted

on a common axis and at the centre of each decade. The decadal aver-

ages are taken from the annual time series that Figure 10.21 is based

on. Figure TS.12 features the multi-model mean as dark blue and dark

red line, while Figure SPM.6 only features the 5–95% conﬁdence inter-

val. Note that the precipitation plot from Figure 10.21 are not included

in the Technical Summary and SPM versions of this ﬁgure.

TS.SM.6.1 Continental Temperatures

The same model simulations and observational data sets are used as

for Figure 10.21. Continental land areas are based on the IPCC Special

Report on Managing the Risks of Extreme Events and Disasters to

Advance Climate Change Adaptation (SREX) deﬁned regions (IPCC,

2012) shown pictorially in the bottom right most panel of Figure 10.7.

Temperature anomalies in Figure SPM.6 are with respect to 1880–1919

(except for Antarctica where anomalies are relative to 1950–2010).

TSSM

Technical Summary Supplementary Material

TSSM-7

TS.SM.6.2 Ocean Heat Content

The same model simulations and observational data sets are used as

for Figure 10.21.

TS.SM.6.3 Sea Ice

The same model simulations and observational data sets are used as

for Figure 10.21.

TS.SM.6.4 Data Quality

For land and ocean surface temperatures panels, solid lines indicate

where data spatial coverage of areas being examined is above 50%

coverage and dashed lines where coverage is below 50%. For example,

data coverage of Antarctica never goes above 50% of the land area of

the continent. For ocean heat content and sea-ice panels, the solid line

is where the coverage of data is good and higher in quality, and the

dashed line is where the data coverage is only adequate, based respec-

tively on the spatial coverage and instrument type and on the presence

of satellite measurements.

TS.SM.7 Notes and Technical Details on CMIP5

Simulated Time Series Figures in the

Summary for Policymakers –

Figure SPM.7

This material documents the provenance of the data and plotting

procedures that were used to create Figure SPM.7, based on Climate

Model Intercomparison Project Phase 5 (CMIP5) model results as of

March, 2013. This ﬁgure is closely derived from Figures12.5 and TS.15

(global average surface temperature), 12.28 and TS.17 (sea ice), 6.28

and TS.20a (ocean surface pH), but includes fewer model scenarios. The

reader is referred to the Technical Summary and the Chapters 12 and 6

where all RCP scenarios are given for the respective quantity.

TS.SM.7.1 Global Average Surface Temperature Change

(Figure SPM.7a) and Global Ocean Surface pH

(Figure SPM.7c)

Step 1 – Analyzed simulations

The simulations considered are annual or monthly mean ﬁelds from

different model simulations carried out as part of the CMIP5 project

(when applicable the variable name as given in the CMIP5 archive is

indicated in square brackets). The time series between 1850 and 2005

originate from the historical simulations. The two time series of the

future projections are from RCP2.6 and RCP8.5. The box plots show-

ing the change at the end of the century additionally use RCP4.5 end

RCP6.0. Table TS.SM.1 lists the models and ensemble simulations used

for panels (a) and panel (c). Only one ensemble simulation per model

is used. All models are weighted equally except for sea ice (panel (b))

where a subset of models is considered.

Step 2a – Interpolation

For panel (a), the monthly temperature ﬁelds [tas] are re-gridded to a

2.5° × 2.5° grid using bilinear interpolation. No special treatment is

used at the land-sea border.

For panel (c), the monthly temperature [tos] and salinity [sos] ﬁelds are

ﬁrst averaged to yield annual means. Then, annual-mean temperature,

salinity, dissolved inorganic carbon [dissic] and alkalinity [talk] ﬁelds

are re-gridded to a 1° × 1° using bilinear interpolation. For the model

MIROC-ESM-CHEM the upper-most layers of the 3-dimensional ﬁelds

of monthly sea water potential temperature [thetao] and monthly sea

water salinity [so] are used.

Step 2b – Derivation of pH

For each model, surface pH was computed from simulated DIC, alka-

linity, temperature, and salinity. Before computation each simulated

input ﬁeld was corrected for its decadal mean bias relative to modern

observations, using the approach of Orr et al. (2005) and Orr (2011).

That is, pH was computed after ﬁrst removing from each model ﬁeld,

the average difference between the model mean during 1989–1998

Table TS.SM.1 | Models and ensembles used for panels (a) and (c).

Model Ensemble Member Historical RCP2.6 RCP4.5 RCP6.0 RCP8.5

ACCESS1.0 r1i1p1 (a) (a) (a)

ACCESS1.3 r1i1p1 (a) (a) (a)

BCC-CSM1.1 r1i1p1 (a) (a) (a) (a) (a)

BCC-CSM1.1(m) r1i1p1 (a) (a) (a) (a)

BNU-ESM r1i1p1 (a) (a) (a) (a)

CanESM2 r1i1p1 (a) (c) (a) (c) (a) (c) (a) (c)

CCSM4 r1i1p1 (a) (a) (a) (a) (a)

CESM1(BGC) r1i1p1 (a) (a) (a)

CESM1(CAM5) r1i1p1 (a) (a) (a) (a) (a)

CMCC-CM r1i1p1 (a) (a) (a)

CMCC-CMS r1i1p1 (a) (a) (a)

CNRM-CM5 r1i1p1 (a) (a) (a)

CSIRO-Mk3.6.0 r1i1p1 (a) (a) (a) (a) (a)

EC-EARTH r8i1p1 (a) (a) (a) (a)

(continued on next page)

TSSM

Technical Summary Supplementary Material

TSSM-8

Model Ensemble Member Historical RCP2.6 RCP4.5 RCP6.0 RCP8.5

FGOALS-g2 r1i1p1 (a) (a) (a) (a)

FIO-ESM r1i1p1 (a) (a) (a) (a) (a)

GFDL-CM3 r1i1p1 (a) (a) (a) (a) (a)

GFDL-ESM2G r1i1p1 (a) (c) (a) (c) (a) (c) (a) (c) (a) (c)

GFDL-ESM2M r1i1p1 (a) (c) (a) (c) (a) (c) (a) (c) (a) (c)

GISS-E2-H r1i1p1 (a) (a) (a) (a) (a)

GISS-E2-H r1i1p2 (a) (a) (a) (a) (a)

GISS-E2-H r1i1p3 (a) (a) (a) (a) (a)

GISS-E2-H-CC r1i1p1 (a) (a)

GISS-E2-R r1i1p1 (a) (a) (a) (a) (a)

GISS-E2-R r1i1p2 (a) (a) (a) (a) (a)

GISS-E2-R r1i1p3 (a) (a) (a) (a) (a)

GISS-E2-R-CC r1i1p1 (a) (a)

HadGEM2-AO r1i1p1 (a) (a) (a) (a) (a)

HadGEM2-CC r1i1p1 (a) (c) (a) (c) (a)

HadGEM2-ES r2i1p1 (a) (a) (a) (a) (a)

INM-CM4 r1i1p1 (a) (a) (a)

IPSL-CM5A-LR r1i1p1 (a) (c) (a) (c) (a) (c) (a) (c) (a) (c)

IPSL-CM5A-MR r1i1p1 (a) (c) (a) (c) (a) (c) (a) (a) (c)

IPSL-CM5B-LR r1i1p1 (a) (c) (a) (a) (c)

MIROC5 r1i1p1 (a) (a) (a) (a) (a)

MIROC-ESM r1i1p1 (a) (c) (a) (c) (a) (c) (a) (a) (c)

MIROC-ESM-CHEM r1i1p1 (a) (c) (a) (c) (a) (c) (a) (c) (a) (c)

MPI-ESM-LR r1i1p1 (a) (c) (a) (c) (a) (c) (a) (c)

MPI-ESM-MR r1i1p1 (a) (c) (a) (c) (a) (c) (a) (c)

MRI-CGCM3 r1i1p1 (a) (a) (a) (a) (a)

NorESM1-M r1i1p1 (a) (a) (a) (a) (a)

NorESM1-ME r1i1p1 (a) (c) (a) (a) (c) (a) (a)

Table TS.SM.1 (continued)

and the observational reference. For observed ﬁelds, the GLODAP grid-

ded data product (Key et al., 2004) for DIC and alkalinity along with

the 2009 World Ocean Atlas climatology for temperature, salinity, and

concentrations of phosphate and silica (Locarnini et al., 2010; Antonov

et al., 2010; Garcia et al., 2010) were used. Changes to the concentra-

tions of phosphate and silica were assumed to be zero, because not

all models provided those variables. pH was computed using routines

based on the standard OCMIP carbonate chemistry adapted for earlier

studies (Orr, 2011) to compute all carbonate system variables and use

recommended constants from the Guide to Best Practices for Ocean

Measurements (Dickson et al., 2007).

Step 3 – Global and annual mean

The monthly (temperature) or annual (pH) surface ﬁelds are averaged

(weighted by the cosine of the latitude) to obtain the global mean

values. The monthly global mean temperature values are averaged to

annual means.

Step 4 – Reference period

The average from 1986 to 2005 of the annual means for each model is

computed and is subtracted from the respective model time series to

obtain the corresponding temperature anomalies.

Step 5 – Mean and standard deviation

The mean and standard deviation over all the models is calculated. For

the time period after 2006 all the possible models that are listed in

Table TS.SM.1 are used. If a model provided several RCPs based on the

same historical simulation, that historical simulation is counted only

once.

Step 6 – Uncertainty estimates

First, for each model the average from 2081 to 2100 is computed from

the above mentioned time series. Then, in a second step, the mul-

ti-model average and standard deviation over all model averages are

calculated. The likely ranges on the right of the ﬁgure show the mean

plus/minus 1.64 times the standard deviation across the model averag-

es. The shading on the time series indicates the mean value plus/minus

1.64 times the standard deviation across the models for each year.

Step 7 – Graphical display

To close the multi-model mean time series at the year 2005 when the

historical simulation ends and the RCP begins, the value at year 2005

is assigned to belong to both the historical time series and also to the

corresponding RCP.

TSSM

Technical Summary Supplementary Material

TSSM-9

TS.SM.7.2 Northern Hemisphere September Sea Ice

Extent – Figure SPM.7b

Step 1 – Analyzed simulations

Table TS.SM.2 provides the model and RIP ensemble member included

from each RCP to create the multi-model mean time series of the NH

September sea ice extent [sic] shown in Figure SPM.7b. In most cases,

the ﬁrst ensemble member (r1i1p1) was used. A selection algorithm

produces a subset of models that most closely match observations,

and is detailed below. The corresponding historical ensemble member

Model Ensemble Member RCP2.6 Historical/RCP4.5 RCP6.0 RCP8.5

ACCESS1.0 r1i1p1 x x

ACCESS1.3 r1i1p1 x x

BCC-CSM1.1 r1i1p1 x x x x

BCC-CSM1.1(m) r1i1p1 x x x x

BNU-ESM r1i1p1 x x x

CanESM2 r1i1p1 x x x

CCSM4 r1i1p1 x x x x

CESM1(BGC) r1i1p1 x x

CESM1(CAM5) r1i1p1 x x x x

CESM1(WACCM) r2i1p1 x x x

CMCC-CM r1i1p1 x x

CMCC-CMS r1i1p1 x x

CNRM-CM5 r1i1p1 x x x

CSIRO-Mk3.6.0 r1i1p1 x x x x

EC-EARTH r1i1p1 x x

r8i1p1 x

FGOALS-g2 r1i1p1 x x x

FIO-ESM r1i1p1 x x x x

GFDL-CM3 r1i1p1 x x x x

GFDL-ESM2G r1i1p1 x x x x

GFDL-ESM2M r1i1p1 x x x x

GISS-E2-H r1i1p1 x x x x

GISS-E2-H-CC r1i1p1 x

GISS-E2-R r1i1p1 x x x x

GISS-E2-R-CC r1i1p1 x

HadGEM2-AO r1i1p1 x x x x

HadGEM2-CC r1i1p1 x x

HadGEM2-ES r2i1p1 x x x x

INM-CM4 r1i1p1 x x

IPSL-CM5A-LR r1i1p1 x x x x

IPSL-CM5A-MR r1i1p1 x x x x

IPSL-CM5B-LR r1i1p1 x x

MIROC5 r1i1p1 x x x x

MIROC-ESM r1i1p1 x x x x

MIROC-ESM-CHEM r1i1p1 x x x x

MPI-ESM-LR r1i1p1 x x x

MPI-ESM-MR r1i1p1 x x x

MRI-CGCM3 r1i1p1 x x x x

NorESM1-M r1i1p1 x x x x

NorESM1-ME r1i1p1 x x x x

is catenated with the respective RCP scenario ensemble member to

create a continuous time series from 1850–2100.

Step 2 – Time series of NH September sea ice extent

Using the sea ice concentration ﬁeld, a mask of the sea ice concentra-

tion >15% for each month of data for the Northern Hemisphere was

created. For each month, the sea ice extent is the sum of the area of

the ocean [areacello] times the ocean fraction [sftof] times the mask

of sic >15% at each grid point. The time series are computed on the

original model grids, which is usually the ocean grid. In some cases,

Table TS.SM.2 | Models and ensemble members used.

TSSM

Technical Summary Supplementary Material

TSSM-10

sea ice concentration is on the atmospheric grid. In cases where the

grid area was not available for regular grids, a regular lat-lon grid was

constructed based on the grid dimensions following

areacello=((dlat*2π/360)*R_earth) .* ((dlon*2 π/360).*(R_earth*-

cos(LAT))),

with R_earth being the radius of Earth (6,371,000 m), dlat and dlon

being the differentials in lat/lon in each dimension, and LAT being the

latitude in radians.

If the ocean fraction was unavailable, it was assumed that the ocean

fraction was 1 where the sea ice concentration was greater than 0%.

Step 3 – Create multi-model mean time series

The multi-model mean time series of sea ice extent is computed

across all model members in Table TS.SM.2. A ﬁve-year running mean

is applied to this time series. This is plotted as the dotted line in the

ﬁgure. Some time series start later than 1850 or end earlier than 2100,

and these are treated as missing values for those years.

Step 4 – Select models that most closely match observations

The selection process is done in a series of steps which compare the

models to observed/reanalyzed data. This selection process is based on

the underlying assessment of Chapter 12 and referenced therein. The

method proposed by Massonnet et al. (2012) is applied here to the full

set of models that provided sea ice output ﬁelds to the CMIP5 data-

base. For the model selection, all available ensemble members are used

for all of the models that provide simulations for Historical and RCP4.5.

These ensemble members are listed in Table TS.SM.3.

Four diagnostics from the models are compared to the same quantities

in observations or reanalyses. The diagnostics are: (a) September Arctic

sea ice extent (1986–2005), (b) Annual mean Arctic sea ice volume

(1986–2005), (c) Amplitude of the seasonal cycle of Arctic sea ice

extent (1986–2005), and (d) Trend in September Arctic sea ice extent

(1979–2012). Computation of each diagnostic is described and then

the method for comparison is described below.

Step 4a – Computation of diagnostic quantities

(a) Sea ice extent is computed for each model ensemble member as

outlined above to get the total area where sea ice concentration is

>15%. For each ensemble member, an average September sea ice

extent is then computed for the years 1986–2005. Observations for

sea ice extent use the monthly mean sea ice extents from Comiso and

Nishio (2008, updated 2012). The observations were computed in the

same way as in the models (i.e., these are the monthly mean extents

computed from the observed monthly mean sea ice concentration).

(b) Sea ice volume is computed as the sum of the sea ice thickness ﬁeld

[sit] times the ocean area [areacello] times the ocean fraction [sftof],

since the sea ice thickness is given as thickness averaged over the

entire ocean grid cell. Caveats for the grids are the same as discussed

in Step 2 above. The time series of monthly sea ice volume for each

ensemble member is then annually averaged for the period 1986–

2005. The bias-adjusted PIOMAS (Pan-Arctic Ice-Ocean Modelling and

Assimilation System) reanalysis data (Schweiger et al., 2011) is used

to provide estimates for sea ice volume for comparison to the models.

puted for each model from a climatology of monthly sea ice extent

for 1986–2005. The amplitude is the difference between the maximum

(March) and minimum (September) sea ice extent for each model

ensemble member. Amplitude of seasonal cycle for observations are

computed in the same way from Comiso and Nishio (2008, updated

2012).

(d) The linear trend in September sea ice extent is computed for the

period 1979–2012. Again observations are taken from Comiso and

Nishio (2008, updated 2012).

Step 4b – Estimation of natural variability for model ensembles

For models with multiple ensemble members, a standard deviation

is computed for each of the diagnostics for each ensemble member.

Then the mean of all the standard deviations is computed, and using

this value, a ±2 standard deviation interval is constructed around the

ensemble mean or single realization of each diagnostic for each model.

Step 4c – Model selection - Comparison of modeled diagnostics

to observed/reanalyzed diagnostic

For each of the observed/reanalyzed diagnostics, a ±–20% interval is

constructed around the mean value for the given period. A model is

retained in the selection if, for each diagnostic, either the ±2 standard

deviation around the model ensemble mean diagnostic overlaps the

±20% interval around the observed/reanalysed value of the diagnostic

OR at least one ensemble member from that model gives a value for

the diagnostic that falls within ±20% of the observed/reanalysed data.

A model is selected only if all four diagnostic values meet this criterion.

Results of the selection

The model diagnostics are calculated using RCP4.5 which has the

largest number of models. Five models are selected by this process:

ACCESS1.0, ACCESS1.3, GFDL-CM3, IPSL-CM5A-MR, MPI-ESM-MR,

and all ﬁve models have simulations for both RCP8.5 and RCP4.5. For

RCP2.6 only three of this subset have simulations (GFDL-CM3, IPSL-

CM5A-MR, MPI-ESM-MR), and for RCP6.0, only two models have sim-

ulations (GFDL-CM3, IPSL-CM5A-MR).

Step 5 – Time series of sea ice extent for the selected models

The multi-model mean time series of September sea ice extent is cal-

culated for the selected models. The solid line shows the multi-model

mean smoothed with a ﬁve-year running mean, and the shading rep-

resents the minimum and maximum range of the selected model time

series, also smoothed by the same ﬁve year running mean.

The shaded bars on the right are the multi-model mean and the mean

of the maximum and minimum range for the selected models for the

period 2081–2100.

TSSM

Technical Summary Supplementary Material

TSSM-11

Model Ensemble Member RCP4.5

ACCESS1.0 r1i1p1

ACCESS1.3 r1i1p1

BCC-CSM1.1 r1i1p1

BCC-CSM1.1(m) r1i1p1

BNU-ESM r1i1p1

CanESM2 r1i1p1

r2i1p1

r3i1p1

r4i1p1

r5i1p1

CCSM4 r1i1p1

r2i1p1

r3i1p1

r4i1p1

r5i1p1

r6i1p1

CESM1(BGC) r1i1p1

CESM1(CAM5) r1i1p1

r2i1p1

r3i1p1

CESM1(WACCM) r2i1p1

CMCC-CM r1i1p1

CMCC-CMS r1i1p1

CNRM-CM5 r1i1p1

CSIRO-Mk3.6.0 r1i1p1

r2i1p1

r3i1p1

r4i1p1

r5i1p1

r6i1p1

r7i1p1

r8i1p1

r9i1p1

10i1p1

EC-EARTH r1i1p1

r2i1p1

r3i1p1

r6i1p1

r7i1p1

r8i1p1

r9i1p1

10i1p1

11i1p1

12i1p1

13i1p1

14i1p1

FGOALS-g2 r1i1p1

FIO-ESM r1i1p1

r2i1p1

r3i1p1

GFDL-CM3 r1i1p1

r3i1p1

r5i1p1

GFDL-ESM2G r1i1p1

GFDL-ESM2M r1i1p1

GISS-E2-H r1i1p1

r2i1p1

r3i1p1

r4i1p1

r5i1p1

GISS-E2-H-CC r1i1p1

Model Ensemble Member RCP4.5

GISS-E2-R r1i1p1

r2i1p1

r3i1p1

r4i1p1

r5i1p1

r6i1p1

GISS-E2-R-CC r1i1p1

HadGEM2-AO r1i1p1

HadGEM2-CC r1i1p1

HadGEM2-ES r2i1p1

r3i1p1

r4i1p1

INM-CM4 r1i1p1

IPSL-CM5A-LR r1i1p1

r2i1p1

r3i1p1

r4i1p1

IPSL-CM5A-MR r1i1p1

IPSL-CM5B-LR r1i1p1

MIROC5 r1i1p1

r2i1p1

r3i1p1

MIROC-ESM r1i1p1

MIROC-ESM-CHEM r1i1p1

MPI-ESM-LR r1i1p1

r2i1p1

r3i1p1

MPI-ESM-MR r1i1p1

r2i1p1

r3i1p1

MRI-CGCM3 r1i1p1

NorESM1-M r1i1p1

NorESM1-ME r1i1p1

Table TS.SM.3 | Models and ensembles used for model selection, RCP4.5.

TS.SM.8 Notes and Technical Details on Maps

Showing CMIP5 Results in the Summary

for Policymakers – Figure SPM.8

This material documents the provenance of the data and plotting

procedures that were used to create Figure SPM.8, based on CMIP5

model results as of March, 2013. This ﬁgure is closely derived from Fig-

ures12.11 and TS.15 (global average surface temperature), TS.16 (pre-

cipitation), 12.29 and TS.17 (sea ice), 6.28 and TS.20b (ocean surface

pH), but includes fewer model scenarios. The reader is referred to the

Technical Summary or the Chapters 12 and 6 where all RCP scenarios

are given for the respective quantity.

TS.SM.8.1 Change in Average Surface Temperature

(Figure SPM.8a) and Change in Average

Precipitation (Figure SPM.8b)

Step 1 – Analyzed simulations

The simulations considered are monthly mean ﬁelds of surface tem-

perature [tas] and precipitation [pr] from different model simulations

carried out as part of the CMIP5 project (when applicable the variable

name as given in the CMIP5 archive is indicated in square brackets).

Table TS.SM.4 lists the models and ensemble members used for these

panels. Only one ensemble member per model is used.

TSSM

Technical Summary Supplementary Material

TSSM-12

Step 2 – Interpolation

In a ﬁrst step the monthly ﬁelds are re-gridded to a 2.5° × 2.5° grid

using bilinear interpolation. No special treatment is used at the land-

sea border.

Table TS.SM.4 | Models and ensemble members used.

Model Ensemble Member RCP2.6 Historical/RCP4.5 RCP6.0 RCP8.5

ACCESS1.0 r1i1p1 x x

ACCESS1.3 r1i1p1 x x

BCC-CSM1.1 r1i1p1 x x x x

BCC-CSM1.1(m) r1i1p1 x x x

BNU-ESM r1i1p1 x x x

CanESM2 r1i1p1 x x x

CCSM4 r1i1p1 x x x x

CESM1(BGC) r1i1p1 x x

CESM1(CAM5) r1i1p1 x x x x

CMCC-CM r1i1p1 x x

CMCC-CMS r1i1p1 x x

CNRM-CM5 r1i1p1 x x

CSIRO-Mk3.6.0 r1i1p1 x x x x

EC-EARTH r8i1p1 x x x

FGOALS-g2 r1i1p1 x x x

FIO-ESM r1i1p1 x x x x

GFDL-CM3 r1i1p1 x x x x

GFDL-ESM2G r1i1p1 x x x x

GFDL-ESM2M r1i1p1 x x x

GISS-E2-H r1i1p1 x x x x

GISS-E2-H r1i1p2 x x x x

GISS-E2-H r1i1p3 x x x x

GISS-E2-H-CC r1i1p1 x

GISS-E2-R r1i1p1 x x x x

GISS-E2-R r1i1p2 x x x x

GISS-E2-R r1i1p3 x x x x

GISS-E2-R-CC r1i1p1 x

HadGEM2-AO r1i1p1 x x x x

HadGEM2-CC r1i1p1 x x

HadGEM2-ES r2i1p1 x x x x

INM-CM4 r1i1p1 x x

IPSL-CM5A-LR r1i1p1 x x x x

IPSL-CM5A-MR r1i1p1 x x x x

IPSL-CM5B-LR r1i1p1 x x

MIROC5 r1i1p1 x x x x

MIROC-ESM r1i1p1 x x x x

MIROC-ESM-CHEM r1i1p1 x x x x

MPI-ESM-LR r1i1p1 x x x

MPI-ESM-MR r1i1p1 x x x

MRI-CGCM3 r1i1p1 x x x x

NorESM1-M r1i1p1 x x x x

NorESM1-ME r1i1p1 x x x x

Step 3 – Annual average and period

The monthly mean values are averaged to annual means. Then in a

second step the time mean is computed over the 20-year period of

interest.

TSSM

Technical Summary Supplementary Material

TSSM-13

Step 4 – Time average and anomalies

The average from 1986 to 2005 of the annual means for each model

is computed as the reference value and the annual mean from 2081 to

2100 are computed as the future period for the two RCPs. For each model

the reference value is then subtracted from the future period value.

Step 5 – Calculation of the signiﬁcance

Step 5a – Natural variability

To compute the natural variability all the models that provide more

than 500 years of pre-industrial control simulation [piControl] are

used. A list of these models is given in Table TS.SM.5. For each model

the ﬁrst 100 years are discarded to minimize problems with model

initialization. Re-gridding and calculation of annual means is done as

described in steps 2 and 3. The control runs are divided into 20-year

non-overlapping periods. If the available data are not a multiple of

20-year the remaining years after the last 20-year period are not used

in the calculation.

Averages over the 20-year periods are computed for every grid point.

A quadratic trend is subtracted from this time series of 20-year aver-

aged periods to remove potential model drift at each grid point. Finally

Model Ensemble Member

ACCESS1.0 r1i1p1

ACCESS1.3 r1i1p1

BCC-CSM1.1 r1i1p1

BNU-ESM r1i1p1

CanESM2 r1i1p1

CCSM4 r1i1p1

CESM1(BGC) r1i1p1

CMCC-CMS r1i1p1

CNRM-CM5 r1i1p1

CSIRO-Mk3-6-0 r1i1p1

FGOALS-g2 r1i1p1

FIO-ESM r1i1p1

GFDL-CM3 r1i1p1

GFDL-ESM2G r1i1p1

GFDL-ESM2M r1i1p1

GISS-E2-H r1i1p2

GISS-E2-H r1i1p3

GISS-E2-R r1i1p2

GISS-E2-R r1i1p3

INM-CM4 r1i1p1

IPSL-CM5A-LR r1i1p1

MIROC5 r1i1p1

MIROC-ESM r1i1p1

MPI-ESM-LR r1i1p1

MPI-ESM-MR r1i1p1

MPI-ESM-P r1i1p1

MRI-CGCM3 r1i1p1

NorESM1-M r1i1p1

Table TS.SM.5 | Models and ensemble members from the piControl experiments used

for the calculation of the natural variability.

for each model the standard deviation is computed over the different

20-year periods and for each grid point.

To obtain the ﬁnal value of the natural variability the median of the

standard deviations of the different models is multiplied with the

square root of 2 (the natural variability characterizes the typical dif-

ference between two 20-year periods, rather than the difference of

one period from the long-term mean, the former being larger than the

latter by the square root of two).

Step 5b – Testing for signiﬁcance

For each model the projected change is taken relative to its reference

period and then the multi-model average at every grid point is com-

puted. In a second step, at each grid point the number of models with

positive and negative change are counted.

If more than 90% of the models agree on the sign of the change and

the multi-model mean change is larger than 2 times the natural var-

iability (as deﬁned above) this grid point is said to be signiﬁcant and

robust across models.

Step 5c – Check for non-signiﬁcance

Again, for each model the projected change is taken relative to the

reference period and then the multi-model average at every grid point

is computed.

If the multi-model mean change at one grid point is less than the natu-

ral variability (as deﬁned above) the value is said to be non-signiﬁcant.

Step 6 – Graphical display

For each model the projected change is taken relative to the reference

period and then the multi-model average at every grid point is comput-

ed. The locations that are signiﬁcant and robust (as described in step

5b) are marked by small black dots and the locations that are non-sig-

niﬁcant (as described in step 5c) are marked by hatching.

For panel b, all calculations are performed as absolute changes. To

show the relative changes, the multi-model mean precipitation change

is divided by the multi-model mean of the reference period.

TS.SM.8.2 Northern Hemisphere September Sea Ice

Extent (Figure SPM.8c)

Step 1 – Analyzed simulations and subset of models

The simulations analyzed here are the same as those listed for Figure

SPM.7b. The subset of models are the same that are selected for Figure

SPM.7b outlined in the following Step 4. Only one ensemble member

from each model is used to create these ﬁgures.

Step 2 – Computation of mean sea ice concentration

For each model ensemble member, the mean sea ice concentration [sic]

is calculated for the two periods, 1986–2005 and 2081–2100, on the

native model grid (see also recipe for Figure SPM.7b).

Step 3 – Regrid sea ice concentration to common grid

SOSIE (http://sosie.sourceforge.net/) is used to regrid the mean sea

ice concentration to a common 1° × 1° grid, applying the bilinear

TSSM

Technical Summary Supplementary Material

TSSM-14

interpolation scheme (SOSIE: cmethod = ‘bilin’). Further, the regridded

sea ice concentrations are ‘drowned’ across the land-sea boundary to

eliminate low-biased interpolated values in the area of land-sea tran-

sition (SOSIE: ldrown = T). With this approach, interpolation artifacts

can occur throughout the Canadian Archipelago, since each model rep-

resents this area quite differently. Comparison of individual models on

their native grid allows to identify and mask such areas. Note that, for

these reasons the interpolated sea ice concentrations shall not be used

for quantitative interpretation, but only for visualization purposes. For

visualization the MATLAB land-ocean mask is overlaid.

Step 4 – Calculate multi-model mean sea ice concentration

For each RCP, RCP2.6 and RCP8.5, and each period, 1986–2005 and

2081–2100, the mean sea ice concentration is calculated in each grid

cell on the common grid. The same is done for the subset of models

for each period. For RCP2.6 this subset is GFDL-CM3, IPSL-CM5A-MR,

MPI-ESM-MR. For RCP8.5 this subset is ACCESS1.0, ACCESS1.3, GFDL-

CM3, IPSL-CM5A-MR, MPI-ESM-MR.

Step 5 – Contour the multi-model mean sea ice concentration

of 15%

The multi-model mean sea ice concentration is contoured at 15%

according to the following:

1986–2005: multi-model mean all models: white line

1986–2005: subset models: light blue line

2081–2100: multi-model mean all models: white ﬁlled patch

2081–2100: subset models: light blue ﬁlled patch

Note for RCP8.5 there is no sea ice concentration >15% for the subset

of models.

The decision was taken to contour the 15% contour of mean sea ice

concentration to make this ﬁgure consistent with Figure 12.29, which

shows a contour plot of the multi-model mean sea ice concentrations.

It is also possible to make binary ﬁelds of sea ice concentration >15%,

take the mean of those binary ﬁelds (for both 20 year averages and

then in multi-model averages), and contour the 50% contour of the

mean binary ﬁeld as the mean sea ice extent. This option was not

chosen here.

Model Ensemble Member Historical RCP2.6 RCP4.5 RCP6.0 RCP8.5

CanESM2 r1i1p1 d d d d

GFDL-ESM2G r1i1p1 d d d d d

GFDL-ESM2M r1i1p1 d d d d d

HadGEM2-CC r1i1p1 d d d

IPSL-CM5A-LR r1i1p1 d d d d d

IPSL-CM5A-MR r1i1p1 d d d d

IPSL-CM5B-LR r1i1p1 d d d

MIROC-ESM r1i1p1 d d d d d

MIROC-ESM-CHEM r1i1p1 d d d d d

MPI-ESM-LR r1i1p1 d d d d

MPI-ESM-MR r1i1p1 d d d d

NorESM1-ME r1i1p1 d d

Table TS.SM.6 | Models and ensemble members used.

TS.SM.8.3 Change in Ocean Surface pH (Figure SPM.8d)

Step 1 – Analyzed simulations

The simulations considered are annual or monthly mean ﬁelds from

different model simulations carried out as part of the CMIP5 project

(when applicable the variable name as given in the CMIP5 archive

is indicated in square brackets). Table TS.SM.6 lists the models and

ensemble members used for these panels. Only one ensemble member

per model is used.

Step 2a – Interpolation

In a ﬁrst step, the monthly temperature [tos] and salinity [sos] ﬁelds

are ﬁrst averaged to yield annual means. For the model MIROC-ESM-

CHEM the upper-most layer of the 3-dimensional ﬁelds of monthly sea

water potential temperature [thetao] and monthly sea water salinity

[so] are used. Then, annual-mean temperature, salinity, dissolved inor-

ganic carbon [dissic] and alkalinity [talk] ﬁelds are re-gridded to a 1° ×

1° using bilinear interpolation.

Step 2b – Derivation of pH

For each model, surface pH was computed from simulated DIC, alkalini-

ty, temperature, and salinity. Before computation each simulated input

ﬁeld was corrected for its decadal mean bias relative to modern obser-

vations, using the approach used in Orr et al. (2005) and Orr (2011).

That is, pH was computed after ﬁrst removing from each model ﬁeld,

the average difference between the model mean during 1989–1998

and the observational reference. For observed ﬁelds, we used the

GLODAP gridded data product (Key et al., 2004) for DIC and alkalinity

along with the 2009 World Ocean Atlas climatology for temperature,

salinity, and concentrations of phosphate and silica (Locarnini et al.,

2010; Antonov et al., 2010; Garcia et al., 2010). Changes to the con-

centrations of phosphate and silica were assumed to be zero, because

all models did not provide those variables. pH was computed using

routines based on the standard OCMIP carbonate chemistry adapted

for earlier studies (Orr, 2011) to compute all carbonate system varia-

bles and use recommended constants from the Guide to Best Practices

for Ocean CO

Measurements (Dickson et al., 2007).

Step 3 – Average of 20-year period

The time mean is computed over the 20-year period of interest.

TSSM

Technical Summary Supplementary Material

TSSM-15

Step 4 – Time average and anomalies

The average from 1986 to 2005 of the annual means for each model

is computed as the reference value and the annual mean from 2081

to 2100 is computed as the future period for the two RCPs. For each

model the reference value is then subtracted.

Step 5 – Graphical display

For each model the projected change is taken relative to the reference

period and the multi-model mean at every grid point is computed.

TS.SM.9 Notes and Technical Details on the Sea

Level Projection Figure for the Summary

for Policymakers – Figure SPM.9

A full and comprehensive description of the methods used in the pro-

jections of global mean sea level for the 21st century is provided in the

Supplementary Material to Chapter 13 (see Section 13.SM.1). Further

plotting details used to produce Figure SPM.9, and the related Figure

TS.22 are provided here.

TS.SM.9.1 Projected Global Mean Sea Level Rise

Projections are given from process-based models of global mean

sea level rise relative to 1986–2005 for the four emissions scenarios

RCP2.6, RCP4.5, RCP6.0 and RCP8.5.

The likely range for each RCP timeseries is delimited by the data in ﬁles

rcpXX_sumlower and rcpXX_sumupper, while the median timeseries

is the data in ﬁle rcpXX_summid, where ‘XX’ stands for the respective

RCP scenario. These data ﬁles are available from the WGI AR5 website

www.climatechange2013.org. The coloured vertical bars with horizon-

tal lines for the four RCP scenarios indicate the likely ranges and medi-

ans for these scenarios as given in Table 13.5 of Chapter 13.

Note that in Figure SPM.9, projected time series are shown only for

RCP2.6 and RCP8.5. Figure TS.22 include time series for all four RCP

scenarios.

Projected contributions to sea level rise in 2081–2100 relative to

1986–2005 for the four RCP scenarios are provided in Figure TS.21.

TS.SM.10 Notes and Technical Details on the

Summary for Policymakers Figure

Plotting Global Mean Temperature

Increase as a Function of Cumulative

Total Global CO

Emissions – Figure

SPM.10

Figure SPM.10 contains data from CO

only simulations and the RCP

simulations. This ﬁgure is closely derived from TS TFE.8, Figure 1. CO

only simulations are represented by grey-shaded patches and thin

black lines, RCP data by coloured lines and patches. CMIP5 results are

taken from the archive as of March 15, 2013. Note that the thick black

line represents the historical time period of the RCP runs.

TS.SM.10.1 Part A – CO

Only Runs

The thin black line represents the multi-model mean of the decadal

averaged global-mean temperature response of the models listed in

Table TS.SM.7 to a global 1% CO

only forcing increase as performed

as part of CMIP5, as a function of the decadal averaged global-mean

diagnosed carbon emissions.

The dark grey patch represents the 90% range surrounding the dec-

adal averaged model response of the CMIP5 models listed in Table

TS.SM.7 and is calculated as follows: Diagnosed carbon emissions and

temperature response data of the above-deﬁned CMIP5 models (com-

puted as in Gillett et al., 2013) is scaled, respectively, by dividing by the

standard deviation over all available decadal-averaged data points for

a speciﬁc scenario. The 90% range is computed in polar coordinates.

The radius stretches along the x-axis (cumulative emissions) and the

angle is the one between the slope from (0, 0) to a respective scaled

point (cumulative emissions, temperature anomaly) and the x-axis. For

each scaled point the radius and angle are computed. A number of

n (n = 20) segments are deﬁned by regularly spaced steps along the

maximum radius of all available decadal-averaged data points of a

speciﬁc scenario (scaled as described earlier). From all points that fall

within the boundaries of each respective radius segment, the 5th and

95th percentiles in terms of available angles is computed. These per-

centiles are then assigned to the radius corresponding to the middle

of the current radius segment. Each of these mid-segment radii and its

corresponding pair of angles are then transformed back to Cartesian

coordinates. Finally, the 90% range is drawn by connecting all 5th and

95th percentile points of a speciﬁc scenario in a hull.

Model Ensemble Member

GFDL-ESM2G r1i1p1

INM-CM4 r1i1p1

GFDL-ESM2M r1i1p1

IPSL-CM5B-LR r1i1p1

BCC-CSM1.1 r1i1p1

MPI-ESM-MR r1i1p1

IPSL-CM5A-MR r1i1p1

IPSL-CM5A-LR r1i1p1

MPI-ESM-LR r1i1p1

NorESM1-ME r1i1p1

CESM1(BGC) r1i1p1

HadGEM2-ES r1i1p1

MIROC-ESM r1i1p1

CanESM2 r1i1p1

BNU-ESM r1i1p1

Table TS.SM.7 | Models that were included in the shown results of the CO2 only 1%

increase CMIP5 runs (dark grey patch and thin black line).

TSSM

Technical Summary Supplementary Material

TSSM-16

TS.SM.10.2 Part B – RCP Runs

Data of the RCP runs (coloured lines and patches) is prepared with

the same methodology as the data for the CO

only runs as described

in the previous section. Note that markers show decadal time steps,

and that the labels in Figure SPM.10 (and TS TFE.8, Figure 1) denote

the cumulative global carbon emissions from 1870 until (but not

including) that year (i.e., label 2050 is placed next to the marker of the

2040–2049 decade). The 90% range is computed for n (n = 12) regu-

larly spaced steps along the maximum radius available for each RCP

(scaled as described earlier). Available Earth System Models (ESM) for

the respective RCP are listed in Table TS.SM.8, available Earth System

Models of Intermediate Complexity (EMIC) in Table TS.SM.9.

Following operations are carried out onto the data:

• Decadal means of global-mean temperature change are computed

relative to the 1861–1880 base period.

• Emissions from the ESMs for the different scenarios are computed

as in Jones et al. (2013).

Model Ensemble Member RCP2.6 RCP4.5 RCP6.0 RCP8.5

BCC-CSM1.1 r1i1p1 x* x* x* x*

CanESM2 r1i1p1 x x x

CESM1(BGC) r1i1p1 x x

GFDL-ESM2G r1i1p1 x x x x

GFDL-ESM2M r1i1p1 x x x x

HadGEM2-CC r1i1p1 x x

HadGEM2-ES r2i1p1 x x x x

INM-CM4 r1i1p1 x* x*

IPSL-CM5A-LR r1i1p1 x x x x

IPSL-CM5A-MR r1i1p1 x x x

IPSL-CM5B-LR r1i1p1 x x

MIROC-ESM r1i1p1 x x x x

MIROC-ESM-CHEM r1i1p1 x x x x

MPI-ESM-LR r1i1p1 x x x

NorESM1-ME r1i1p1 x x x x

• Land-use change emission estimated for each RCP are added to

all EMICs, and to the ESMs that diagnose fossil-fuel emission only

(see Table TS.SM.8). Land-use change emissions are obtained from

http://www.pik-potsdam.de/~mmalte/rcps/ for each RCP, respec-

tively. Note that the data for Figure SPM.10 provided at the exter-

nal sources cited above may be subject to changes in the future by

the owners. Furthermore, no guarantee is provided that the web-

links cited above remain active.

• Decadal-mean cumulative emissions are computed from cumula-

tive carbon emissions relative to 1870.

• Each RCP range is drawn as long as data is available for all models

or until temperatures have peaked. The encompassing range shown

in Figure SPM.10 (and TS TFE.8, Figure 1) is constructed by con-

necting the outer last points of each single RCP range and is ﬁlled

as long as data are available for all models for RCP8.5. Beyond

this point, the range is illustratively extended by further progressing

along the radius while keeping the angles ﬁxed at those available

at the last point with data from all models for RCP8.5. The fading

out of the range is illustrative.

Model RCP2.6 RCP4.5 RCP6.0 RCP8.5

Bern3D x x x x

DCESS x x x x

GENIE x x x x

IGSM x x x x

UVic x x x x

Table TS.SM.8 | Overview of RCP model runs available in the CMIP5 archive, as used in Figure SPM.10 (and TS TFE.8, Figure 1).

Table TS.SM.9 | Overview of EMIC RCP model runs from (Eby et al. 2013; Zickfeld et al. 2013), as used in Figure SPM.10 (and TS TFE.8, Figure 1). EMICs output is available from

http://www.climate.uvic.ca/EMICAR5.

Notes:

* runs do not include explicit land-use change modelling. Models diagnose fossil-fuel and land-use change emissions jointly and therefore do not require adding land-use change emissions.

TSSM

Technical Summary Supplementary Material

TSSM-17

References

Antonov, J. I., et al., 2010: World Ocean Atlas 2009, Volume 2: Salinity [Levitus, S.

(Ed.)]. NOAA Atlas NESDIS 69, 184 pp.

Bates , N .R ., 2007: Interannual variability of the oceanic CO

sink in the subtropical

gyre of the North Atlantic Ocean over the last two decades. J. Geophys. Res.

Oceans, 112, C09013.

Becker, A., et al., 2013: A description of the global land-surface precipitation

data products of the Global Precipitation Climatology Centre with sample

applications including centennial (trend) analysis from 1901–present. Earth Sys.

Sci. Data, 5, 71–99.

Brown, R. D., and D. A. Robinson, 2011: Northern Hemisphere spring snow cover

variability and change over 1922–2010 including an assessment of uncertainty.

Cryosphere, 5, 219–229.

Cavalieri, D. J., and C. L. Parkinson, 2012: Arctic sea ice variability and trends, 1979-

2010. Cryosphere, 6, 957–979.

Cavalieri, D. J., P. Gloersen, and W. J. Campbell, 1984: Determination of sea ice

parameters with the Nimbus-7 SMMR. J. Geophys. Res. Atmos., 89, 5355–5369.

Church, J. A., and N. J. White, 2011: Sea-Level Rise from the Late 19th to the Early

21st Century. Surv. Geophys., 32, 585–602.

Comiso, J. C., and F. Nishio, 2008: Trends in the sea ice cover using enhanced and

compatible AMSR-E, SSM/I, and SMMR data. J. Geophys. Res. Oceans, 113,

C02S07.

Dickson, A.G., C. L. Sabine, and J. R. Christian, J.R., (eds.), 2007: Guide to best

practices for ocean CO

measurements. PICES Special Publication 3, 191 pp.

Domingues, C. M., J. A. Church, N. J. White, P. J. Gleckler, S. E. Wijffels, P. M. Barker,

and J. R. Dunn, 2008: Improved estimates of upper-ocean warming and multi-

decadal sea-level rise. Nature, 453, 1090–1094.

Dore, J. E., R. Lukas, D. W. Sadler, M. J. Church, and D. M. Karl, 2009: Physical and

biogeochemical modulation of ocean acidiﬁcation in the central North Paciﬁc.

Proc. Natl. Acad. Sci. U.S.A., 106, 12235–12240.

Eby, M., et al., 2013: Historical and idealized climate model experiments: an

intercomparison of Earth system models of intermediate complexity. Clim. Past,

9, 1111–1140.

Garcia, H. E., R. A. Locarnini, T. P. Boyer, J. I. Antonov, O. K. Baranova, M. M. Zweng,

and D. R. Johnson, 2010: World Ocean Atlas 2009, Volume 3: Dissolved Oxygen,

Apparent Oxygen Utilization, and Oxygen Saturation [ Levitus, S. (Ed.)]. NOAA

Atlas NESDIS 70, 344 pp.

Gillett, N. P., V. K. Arora, D. Matthews, and M. R. Allen, 2013: Constraining the Ratio

of Global Warming to Cumulative CO

Emissions Using CMIP5 Simulations. J.

Clim., 26, 6844–6858.

González-Dávila, M., and J. M. Santana-Casiano, 2009: Sea Surface and Atmospheric

fCO

data measured during the ESTOC Time Series cruises from 1995–2009. Oak

Ridge National Laboratory, US Department of Energy, Oak Ridge, Tennessee.

http://cdiac.ornl.gov/ftp/oceans/ESTOC_data/.

González-Dávila, M., J. M. Santana-Casiano, J. M. Rueda, and O. Llinás, 2010: Water

column distribution of the carbonate system variables in the ESTOC site from

1995 to 2004. Biogeosciences, 7, 3067–3081.

IPCC, 2012: Managing the Risks of Extreme Events and Disasters to Advance

Climate Change Adaptation. A Special Report of Working Groups I and II of the

Intergovernmental Panel on Climate Change [Field, C.B., et al., (eds.)]. Cambridge

University Press, Cambridge, UK, and New York, NY, USA, 582 pp.

Ishii, M., and M. Kimoto, 2009: Reevaluation of historical ocean heat content

variations with time-varying XBT and MBT depth bias corrections. J. Oceanogr.,

65, 287–299.

Jevrejeva, S., J. C. Moore, A. Grinsted, and P. L. Woodworth, 2008: Recent global sea

level acceleration started over 200 years ago? Geophys. Res. Lett., 35, L08715.

Jones, C., et al., 2013: Twenty-First Century Compatible CO

Emissions and Airborne

Fraction Simulated by CMIP5 Earth System Models under Four Representative

Concentration Pathways. J. Clim., 26, 4398–4413.

Keeling, C., R. Bacastow, A. Bainbridge, C. Ekdahl, P. Guenther, L. Waterman, and J.

Chin, 1976a: Atmospheric Carbon-Dioxide Variations at Mauna-Loa Observatory,

Hawaii. Tellus, 28, 538–551.

Keeling, C. D., J. A. Adams, and C. A. Ekdahl, 1976b: Atmospheric Carbon-Dioxide

Variations at South Pole. Tellus, 28, 553–564.

Keeling, C. D., S. C. Piper, R. B. Bacastow, M. Wahlen, T. P. Whorf, M. Heimann, and

H. A. Meijer, 2001: Exchanges of atmospheric CO

and

with the terrestrial

biosphere and oceans from 1978 to 2000. I. Global aspects, SIO Reference

Series, No. 01–06. Scripps Institution of Oceanography, San Diego, 88 pp.

Key, R. M., et al., 2004: A global ocean carbon climatology: Results from Global Data

Analysis Project (GLODAP). Glob. Biogeoch. Cycles, 18, GB4031.

Le Quéré, C., et al., 2013: The global carbon budget 1959–2011. Earth Syst. Sci. Data,

5, 165–185.

Levitus, S., et al., 2012: World ocean heat content and thermosteric sea level change

(0-2000 m), 1955–2010. Geophys. Res. Lett., 39, L10603.

Locarnini, R. A., et al., 2010: World Ocean Atlas 2009, Volume 1: Temperature

[Levitus, S. (Ed.)]. NOAA Atlas NESDIS 68, 184 pp.

Markus, T., and D. J. Cavalieri, 2000: An enhancement of the NASA Team sea ice

algorithm. IEEE Trans. Geosci. Remote Sens., 38, 1387–1398.

Massonnet, F., T. Fichefet, H. Goosse, C. M. Bitz, G. Philippon-Berthier, M. M. Holland,

and P.-Y. Barriat, 2012: Constraining projections of summer Arctic sea ice.

Cryosphere, 6, 1383–1394.

Mitchell, T. D., and P. D. Jones, 2005: An improved method of constructing a database

of monthly climate observations and associated high-resolution grids. Int. J.

Climatol., 25, 693–712.

Morice, C. P., J. J. Kennedy, N. A. Rayner, and P. D. Jones, 2012: Quantifying

uncertainties in global and regional temperature change using an ensemble of

observational estimates: The HadCRUT4 data set. J. Geophys. Res. Atmos., 117,

22.

Nerem, R. S., D. P. Chambers, C. Choe, and G. T. Mitchum, 2010: Estimating Mean

Sea Level Change from the TOPEX and Jason Altimeter Missions. Mar. Geod.,

33, 435–446.

Orr, J.C. et al., 2005: Anthropogenic ocean acidiﬁcation over the twenty-ﬁrst century

and its impact on calcifying organisms. Nature, 437, 681-686.

Orr, J. C., 2011: Recent and future changes in ocean carbonate chemistry. In: Ocean

Acidiﬁcation [Gattuso, S.-P., and L. Hansson (eds.)]. Oxford University Press,

Oxford, United Kingdom and New York, NY, USA, 352 pp.

Palmer, M. D., K. Haines, S. F. B. Tett, and T. J. Ansell, 2007: Isolating the signal of

ocean global warming. Geophys. Res. Lett., 34, L23610.

Parkinson, C. L., and D. J. Cavalieri, 2012: Antarctic Sea Ice Variability and Trends,

1979-2010. Cryosphere, 6, 871–880.

Ray, R. D., and B. C. Douglas, 2011: Experiments in reconstructing twentieth-century

sea levels. Prog. Oceanogr., 91, 496–515.

Rayner, N. A., et al., 2003: Global analyses of sea surface temperature, sea ice, and

night marine air temperature since the late nineteenth century. J. Geophys. Res.

Atmos., 108, 4407.

Schweiger, A., R. Lindsay, J. L. Zhang, M. Steele, H. Stern, R. Kwok, 2011: Uncertainty

in modeled Arctic sea ice volume. J. Geophys. Res. Oceans, 116, C00D06.

Smith, D. M., and J. M. Murphy, 2007: An objective ocean temperature and salinity

analysis using covariances from a global climate model. J. Geophys. Res. Oceans,

112, C02022.

Thoning, K. W., P. P. Tans, and W. D. Komhyr, 1989: Atmospheric carbon dioxide at

Mauna Loa Observatory 2. Analysis of the NOAA GMCC data, 1974-1985. J.

Geophys. Res., 94, 8549–8565.

Vose, R. S., Oak Ridge National Laboratory. Environmental Sciences Division.,

U.S. Global Change Research Program, United States. Dept. of Energy. Ofﬁce

of Health and Environmental Research., Carbon Dioxide Information Analysis

Center (U.S.), and Martin Marietta Energy Systems Inc., 1992: The global

historical climatology network: long-term monthly temperature, precipitation,

sea level pressure, and station pressure data. Carbon Dioxide Information

Analysis Center. Available to the public from N.T.I.S., 1 v.

Walsh, J. E., and W. L. Chapman, 2001: 20th-century sea-ice variations from

observational data. Ann. Glaciol., 33, 444–448.

Zickfeld, K., et al., 2013: Long-Term Climate Change Commitment and Reversibility:

An EMIC Intercomparison. J. Clim., 26, 5782–5809.

Please note that all external web-links cited in this document were active at the time

of publication, but no guarantee is provided that these links remain active.