SLIDE 1 Cristian Proistosescu1 Kyle Armour1 Gerard Roe1 Peter Huybers2
1University of Washington 2Harvard University
AGU Fall Meeting 2017
Courtesy of NASA’s Earth Observatory
When are we committed to crossing critical (1.5 or 2 °C) temperature thresholds?
SLIDE 2 Two questions
- 1. When will we cross 1.5 or 2 °C global warming thresholds (e.g., following
high or low emission scenarios) – subject to constraints from the observed global energy budget?
- 2. When will we be geophysically committed to crossing 1.5 or 2 °C global
warming thresholds?
SLIDE 3
What do CMIP5 models say?
CMIP5 projections (IPCC AR5)
2 °C
SLIDE 4
What do CMIP5 models say?
CMIP5 projections (IPCC AR5)
§ Models may not agree with observed global warming and energy budget constraints § Models may not span full range of plausible future warming § Computationally expensive to run different emissions scenarios, so can’t ask questions like, when are we geophysically committed to 2 °C? § Not clear which physical factors are contributing to uncertainty in projected warming
2 °C
SLIDE 5
Our approach
upper ocean Tu deep ocean Td
cu dTu dt = Tu + F + "(Td Tu) ( cd dTd dt = (Tu Td) ( + F +
radiative forcing
= Tu +
radiative response
= (Tu Td) (
Ocean heat uptake efficacy
§ Use a 2-layer ocean model (Held et al. 2010; Armour 2017) that includes the essential physics governing global-mean surface warming:
SLIDE 6
50 100 150 2 4 6
Our approach
cu dTu dt = Tu + F + "(Td Tu) ( cd dTd dt = (Tu Td) (
Temperature change T (°C) Global surface temperature response to abrupt CO2 quadrupling Year after CO2 quadrupling
50 100 150 1 2 3
Fast warming on timescale of the surface Slow warming on timescale of the deep ocean
§ Use a 2-layer ocean model (Held et al. 2010; Armour 2017) that includes the essential physics governing global-mean surface warming:
SLIDE 7
50 100 150 2 4 6
Our approach
cu dTu dt = Tu + F + "(Td Tu) ( cd dTd dt = (Tu Td) (
Temperature change T (°C) Global surface temperature response to abrupt CO2 quadrupling Year after CO2 quadrupling
50 100 150 1 2 3
§ Use a 2-layer ocean model (Held et al. 2010; Armour 2017) that includes the essential physics governing global-mean surface warming:
SLIDE 8 Our approach
Step 1: Draw priors of , , , and from fits
- f 2-layer model to CMIP5 model response to
CO2 forcing (Geoffroy et al. 2013)
cu dTu dt = Tu + F + "(Td Tu) ( cd dTd dt = (Tu Td) (
cu T
cd (
+ "
§ Use a 2-layer ocean model (Held et al. 2010; Armour 2017) that includes the essential physics governing global-mean surface warming:
SLIDE 9 Our approach
Step 1: Draw priors of , , , and from fits
- f 2-layer model to CMIP5 model response to
CO2 forcing (Geoffroy et al. 2013) Step 2: Drive model with timeseries of historical radiative forcing (Meinshausen et al. 2011), with priors drawn from forcing range in IPCC AR5
cu dTu dt = Tu + F + "(Td Tu) ( cd dTd dt = (Tu Td) (
§ Use a 2-layer ocean model (Held et al. 2010; Armour 2017) that includes the essential physics governing global-mean surface warming:
cu T
cd (
+ "
SLIDE 10 Our approach
Step 1: Draw priors of , , , and from fits
- f 2-layer model to CMIP5 model response to
CO2 forcing (Geoffroy et al. 2013) Step 2: Drive model with timeseries of historical radiative forcing (Meinshausen et al. 2011), with priors drawn from forcing range in IPCC AR5 Step 3: Use Bayesian inference to estimate posterior parameters/forcings based on
- bserved warming and energy budget (see
also: Forest et al. 2002, 2006; Stott & Forest 2007) § Use a 2-layer ocean model (Held et al. 2010; Armour 2017) that includes the essential physics governing global-mean surface warming:
cu dTu dt = Tu + F + "(Td Tu) ( cd dTd dt = (Tu Td) (
= 0.75 ± 0.2 °C = 0.65 ± 0.27 Wm-2 = 2.3 ± 1 Wm-2
(Otto et al. 2013; 2000-2009 relative to 1860-1879)
⇥Tobs
− Qobs =
⇥
Fobs
cu T
cd (
+ "
SLIDE 11 Our approach
Step 1: Draw priors of , , , and from fits
- f 2-layer model to CMIP5 model response to
CO2 forcing (Geoffroy et al. 2013) Step 2: Drive model with timeseries of historical radiative forcing (Meinshausen et al. 2011), with priors drawn from forcing range in IPCC AR5 Step 3: Use Bayesian inference to estimate posterior parameters/forcings based on
- bserved warming and energy budget (see
also: Forest et al. 2002, 2006; Stott & Forest 2007) = 0.75 ± 0.2 °C = 0.65 ± 0.27 Wm-2 = 2.3 ± 1 Wm-2
(Otto et al. 2013; 2000-2009 relative to 1860-1879)
⇥Tobs
− Qobs =
⇥
Fobs
cu T
cd (
+ "
Year 2-layer model historical warming Temperature change [°C]
SLIDE 12 Our approach
Step 1: Draw priors of , , , and from fits
- f 2-layer model to CMIP5 model response to
CO2 forcing (Geoffroy et al. 2013) Step 2: Drive model with timeseries of historical radiative forcing (Meinshausen et al. 2011), with priors drawn from forcing range in IPCC AR5 Step 3: Use Bayesian inference to estimate posterior parameters/forcings based on
- bserved warming and energy budget (see
also: Forest et al. 2002, 2006; Stott & Forest 2007) Step 4: Use the observationally-constrained parameter/forcing estimates to project warming, and committed warming, following RCP2.6 and RCP8.5 emissions scenarios
cu T
cd (
+ "
Temperature change [°C] Year 2-layer model historical warming
SLIDE 13 Our approach
Step 1: Draw priors of , , , and from fits
- f 2-layer model to CMIP5 model response to
CO2 forcing (Geoffroy et al. 2013) Step 2: Drive model with timeseries of historical radiative forcing (Meinshausen et al. 2011), with priors drawn from forcing range in IPCC AR5 Step 3: Use Bayesian inference to estimate posterior parameters/forcings based on
- bserved warming and energy budget (see
also: Forest et al. 2002, 2006; Stott & Forest 2007) Step 4: Use the observationally-constrained parameter/forcing estimates to project warming, and committed warming, following RCP2.6 and RCP8.5 emissions scenarios
cu T
cd (
+ "
Year 2-layer model projections Temperature change [°C]
SLIDE 14 Our approach
Step 1: Draw priors of , , , and from fits
- f 2-layer model to CMIP5 model response to
CO2 forcing (Geoffroy et al. 2013) Step 2: Drive model with timeseries of historical radiative forcing (Meinshausen et al. 2011), with priors drawn from forcing range in IPCC AR5 Step 3: Use Bayesian inference to estimate posterior parameters/forcings based on
- bserved warming and energy budget (see
also: Forest et al. 2002, 2006; Stott & Forest 2007) Step 4: Use the observationally-constrained parameter/forcing estimates to project warming, and committed warming, following RCP2.6 and RCP8.5 emissions scenarios
cu T
cd (
+ "
Year 2-layer model projections Temperature change [°C]
SLIDE 15 Our approach
Year 2-layer model projections Step 1: Draw priors of , , , and from fits
- f 2-layer model to CMIP5 model response to
CO2 forcing (Geoffroy et al. 2013) Step 2: Drive model with timeseries of historical radiative forcing (Meinshausen et al. 2011), with priors drawn from forcing range in IPCC AR5 Step 3: Use Bayesian inference to estimate posterior parameters/forcings based on
- bserved warming and energy budget (see
also: Forest et al. 2002, 2006; Stott & Forest 2007) Step 4: Use the observationally-constrained parameter/forcing estimates to project warming, and committed warming, following RCP2.6 and RCP8.5 emissions scenarios
cu T
cd (
+ "
Temperature change [°C]
SLIDE 16
Our approach
Year 2-layer model projections CMIP5 projections (IPCC AR5) Temperature change [°C]
SLIDE 17
When are we going to cross 1.5 or 2 °C thresholds?
Temperature change [°C] Year 2-layer model projections Probability density [1/yr] Year Year at which 2 °C is crossed
SLIDE 18
When are we going to cross 1.5 or 2 °C thresholds?
Temperature change [°C] Year 2-layer model projections Probability density [1/yr] Year Year at which 2 °C is crossed
Probability of crossing 2.0 °C along RCP 2.6 = 0.13
SLIDE 19
When are we going to cross 1.5 or 2 °C thresholds?
Year 2-layer model projections Temperature change [°C] Probability density [1/yr] Year Year at which 1.5 °C is crossed
Probability of crossing 1.5 °C along RCP 2.6 = 0.41
SLIDE 20 Zero-emissions climate commitment
§ How much more warming will occur given no further human influence on climate?
§ Constant atmospheric composition requires continued emissions; the climate commitment is better defined with respect to past emissions only
1800 1900 2000 2100 2200 2300 Year 1.5 1.0 0.5 0.0 Global temperature change ( °C) Constant composition Zero emissions This study IPCC AR4 models HadCM3LC model BERN2.5CC model
(Matthews and Weaver 2010)
SLIDE 21 Zero-emissions climate commitment
§ How much more warming will occur given no further human influence on climate?
§ Constant atmospheric composition requires continued emissions; the climate commitment is better defined with respect to past emissions only
1800 1900 2000 2100 2200 2300 Year 1.5 1.0 0.5 0.0 Global temperature change ( °C) Constant composition Zero emissions This study IPCC AR4 models HadCM3LC model BERN2.5CC model
Zero-emissions climate commitment assuming cessation of CO2 emissions
- nly, temperatures stay flat or even
decline (see also: Solomon et al. 2009)
(Matthews and Weaver 2010)
§ We are not committed to any more warming in the pipeline § … but what about non-CO2 forcing agents?
SLIDE 22 Zero-emissions climate commitment
Radiative forcing [Wm-2] Year 2-layer model forcing Zero-emissions climate commitment assuming cessation of CO2 emissions
- nly, temperatures stay flat or even
decline (see also: Solomon et al. 2009) § We are not committed to any more warming in the pipeline § … but what about non-CO2 forcing agents?
SLIDE 23 Zero-emissions climate commitment
Radiative forcing [Wm-2] Year 2-layer model forcing
Long lifetime of CO2 (range of carbon cycle parameters from Joos et al. 2013)
Zero-emissions climate commitment assuming cessation of CO2 emissions
- nly, temperatures stay flat or even
decline (see also: Solomon et al. 2009) § We are not committed to any more warming in the pipeline § … but what about non-CO2 forcing agents?
Cessation of emission at 8:22am, Dec 15, 2017 Jump due to loss of tropospheric aerosols
SLIDE 24 Zero-emissions climate commitment
Year 2-layer model temperature response Zero-emissions climate commitment assuming cessation of CO2 emissions
- nly, temperatures stay flat or even
decline (see also: Solomon et al. 2009) § We are not committed to any more warming in the pipeline § … but what about non-CO2 forcing agents? Temperature change [°C]
SLIDE 25
Zero-emissions climate commitment
Year 2-layer model temperature response Temperature change [°C] § Armour & Roe (2011): 0.6 °C (0.3-7.2 °C, 5-95%) based on AR4 values § Mauritsen & Pincus (2017): 1.1 °C (0.7-1.8 °C, 5-95%) based on AR5 values and updated energy budget; 13% that we’re already committed to 1.5 °C § This work: 1.3 °C (0.9-1.9 °C, 5-95%) based on Otto et. al energy budget; 14% that we’re already committed to 1.5 °C, 3% chance that we’re already committed to 2 °C Committed warming (above pre-industrial) inferred from observed warming and global energy budget:
SLIDE 26 When are we committed to crossing 1.5 or 2 °C thresholds?
Probability density [1/yr] Year Year at which 2 °C is crossed
Probability density [1/yr] Year Year at which 1.5 °C is crossed or committed
committed crossed committed crossed
SLIDE 27 When are we committed to crossing 1.5 or 2 °C thresholds?
Probability density [1/yr] Year Year at which 2 °C is crossed
Probability density [1/yr] Year Year at which 1.5 °C is crossed or committed
committed crossed committed crossed 50% chance of being committed to 1.5 °C by 2031 50% chance of being committed to 2°C by 2047
SLIDE 28 When are we committed to crossing 1.5 or 2 °C thresholds?
Probability density [1/yr] Year Year at which 2 °C is crossed
Probability density [1/yr] Year Year at which 1.5 °C is crossed or committed
committed crossed committed crossed
SLIDE 29 Parting thoughts
§ We are committed to crossing warming thresholds up to a couple decades before those temperatures would be reached under ongiong emissions, due to the loss of tropospheric aerosols § Current climate commitment is about 1.3 °C (0.9-1.9 °C, 5-95%), with a14% that we’re already committed to reaching 1.5 °C above pre- industrial § 50% chance of being committed to 1.5 °C by 2031, 2 °C by 2047 following RCP8.5 § Model, constrained by observations of global warming and energy budget, is a great tool for probing sources of uncertainty in future warming and climate commitment (hint: primarily aerosol forcing uncertainty; carbon cycle uncertainty becomes important later in the century under high emissions)
Year Temperature change [°C] Year Probability distribution (1/yr)
Manuscript in preparation: Proistosescu, Armour, Roe & Huybers
