77third that examined Earth’s energy budget in detail over various decadal periods
(Otto et al. 2013 ). The right hand side of Fig. 2.11 shows ECS found using our
EM-GC framework, for the six estimates of OHC that appear in Fig. 2.8.
Figure 2.11 shows that published values of ECS from GCMs (average of the
three best estimates is 3.5 °C) are considerably larger than estimates of ECS from
the actual climate record. This pattern holds upon comparison of GCM-based ECS
to values found using empirically-based estimates of ECS found by other research
groups (mean value 2.1 °C) and using our EM-GC framework (mean value 1.6 °C).
These three estimates of ECS are important for policy. The mean value of ECS
from GCMs (3.5 °C), taken literally and ignoring changes in other GHGs, indicates
CO 2 must be kept far short of the 2 × pre-industrial level to achieve the Paris upper
limit of 2 °C warming. The mean of the three empirically based estimates of ECS
from other groups (2.1 °C) suggests the Paris upper limit can perhaps be achieved if
the rise of CO 2 can be arrested before reaching the 2 × pre-industrial level, whereas
the mean value ECS from our EM-GC framework (1.6 °C) suggests that if society
manages to keep CO 2 from reaching 2 × pre-industrial level, the Paris goal might be
achieved. Of course, these statements are all contingent on minimal future growth
of other GHGs. Also, we stress that all of the estimates of ECS, even those from our
EM-GC framework, are associated with considerable uncertainty. Nonetheless, the
various ECS estimates in Fig. 2.11 suggest climate feedback within GCMs is larger
than in the actual climate system,^22 which would explain the tendency for so many
CMIP5 GCM projections of ΔT to lie above the green trapezoid in Fig. 2.3.
The tendency of CMIP5 GCMs to warm too quickly, with respect to the actual
human influence on ΔT, is probed further in Sect. 2.3. This shortcoming of the
CMIP5 GCMs is crucial to the thesis of this book: that the Paris Climate Agreement,
as presently formulated, could actually limit the growth of GMST to less than 2 °C
above pre-industrial.
2.3 Attributable Anthropogenic Warming Rate
The most important metric for a climate model is how well the prior rise in global
mean surface temperature can be simulated. The green trapezoid used in various
figures throughout this chapter is based on the recognition, by Chap. 11 of IPCC
( 2013 ), that CMIP5 GCMs have warmed too aggressively compared to observations
over the prior several decades. In this section, the Empirical Model of Global
Climate is used to quantify the amount of global warming that can be attributed to
humans, over the time period 1979–2010.^23 These years are chosen because the rise
in ΔT is nearly linear over this interval and this period has been the basis of similar
examination by several other studies (Foster and Rahmstorf 2011 ; Zhou and Tung
2013 ). Our analysis of ΔT is compared to simulations of this quantity provided by
CMIP5 GCMs, and to other analyses of ΔT over this period of time.
(^22) Most estimates of ECS, such as Eq. 2.6, show ECS to be solely a function of climate feedback.
(^23) Specifically all analyses in this section span the start of 1979 to the end of 2010.
2.3 Attributable Anthropogenic Warming Rate