85Analysis of the GISS and NCEI data sets leads to a different conclusion. As
shown in Fig. 2.14b, c, the observed and modeled slope of ΔT, for 1998–2012,
agree extremely well. The GISS record of GMST is based on the same SST record
used by NCEI. Earlier versions of the NCEI record (not shown), released prior to
the update in SST described by Karl et al. ( 2015 ), did support the notion that some
unknown factor was suppressing the rise in ΔT from 1998 to 2012.
Cowtan and Way ( 2014 ) (hereafter, CW2014) suggest the existence of a recent,
cool bias in the CRU estimate of ΔT, due to closure of observing stations in the
Arctic and Africa that they contend has not been handled properly in the official
CRU data release. CW2014 published two alternate versions of the CRU data set,
termed “kriging” and “hybrid”, to account for the impact of these station closures
on ΔT. Figure 2.14d shows that, upon use of the CRU-Hybrid data set of CW2014,
the observed and modeled slope of ΔT are in excellent agreement. Similarly good
agreement between measured and modeled ΔT is obtained for CRU-Kriging (not
shown). It remains to be seen whether CW2014 will impact future versions of ΔT
from CRU. In the interim, the CW2014 analysis supports the finding, from the GISS
and NCEI data sets, that there was no hiatus in the gradual, long-term rise of ΔT.
The EM-GC allows us to extract AAWR for any period of time. For the simulations
shown in four panels of Fig. 2.14, the values of AAWR for 1998–2012 are 0.1075 ±
0.0041, 0.1186 ± 0.004, 0.1089 ± 0.0046, and 0.1039 ± 0.004, respectively, all in units
of °C/decade. The primary factors responsible for the slightly smaller rise in ΔT (black
numbers, Fig. 2.14) compared to AAWR over 1998–2012 is the tendency of the climate
system to be in a more La Niña like state during the latter half of this period of time^31
(Kosaka and Xie 2013 ) and a relatively small value of total solar irradiance during the
most recent solar max cycle (Coddington et al. 2016 ). Our simulations, which include
Kasatochi, Sarychev and Nabro, suggest these recent minor volcanic eruptions played
only a miniscule role (~0.0018 °C/decade cooling) over this period. We conclude
human activity exerted about 0.11 °C/decade warming over 1998–2012, and observa-
tions show a rise of ΔT that is slightly smaller in magnitude, due to natural factors that
are well characterized by the Empirical Model of Global Climate.
2.5 Future Temperature Projections
Accurate projections of the expected future rise of GMST are central for the suc-
cessful implementation of the Paris Climate Agreement. As shown in Sect. 2.2.1.3,
the degeneracy of the climate system coupled with uncertainty in ΔRF due to tropo-
spheric aerosols leads to considerable spread in projections of ΔT (the anomaly of
(^31) This is not particularly surprising given the strong ENSO of 1998. Hindsight is 20:20, but it is
nonetheless remarkable how much attention has been devoted to discussion of ΔT over the 1998–
2012 time period, including within IPCC ( 2013 ), given the unusual climatic conditions known to
have occurred at the start of this time period. Apparently the global warming deniers took the lead
in promulgating the notion that more than a decade had passed without a discernable rise in ΔT,
and the scientific community took that bait and devoted enormous resources to examination of
GMST over this particular 15-year interval.
2.5 Future Temperature Projections