62
Consideration of these three additional ocean proxies improves the simulation of ΔT
around year 1910 and in the mid-1940s (Fig. 2.5) compared to the results shown in Fig.
2.4, which lacked these terms. Most of this improvement is due to the use of AMV as
a proxy for variations in the strength of the Atlantic Meridional Overturning Circulation,
which only recently has been recognized as having a considerable effect on global cli-
mate (Schlesinger and Ramankutty 1994 ; Andronova and Schlesinger 2000 ). In our
approach, the PDO (Zhang et al. 1997 ) and the IOD (Saji et al. 1999 ) have little expres-
sion on global climate, which is a common finding using MLR analysis of the ~150
year long record of ΔT (Rypdal 2015 ; Chylek et al. 2014 ). Also, upon inclusion of the
AMV proxy (Fig. 2.5), the cooling after major volcanic eruptions is diminished by
nearly a factor of two relative to a MLR analysis that neglects this term (volcanic term
in Fig. 2.5 compared to volcanic term in Fig. 2.4). This finding could have significant
implications for the use of volcanic cooling as a proxy for the efficacy of geo-
engineering of climate via stratospheric sulfate injection (Canty et al. 2013 ).
Additional detail on inputs to the Empirical Model of Global Climate is provided
in Sect. 2.2.1.1. More explanation of the model outputs is given in Sect. 2.2.1.2.
Both of these sections are condensed from our model description paper (Canty et al.
2013 ), including a few updates since the original publication.
2.2.1.1 Model Inputs
The ΔRF due to GHGs is based on global, annual mean mixing ratios of CO 2 , CH 4 ,
N 2 O, the class of halogenated compounds known as ozone depleting substances
(ODS), HFCs, PFCs, SF 6 , and NF 3 (Other F-gases) provided by the RCP 4.5 (Thomson
et al. 2011 ) and RCP 8.5 (Riahi et al. 2011 ) scenarios. Annual abundances are inter-
polated to a monthly time grid, because monthly resolution is needed to resolve short-
term impacts on ΔT of processes such as ENSO and volcanic eruptions. Values of
ΔRF for each GHG are computed using formula originally given in Table 6.2 of IPCC
( 2001 ) except the pre-industrial value of CH 4 has been adjusted to 0.722 ppm, follow-
ing Table AII.1.1a of (IPCC 2013 ). The ΔRF due to tropospheric O 3 is based on the
work of Meinshausen et al. ( 2011 ), obtained from a file posted at the Potsdam Institute
for Climate Impact Research website. The sum of ΔRF due to CO 2 , CH 4 , N 2 O, ODS,
Other F-gases, and tropospheric O 3 constitutes GHG ΔRFi in Eq. 2.2.
The ΔRF due to aerosols is the sum of direct and indirect effects of six types of
aerosols, as described in Sect. 3.2.2 of Canty et al. ( 2013 ). The six aerosol types are
sulfate, mineral dust, ammonium nitrate, fossil fuel organic carbon, fossil fuel black
carbon, and biomass burning emissions of organic and black carbon. The direct ΔRF
for all aerosol types other than sulfate is also based on the work of Meinshausen et al.
( 2011 ), again obtained from files posted at the Potsdam Institute for Climate Impact
Research website. Different estimates for RCP 4.5 and RCP 8.5 are used, since it is
assumed that reduction of atmospheric release of aerosol precursors will occur more
quickly in RCP 4.5, in lock-step with the decreased emission of GHGs in this scenario
relative to RCP 8.5. The direct RF due to sulfate is based on the work of Smith
et al. ( 2011 ). Scaling parameters are used to multiply the direct ΔRF of aerosols, to
account for the aerosol indirect effect, as described in Sect. 3.2.2 of Canty et al. ( 2013 ).
2 Forecasting Global Warming