89will occur. In other words, the most probable value of ΔT in year 2060, for the
EM-GC projection that uses RCP 4.5, is 1.2–1.3 °C above pre-industrial, and there
is slightly less than 20 % probability ΔT will actually fall within this range. In con-
trast, the CMIP5 GCMs project ΔT in 2060 will most probably be 2.0–2.2 °C
warmer than pre-industrial, with a ~12 % probability ΔT will actually fall within
this range. A finer spacing for ΔT is used for the EM-GC projection, since we are
a
b
RCP 4.5
RCP 8.5
Paris Upper Limit
Paris TargetFig. 2.17 Probability distribution functions of rise in GMST in year 2060. The line segments
represent a series of histograms (narrow, vertical rectangles) for projections of ΔT in year 2060
relative to the pre-industrial baseline found using our EM-GC (blue) and CMIP5 GCMs (red). The
height of each histogram represents the probability the rise of ΔT will fall within the range of ΔT
that corresponds to the ends of each line segment (see main text). The Paris Climate Agreement
target and upper limit of 1.5 and 2.0 °C warming are denoted. Projections of ΔT 2060 found using the
EM-GC consider only combinations of model parameters λ and AerRF 2011 that fall within the
respective ellipse of Fig. 2.17 (i.e., projections consider only acceptable fits to the climate record)
and the EM-GC values of ΔT 2060 are weighted by 1/χ^2 , so that simulations that provide a better fit
to the climate record are given more credence. Finally, the EM-GC simulations used OHC from the
average of six data records shown in Fig. 2.8 and the aerosol ΔRF time series based on scaling
parameters along the middle road of Fig. 2.21. (a) EM-GC and CMIP5 GCM projections based on
RCP 4.5 (Thomson et al. 2011 ); the GCM projections consider the 41 models represented in Fig.
2.3a ; (b) EM-GC and CMIP5 GCM projections based on RCP 8.5 (Riahi et al. 2011 ); the GCM
projections consider the 38 models represented in Fig. 2.3b
2.5 Future Temperature Projections