Table 1.

Climate sensitivity estimates obtained by minimizing the RMSE between observed (IPCC) global-mean temperature changes and modeled changes over 1861–1994

Forcing			Raw IPCC data			IPCC data, ENSO removed			Low-pass filtered IPCC data
GHG	Aerosol	Solar	ΔT2×, °C	RMSE, °C	R2	ΔT2×, °C	RMSE, °C	R2	ΔT2×, °C	RMSE, °C	R2
Yes	—	—	1.2	0.128	0.60	1.1	0.119	0.61	1.2	0.093	0.74
Yes	Mid-IPCC	—	6.3	0.121	0.64	6.0	0.114	0.65	6.4	0.083	0.79
Yes	Low	—	2.6	0.125	0.62	2.5	0.116	0.63	2.6	0.088	0.76
Yes	Mid-IPCC	Yes	3.0	0.117	0.67	2.8	0.107	0.69	3.0	0.077	0.82
Yes	Low	Yes	1.8	0.118	0.66	1.7	0.109	0.68	1.8	0.079	0.81

Results are given for five different forcing combinations and three different observed data series. The forcings used are IPCC GHG forcing; zero, central (Mid-IPCC), and “low” aerosol forcing (Mid-IPCC aerosol forcing is −1.3 W/m² to 1990; low is −0.9 W/m², a value that is substantially larger in magnitude than the IPCC lower bound); and solar forcing (33). The observed data cases are raw IPCC annual data; IPCC annual data with the ENSO influence factored out; and low-pass filtered IPCC data. The explained variance (R²) was calculated using R² = 1 − (RMSE/SD)², where SD is the standard deviation of the observed data (namely, 0.202°C for the raw IPCC data, 0.192°C for the data with ENSO removed, and 0.181°C for the low-pass filtered data).