Projected changes in the Earth's climate can be driven from a combined
set of forcing factors consisting of regionally heterogeneous anthrop
ogenic and natural aerosols and land use changes, as well as global-sc
ale influences from solar variability and transient increases in human
-produced greenhouse gases. Thus, validation of climate model projecti
ons that are driven only by increases in greenhouse gases can be incon
sistent when one attempts the validation by looking for a regional or
time-evolving ''fingerprint'' of such projected changes in real climat
ic data. Until climate models are driven by time-evolving, combined, m
ultiple, and heterogeneous forcing factors, the best global climatic c
hange ''fingerprint'' will probably remain a many-decades average of h
emispheric-to global-scale trends in surface air temperatures. Century
-long global warming (or cooling) trends of 0.5-degrees-C appear to ha
ve occurred infrequently over the past several thousand years-perhaps
only once or twice a millennium, as proxy records suggest. This implie
s an 80 to 90 percent heuristic likelihood that the 20th-century 0.5 /- 0.2-degrees-C warming trend is not a wholly natural climatic fluctu
ation.