The question as to whether the climatic anomalies associated with the
Medieval Warm Period and the Little Ice Age can be attributed to natur
al climatic variability is explored in this paper. The output from a 5
00-year run with a global climatic model is used for this purpose. The
model exhibits multi-decadal variability in its climatic outputs, whi
ch appears to have many of the characteristics of observed climatic da
ta over the last millennium. Global distributions of surface temperatu
re associated with peak warming and cooling phases of the model run hi
ghlight the spatial variability which occurs, and the lack of synchron
eity in the response from region to region. Considerable year-to-year
variability occurs in temperature anomaly patterns during the warming
and cooling phases, indicating the complexity of the responses. The mo
del results suggest that such climatic phases should not be considered
as lengthy periods of universal warming or cooling. Comparison of obs
erved time series of land surface temperature for the northern hemisph
ere for the last 500 years with model output indicates that most of th
e observed features in this climatic record can be reproduced by proce
sses associated with internal mechanisms of the climatic system as rep
roduced in the model. While the model results do not exclude the possi
ble contribution of external forcing agents as a contributing factor t
o these climatic episodes, the perception is that such agents would en
hance existing naturally-induced climatic features rather than initiat
e them, at least for this time frame. Given the omnipresent nature of
natural climatic variability, it is assumed that such variability rath
er than external forcing agents has primacy in generating and maintain
ing the underlying observed climatic variability. An understanding of
the mechanisms and behaviour of such climatic features is becoming of
increasing importance, in view of their possible role in modulating fu
ture climatic trends given the expected influence of the greenhouse ef
fect.