then suggests that the results be used to examine the contingencies so 
estimated and the associated costs and benefits of providing reasonable 
reserves against these contingencies. 
The Crucial Role of the Oceans: Long-Term Global Climate* 
Aside from the normal large-amplitude year-to-year climatic fluctuations, 
there are longer term variations of smaller amplitude but of far-reaching 
consequence. Over tens and hundreds of thousands of years, these have 
manifested themselves as alternating ice ages and warm interglacial periods. 
But even within the limited range of recent history, significant changes have 
taken place. 
Considering only very recent geological history, the earth has, for the 
past ten thousand years, been experiencing a warm interglacial period. 
Within this period, from roughly 950 to 400 B.C. there was a cool spell, 
a warm spell occurred from 800 to 1200 A.D., a cool spell (the “little ice 
age’) from 1550 to 1850, and from 1900 on another relatively warm spell— 
a climate “optimum,” which may be ending. Average global temperatures 
varied by no more than several degrees in the course of these changes but 
this was sufficient, if they persisted over many years, to create large fluctua- 
tions in living conditions around the globe. For climate changes on this 
scale, the possibility exists to adapt agricultural and other cultural prac- 
tices provided adequate warning can be given. But these changes go far 
beyond the realm of chance and their causes must be understood before 
prediction is possible. 
We are not, at present, capable of predicting such changes. Yet even 
very small variations, such as changes of only several tenths of a degree 
in the earth’s surface temperature, may have significant effect on society. 
This lack has not been unrecognized. The Global Atmospheric Research 
Program (GARP) was begun in the midsixties with the twofold aim of: 
e Achieving understanding of the transient behavior of the atmosphere 
as manifested in the large-scale fluctuations which control changes 
of the weather. This would lead to increasing the accuracy of fore- 
casting over periods from one day to several weeks. 
¢ Achieving understanding of the factors that determine the statistical 
properties of the general circulation of the atmosphere. This would 
lead to better understanding of the physical basis of climate. 
Considerable progress has been made toward the first of these goals 
but not toward the second. This may have come about because, although 
the GARP mandate was addressed to, and has been accepted by, the 
atmospheric science community, the second objective of GARP depends 
* We are indebted to J. O. Fletcher of the National Science Foundation for much 
of the material in this section. 
