In the past two decades methods have been devised for : 
(a) Prediction of surface waves based on observations and pre- 
dictions of surface-wind distribution.. 
(6) Warnings of tsunamis produced by earthquakes which are 
readily detected at great distances. 
These methods have proved vital for safety and economy in coastal 
areas, in commercial shipping and for many military operations. 
Further improvement in wave prediction is tied closely to atmospheric 
prediction, for which atmospheric observations over the oceans are 
required. In a similar way prediction of the depth of the surface 
mixed layer, still in its early stages, is closely tied to the meteorologi- 
cal problem. Understanding the processes occurring in the surface 
mixed layer is important for acoustic-transmission applications within 
the sea and for marine biological problems. 
We have reason to think that these phenomena, for which rather 
simple prediction methods are available, fail to encompass other char- 
acteristic, important features of the ocean. From the fragmentary 
evidence we have at present, it appears that a wide range of time- 
dependent phenomena do indeed occur in the ocean, as our experience 
with stratified fluids in the laboratory or in the atmosphere would lead 
us to expect. Ocean weather may be as varied and complex as the 
weather in the atmosphere. For example, we see indications of inter- 
nal gravity waves, inertial motions associated with the earth’s rotation, 
turbulence, meanders in the Gulf Stream and other currents and 
fluctuations in surface temperature over large areas; but we have not 
yet adequately described any of these phenomena. Whether current 
systems occur which are comparable in size to atmospheric planetary 
waves remains to be discovered. The extent to which prediction of 
these phenomena is inherently feasible and for what scales of time and 
space remains unknown; these problems appear destined to become 
some of the most exciting objectives of ocean research in the next 
decade. The answers are not obvious, for although the governing 
differential equations are well known, we do not know the strength of 
coupling between observable and unobservable scales. 
We do know, however, that lack of ocean surface-layer observations 
restricts effective atmospheric prediction to a few days. 
Until the prediction problem is better understood, the potentialities 
of deliberate ocean modification cannot be determined. Without such 
understanding, large-scale experiments addressed to diverting ocean 
currents, to melting the Arctic ice or to overturning large regions of 
ocean water would be extravagant and highly irresponsible. How- 
ever, inadvertent modification of coastal areas, already of local con- 
cern, is likely to become more serious. In order to plan wisely for 
use and development of coastal areas we must learn to predict such 
43 
