(1) Provide data for support of research; 



(2) Provide data used for survey purposes; and 



(3) Provide current data for operational uses. 



The relative weights given each class of use varies considerably. 

 In oceanography the first two have been predominant and I believe 

 that this has been fortunate. In the field of meteoroiogy the reverse 

 has been true. Here, almost the entire emphasis has been placed 

 on the current use of data for weather forecasting and particularly 

 for aircraft operations. Instrumentation and the more sophisticated 

 observation networks have been created primarily to satisfy these 

 needs. The individuals engaged in atmospheric research and in 

 climatology have had to be satisfied to use data which were largely 

 designed to serve exclusively an entirely different purpose. It is 

 my opinion that the science of meteorology has suffered immeasvir- 

 ably because of this course of evolution. I mentioned previously 

 that it has been fortunate for oceanography that instrumentation and 

 observation practices have been designed primarily to meet the 

 needs of oceanographic research and ocean survey work. We in the 

 Data Center, however, have some reason to be concerned that, 

 under the impetus of newly-developed environmental prediction 

 systems, the emphasis might be changed to the extent that the spec- 

 ifications of the operational interests may be met at the expense 

 of those of the oceanographer. 



It is possible to design instruments to satisfy both classes of 

 requirements. This is done by being certain that the final record 

 is in standard physical units and not in some operational unit that 

 has more than one physical interpretation. 



An example of a device that does not fit this principle is one 

 of the meteorological instruments called a transmissometer which 

 is actually a visibility (or extinction) meter for use at airports. 

 The aircraft pilot is interested in visibility but is not concerned 

 particularly with what meteorological phenomenon it is thg.t reduces 

 the visibility. Since the transmissometer records visibility only, 

 when its record is the only one telemetered from a remote loca- 

 tion, it is n-jeteorologically meaningless to the scientist. There 

 are many opportunities for such types of operational instrumenta- 

 tion in oceanography and these we should avoid except where we 

 can maintain oceanographic integrity through the use of economi- 

 cal and efficient supplemental instrumentation. 



In closing I would like to list very briefly a few other features 

 in instrumentation that would be highly desirable from the stand- 



64 



