analysis, and distribution of the data. 



The assumption is that the participation of the United States 

 in an international program requires the exploration of roughly 30 

 percent of the world's oceans. The vastness of the international 

 program is apparent from the following figures: The oceans cover 

 360 million square kilometers; of these, 36 million comprise the 

 continental and insular shelves; nearly 30 million are covered by 

 Arctic ice; approximately 300 million square kilometers require 

 exploration and investigation. If track lines are run at every 15 

 kilometers, survey ships have a task of sounding along some 20 

 million linear kilometers of ocean exclusive of development of un- 

 usual submarine features and cross check lines. 



For an eight-month operating season, a ship cruising at 12 

 knots might theoretically accomplish this task in 200 years. The 

 National Academy of Sciences' Committee on Oceanography esti- 

 mates for the underway portion (hydrographic, magnetic, gravity, 

 etc.) and the anchored or hove to station ships (physical, chemi- 

 cal, synoptic, biological, and bottom sampling surveys) will 

 require 261 ship-years -- for the United States participation this 

 is about seventy-eight ship-years, eighteen ship-years for the sta- 

 tion ships, and sixty ship-years for the ocean bottom survey ships. 



The great cost in ships and manpower demands that most 

 expeditious and efficient nnethods be used to collect and process 

 the data. Thus, many types of data must be collected simultan- 

 eously, and all instruments must be generating data of sufficient 

 and known accuracy. With a ship cost of several thousand dollars 

 per day, failure of a data system simply means failure of the pro- 

 gram, as it is doubtful that, with the enormity of the task, reruns 

 can be justified. Now this means that each instrument incorporate 

 at each critical point in the operational sequence, alarms, or 

 indicators of some sort, that alert the shipboard personnel when a 

 casualty or a circuit failure occurs which degrades the system. 

 Worse than no data is voluminous data of questionable validity. 

 This quality control feature has been almost totally ignored in 

 instrument design in the past. Frankly, we must accept nothing 

 less than good quality control in the future. 



Before completely dropping the topic of philosophy of instru- 

 ment design (an area upon which I'm trespassing as it rightfully 

 belongs to Mr. Snodgrass, scheduled later this morning) I must 

 point to an instrument concept that is also properly a "Survey 



18 



