107G THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1957 



4.2 Characteristics of Available Information 



4.2.1 Sources of Data 



Observations of temperature in the deep sea were first made in the 

 mid-nineteenth century. The early observations were made with crude 

 instruments and are now of purely historical value. The Challenger 

 expedition of 1872-1876 made several hundred observations but with 

 maximum-minimum thermometers unprotected from pressure. In the 

 late nineteenth century the Richter reversing thermometer was invented 

 and by the turn of the century they were used by nearly all scientific 

 expeditions. Cable ships have taken many observations but almost al- 

 ways with maximum-minimum thermometers. 



Major expeditions have published volumes which included tabulated 

 lists of temperature, salinity, oxygen, etc., for each station occupied, 

 while the shorter expeditions and those institutions which continually 

 collect oceanographic data in the North Atlantic publish their observa- 

 tions in the "Bulletin Hydrographique," a journal published by the 

 International Commission for the Exploration of the Sea, Copenhagen. 

 In addition, unpublished data are available from the files of oceano- 

 graphic institutions. 



Expedition reports give estimates of the reliability and accuracy of' 

 their data and usually describe the calibration tests used to determine ' 

 the accuracy. The "Bulletin Hydrographiciue" merely publishes the data: 

 without comment. The scarcity of data and the tendency to systematic 

 errors in single sets of data coupled with the temperature changes now. 

 being demonstrated for deep ocean water masses tend to frustrate efforts 

 to evaluate the accuracy of data. 



4.2.2 Methods of Evaluation 



Evaluation of temperature data involves comparison of nearby obser-^ 

 vations, verification of the original data sheet, and checking for errors 

 in computation. The calibration of the thermometers is ordinarily done 

 with great care, and observations are generally accurate to ±.05°C. 

 When the thermometer fails to function properly, the temperature is 

 usually so far off that the observation is not reported. The main error 

 comes in the determination of depth of observations. The length of wire 

 paid out to reach a stated depth varies with the magnitude of winds and 

 currents in the area. On early expeditions this led to large errors in 

 observation. The use of two thermometers, one in a pressure case and 

 one unprotected from pressure, allows the calculation of depth of obser- 

 vations with relatively great accuracy. 



