Table of Sigma- 1 with Intervals of 0.1 

 for Temperature and Salinity 



By 



BETTY ANN L. KEALA, Statistician 



Bureau of Commercial Fisheries Biological Laboratory 



Honolulu, Hawaii 



INTRODUCTION 



The following table presents valuesof sigma-t 

 for a range of temperature of -2° to 33° C. at 

 intervals of 0.1° and for a range of salinity of 

 2 parts per mille to 39 parts per mille at inter- 

 vals of 0.1 part per mille. These are essentially 

 the same ranges of salinity and temperature as 

 those covered by Knudsen's tables (Knudsen, 

 Martin, editor. 1901. Hydrographical tables. 

 Copenhagen, G.E.C. Gad, 63 p.). The values 

 were computed by means of a program written 

 for the I.B.M. 7040 computer using the equa- 

 tions for sigma-t found in Knudsen's publica- 

 tion. This same program also prepared the 

 listing of the values, of which the following table 

 is an exact reproduction. 



Knudsen's equations are known to be in error 

 to a significant extent at extremes of the temper- 

 ature and salinity range (Carritt and Carpenter, 

 "The composition of sea water and the salinity- 

 chlorinity-density problems," 1959. NAS-NRC 

 publication 600, p. 67-86). However, until a 

 satisfactory revision is available, Knudsen's 

 equations provide for internal consistency in the 



oceanographic literature, with a degree of accu- 

 racy that is entirely adequate for most pur- 

 poses. 



Although tables similar to the present table 

 have been published before, all require interpo- 

 lation between relatively wider intervals of tem- 

 perature or salinity, or both, than does the pres- 

 ent table, or they cover limited ranges of tem- 

 perature and salinity. Since the dependence of 

 density, and therefore of sigma-t, on tempera- 

 ture and salinity is nonlinear, interpolation by 

 the usual linear method introduces errors which 

 maybe significant where precisionis important. 

 The present table minimizes the effects of inter- 

 polation errors by limiting the ranges over which 

 linear interpolations must be made and by using 

 five significant figures. The tablets most useful 

 for precise work, particularly when salinity de- 

 terminations are made by conductivity methods, 

 whichcan be almost an order of magnitude more 

 precise than titrations. 



Richard A. Barkley provided advice and en- 

 couragement. 



