                     CTD CALIBRATION: CRUISE CIROLANA 6/1981               
                                                                           
Thirty-one CTD profiles and associated calibration data were available from
this  cruise.  Suitable calibration coefficients have  been  calculated  as
described  below.  Only one profile descended to below 3000m but  this  did
indicate  good  agreement  between  the  corrected  data  and  the  theta-S
relationship suggested by Peter Saunders.                                  
                                                                           
                                                                           
Pressure                                                                   
No  pre-cruise  or  post-cruise calibration  was  available.  The  earliest
pressure  calibration  of this CTD (S/N 45056) using a dead  weight  tester
dates from April/May 1983 and this was applied to the pressure data:       
                                                                           
      P(COR) = P(CTD) + deltaP                                             
where deltaP = a * T(COR) + b * P(CTD) + c                                 
and        a = 0.857466                                                    
           b = 0.331988 E-2                                                
           c = -11.96                                                      
                                                                           
                                                                           
Temperature                                                                
In  the  absence  of post- or pre-cruise calibrations  of  the  temperature
sensor,  a comparison was made between recorded CTD and DSRT  temperatures.
Differences ranged between -10 and +60 mK,  but it is believed likely  that
errors  associated  with thermometers,  rather than variability in the  CTD
sensor,  caused this spread (at this time no restricted range  thermometers
were available; they would have had ranges -2 to 12degC or -2 to 25degC). A
mean difference was calculated and applied as a correction:                
                                                                           
        T(COR) = T(CTD) + deltaT                                           
where   deltaT = 0.164 E-1                                                 
                                                                           
                                                                           
Conductivity Ratio                                                         
A  comparison between CTD conductivity ratio and the conductivity ratio  of
water  samples  at the corrected CTD pressure and temperature was  used  to
determine suitable calibration coefficients:                               
                                                                           
        CR(WS)                                                             
        -----------     = a * T(COR) + b * P(COR) + c                      
        CR(CTD)                                                            
                                                                           
Then corrected CTD conductivity ratio is given by:                         
                                                                           
        CR(COR) = CR(CTD) * (a * T(COR) + b * P(COR) + c).                 
                                                                           
From  the  distribution  of the ratio CR(CTD): CR(WS) it  seemed  that  the
calibration  was  best served by dividing the data into 3 groups (see  Fig.
1):                                                                        
                                                                           
stations 49-62:                                                            
        a = -0.13598 E-3                                                   
        b = -0.72673 E-6                                                   
        c = 1.001855                                                       
                                                                           
stations 63-84:                                                            
        a = -0.23385 E-4                                                   
        b = -0.102509 E-6                                                  
        c = 1.000283                                                       
                                                                           
stations 85-106:                                                           
        a = 0.271242 E-4                                                   
        b = 0.949879 E-8                                                   
        c = 0.999224                                                       
                                                                           
A  comparison  between  water   sample   and  CTD  salinity  indicated  rms
differences of 0.005 (N=25), 0.003 (N=32) and 0.003 (N=19) respectively.   
                                                                           
Fig.  2 compares the CTD and water sample salinity differences  before  and
after  this calibration has been applied.  After correction of the  sensors
89% of the differences are within +/- 0.006.                               
                                                                           
One   profile  descended  beyond  3000m.  Fig.  3  compares   the   theta-S
distribution with the linear relationship suggested by Peter Saunders.  (No
data  were  obtained  between  3582  and 4207 db).  It  suggests  that  the
coefficients are providing an effective calibration of the CTD sensors.    
                                                                           
Ken Medler                                                                 
28 February 1990                                                           
