                   CTD CALIBRATION: CRUISE CIROLANA 8/1992                 
                                                                           
                                                                           
CTD: 8705                                                                  
                                                                           
(1)  Sixty-five  CTD profiles were obtained during  the  cruise.  Reversing
thermometers were fitted to the Niskin used to collect samples. Samples for
salinity  analysis  were collected from both the near surface and close  to
the seabed.                                                                
                                                                           
Not  all  the  calibration data collected were used to  calibrate  the  CTD
sensors:                                                                   
   (i)   Thermometer differences too large, > 0.04.                        
   (ii)  It was clear from the profiles obtained at time of station that   
some  T and/or S values measured by the CTD were variable,  believed  real,
not an instrument malfunction, and unsuitable for calibration use.  21 data
sampling points were not used because of this.                             
   (iii) The salinity analysis of water samples from station 167 (top)     
onwards  was suspect,  possibly due to room temperature changes.  These  28
samples were not used for the calibration.                                 
                                                                           
                                                                           
(2)  The pressure sensor was corrected using the laboratory calibration  of
August 1992 at T = 16.2 deg C.                                             
                                                                           
 P(cor) = P(ctd) - 1.4                                                     
                                                                           
                                                                           
(3)  The temperature sensor was corrected using laboratory  calibration  of
August 1992.                                                               
                                                                           
 T(cor) = T(ctd) + dT                                                      
                                                                           
     dT = a*T(ctd)*T(ctd) + b*T(ctd) + c                                   
                                                                           
where a = -1.18042e-05                                                     
      b =  3.74e-04                                                        
      c =  1.0502e-02                                                      
                                                                           
This  amounts to 13 mK for temperatures encountered during the  cruise,  of
about 18 deg C.                                                            
                                                                           
A comparison between thermometer and uncorrected CTD temperatures is  shown
in Fig. 1.  If it is assumed that the thermometer temperatures are accurate
to 0.02 deg C and the CTD to 0.01 deg C then all differences to within 0.03
deg C of the Laboratory calibration are acceptable i.e.  those within 43 mK
and  - 17 mK.  Nine out of ninety-five values were found to be outside this
range.                                                                     
                                                                           
The  mean of the eighty-six differences was found to be 15 mK which  is  in
excellent agreement with the laboratory calibration.                       
                                                                           
                                                                           
(4) Fig.  2 shows how the uncorrected CTD salinity estimates compared  with
salinity measurements from the salinometer.  (Note that CTD measurements of
pressure,  temperature  and conductivity have not been corrected;  the  CTD
salinity is derived from uncorrected observations).                        
                                                                           
This  plot suggests a different CTD response before and after station  101,
also seen in Fig.  3, which shows how the conductivity ratio of the CTD and
water samples varies during the cruise.                                    
                                                                           
                                                                           
(5) Fig. 2 and Fig.  3 suggest a change in response of the CTD near station
102. Consequently, the salinity samples were divided into two parts for the
calibration.                                                               
                                                                           
CR(cor) = CR(ctd)*[ a*T(cor) + b*P(cor) + c ]                              
                                                                           
where:                                                                     
Stations 10 to 101 inclusive                                               
      a = -0.267503539e-04                                                 
      b = -0.200824708e-05                                                 
      c =  0.100047932e+01                                                 
                                                                           
The rms difference between salinometer and corrected CTD                   
salinity was 0.005 for 50 observations.                                    
                                                                           
Stations 102 to 167                                                        
      a =  0.761912360e-04                                                 
      b = -0.215006427e-05                                                 
      c =  0.998961505e+00                                                 
                                                                           
The rms difference between salinometer and corrected CTD salinity was 0.005
for 44 observations.                                                       
                                                                           
Fig.  4 shows how effective the sensor coefficients are in  correcting  the
CTD  salinity  estimates  for  the  calibration  samples.  The  offset  and
discontinuity seen in Fig. 2 and Fig. 3 have been removed.                 
                                                                           
Fig.  5  shows the difference between water sample  (salinometer)  and  CTD
salinity  before  and after the calibration of the conductivity sensor  has
been applied.  Note that for the upper histogram the CTD salinity has  been
derived  from  corrected  CTD  pressure  and  temperature  but  uncorrected
conductivity  ratio.  This  figure therefore indicates  how  effective  the
conductivity  calibration has been.  If it is assumed that the CTD salinity
is  accurate  to  0.01  and the salinometer  is  accurate  to  0.006,  then
differences less than or equal to 0.016 are acceptable. Only one of ninety-
four  data values is outside this range after the sensor calibrations  have
been applied.                                                              
                                                                           
CTD stations 167 to 213 were corrected using the coefficients for  stations
102 to 167.                                                                
                                                                           
                                                                           
(6)  The  %transmission sensor readings were converted to  suspended  loads
using the following calibration:                                           
                                                                           
     SUSP.LOAD = -8.61 * LOGe(%TRANSMISSION) + 33.58                       
