OCEAN SEVEN 304 CTD LOGGER

LOW POWER MICRO CTD SELF-RECORDING CAPABILITY

The Ocean Seven 304 CTD offers combinations of 16-bit high-resolution data
accuracy with long term sensor stability, making this CTD an ideal choice for self-
recording profiling and moored applications. The CTD uses state-of-the-art
electronics and it is equipped with up to 64Mbytes logging memory. The present
report demonstrates the OCEAN SEVEN 304 measurement performances
comparing them against a SBE-91 CTD.

Report on performance at sea June-2004

Foreword
An IDRONAUT OCEAN SEVEN 304 was tested
in the Western North Atlantic in April 2004. It
was mounted on a CTD and Rosette frame with
the OCEAN SEVEN 304 sensors at the height of
the SBE 911 C/T sensors. The OCEAN SEVEN
304 was set up to log P, T, and C versus time
internally at approximately 8 samples per
second with no averaging applied to the data.
A single profile was collected to a depth of
approximately 3500 meters with 22 water
samples collected on the up cast. This report
compares the OCEAN SEVEN 304 with a
Seabird 911 CTD and water sample salinities.
Pictured in figure 1 is the mounting
arrangement of the OCEAN SEVEN 304 with
the SBE 911 C/T sensors at the same height as
the OCEAN SEVEN 304 temperature and
conductivity sensors. Data from both the down
and up profiles of the Seabird 911 have
appropriate sensor lags applied and are
processed to a uniform pressure series at 2
decibar intervals. The Seabird 911 CTD data files
have a file name of OC401010 (station 10) while
the OCEAN SEVEN 304 file name is
I304_st10.txt. The OCEAN SEVEN 304 data is
processed as indicated on the plots.
Figure 1 - OCEAN SEVEN 304 in CTD frame with SBE 911 C/T sensors

Data comparison
The up cast OCEAN SEVEN 304 bottle locations were found using a lowering rate (dp/dt) criteria applied to
the OCEAN SEVEN 304 pressure sensor data. The pressure data was filtered with a cut-off of approximately
20 second before differencing to create lowering rate. The rosette bottle closing occurs at the end of a bottle
stop. The OCEAN SEVEN 304 bottle stop observations shown in Tables II & I are an average of the 45

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 1/10

observations immediately before the completion of a bottle stop ends. The bottle stops locations are
indicated by black “*” in the upper panel of figure 2.

Figure 2 lowering rate (dp/dt) for upcast with (*) indicating assumed bottle stop positions in
OCEAN SEVEN 304

Table I
OCEAN SEVEN 304 data are an average of 45 observations of the bottle stop pressure, temperature,
conductivity and salinity as located on figure 2 merged together with corresponding Seabird 911 bottle stop
average pressure, primary temperature (T1) and salinity data plus water sample salinities collected from 10
liter rosette sampler water sample bottles. The difference of the secondary (T2) and primary temperature and
the water sample salinity complete Table I. The OCEAN SEVEN 304 salinity in Table I is calculated from
uncorrected OCEAN SEVEN 304 pressure (P), temperature (T), and conductivity (C) as shown in Table I
using the 1978 Practical Salinity Scale (PSS-78). CTD salinities (S) from both instruments are calculated using PSS-
1978 :

S = sw_salt(C/42.914,t90tot68(T),P)

Where function t90tot68 converts temperatures on ITS-90 back to IPTS-68 (see Matlab function in Appendix).

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 2/10

 TABLE I Uncorrected OCEAN SEVEN 304 measurements
% OCEAN SEVEN 304 uncorrected SBE 911 Shipboard CTD
% Scan # Pres. Temp. Cond. Salt Theta Pres. Tmp. Saltsbe T1-T2 SaltWS*
% Numb. DBAR ITS-90 mS/cm PSS Deg C Dbar Deg.C psu psu*
% ******* ******* ******* ******* ****** ******* ******* ***** *
697 3518.1 2.1877 32.3221 34.9166 1.9010 3508.6 2.1872 34.8930 -0.0005 34.8908
2159 3362.0 2.2254 32.3013 34.9199 1.9543 3351.9 2.2247 34.8960 0.0011 34.8943
4471 3109.0 2.3010 32.2799 34.9263 2.0543 3098.3 2.3023 34.9008 -0.0002 34.8997
6253 2907.1 2.5110 32.3965 34.9355 2.2802 2897.8 2.5108 34.9115 -0.0003 34.9091
8182 2708.9 2.6637 32.4622 34.9433 2.4495 2700.0 2.6633 34.9186 -0.0007 34.9166
10079 2508.8 2.8874 32.5905 34.9528 2.6886 2500.3 2.8857 34.9277 0.0001 34.9268
11985 2308.3 3.1218 32.7262 34.9608 2.9383 2301.1 3.1207 34.9355 -0.0001 34.9347
13848 2109.0 3.2869 32.7962 34.9640 3.1197 2102.4 3.2853 34.9380 -0.0003 34.9366
15841 1907.7 3.4021 32.8156 34.9617 3.2521 1900.3 3.4032 34.9348 -0.0003 34.9317
16950 1807.5 3.4512 32.8168 34.9596 3.3098 1800.9 3.4514 34.9320 -0.0002 34.9299
18001 1706.3 3.5185 32.8381 34.9632 3.3854 1700.4 3.5179 34.9336 -0.0002 34.9309
19930 1505.3 3.7449 32.9728 34.9822 3.6270 1499.7 3.7468 34.9522 0.0000 34.9502
21298 1377.1 3.8677 33.0349 34.9894 3.7599 1371.0 3.8702 34.9598 0.0000 34.9577
23573 1145.9 4.0503 33.1053 34.9953 3.9610 1141.2 4.0504 34.9633 -0.0002 34.9614
25611 940.0 4.4786 33.4316 35.0305 4.4036 937.9 4.4780 34.9985 -0.0002 34.9967
27772 707.8 4.8903 33.7036 35.0326 4.8328 705.1 4.8877 34.9998 -0.0002 34.9974
29727 503.7 6.3098 34.9436 35.0739 6.2641 501.4 6.3160 35.0411 -0.0009 35.0411
31561 301.6 9.8096 38.2894 35.2505 9.7747 300.4 9.8223 35.2175 0.0019 35.2182
32351 224.0 11.5347 40.1429 35.4838 11.5061 223.4 11.5443 35.4523 0.0016 35.4466
33115 150.3 11.1913 39.5928 35.3004 11.1725 149.6 11.1871 35.2655 -0.0041 35.2698
33880 73.9 16.8604 46.2679 36.2960 16.8482 73.4 16.8355 36.2520 0.0106 36.2596
35080 2.3 21.0714 50.4429 36.1511 21.0710 2.8 21.0717 36.1180 0.0031 36.1207

Figure 3 Difference (SBE - OCEAN SEVEN 304 uncorrected) pressure, temperature, and
salinity plotted versus pressure.
The field temperature data of the OCEAN SEVEN 304 compares very well with both SBE 911 CTD
temperatures. Only those temperature differences in higher temperature gradient regions (i.e. Shallower
than 500 decibars) do not match within a millidegree. The Idronaut pressure sensor calibration reads
increasingly deeper than the SBE 911 pressure. A least squares linear fit between the OCEAN SEVEN 304
and SBE 911 pressure values gives the following slope and bias coefficients; pcof =[0.99693925453714 -
0.45584242461625] which are applied to the OCEAN SEVEN 304 pressure data in Table II. The OCEAN
SEVEN 304 conductivity is found to read higher than the SBE 911 and is corrected to the SBE conductivity
using a conductivity slope adjust factor cslop=0.99915. There is also a vertical dependence seen in bottom
panel (Ds) of figure 3 that requires the following quartz glass conductivity cell geometry corrections of;

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 3/10

 beta = 3.0e-8;
alpha = -5.5e-7
to remove: C = C*(1+alpha.*(t-T0)+beta.*(p-P0))
A room temperature laboratory OCEAN SEVEN 304 conductivity check prior to the cruise indicated that the
OCEAN SEVEN 304 conductivity read higher than the bath salinity and required a conductivity
slope=0.99947 adjustment to match as compared to the field adjustment of 0.99915.

TABLE II with Corrected OCEAN SEVEN 304 P & C

% OCEAN SEVEN 304 Corrected Shipboard CTD
% Scan # Pres. Temp. Cond. Salt Theta Pres. Tmp. Saltsbe T1-T2 Salt WS*
% Numb. DBAR Deg. C mS/cm PSS Deg C Dbar Deg.C psu psu *
% ITS-90 ITS-90
% ******* ******* ******* ******* ****** ******* ******* *******
697 3506.8 2.1877 32.3259 34.8880 1.9012 3508.6 2.1872 34.8930 -0.0005 34.8908
2159 3351.3 2.2254 32.3049 34.8911 1.9545 3351.9 2.2247 34.8960 0.0011 34.8943
4471 3099.0 2.3010 32.2832 34.8972 2.0545 3098.3 2.3023 34.9008 -0.0002 34.8997
6253 2897.7 2.5110 32.3997 34.9061 2.2803 2897.8 2.5108 34.9115 -0.0003 34.9091
8182 2700.2 2.6637 32.4652 34.9137 2.4496 2700.0 2.6633 34.9186 -0.0007 34.9166
10079 2500.7 2.8874 32.5933 34.9230 2.6887 2500.3 2.8857 34.9277 0.0001 34.9268
11985 2300.7 3.1218 32.7288 34.9307 2.9384 2301.1 3.1207 34.9355 -0.0001 34.9347
13848 2102.0 3.2869 32.7986 34.9337 3.1198 2102.4 3.2853 34.9380 -0.0003 34.9366
15841 1901.4 3.4021 32.8178 34.9312 3.2522 1900.3 3.4032 34.9348 -0.0003 34.9317
16950 1801.5 3.4512 32.8189 34.9290 3.3099 1800.9 3.4514 34.9320 -0.0002 34.9299
18001 1700.6 3.5185 32.8401 34.9324 3.3854 1700.4 3.5179 34.9336 -0.0002 34.9309
19930 1500.2 3.7449 32.9746 34.9512 3.6271 1499.7 3.7468 34.9522 0.0000 34.9502
21298 1372.4 3.8677 33.0366 34.9583 3.7600 1371.0 3.8702 34.9598 0.0000 34.9577
23573 1142.0 4.0503 33.1068 34.9639 3.9610 1141.2 4.0504 34.9633 -0.0002 34.9614
25611 936.7 4.4786 33.4328 34.9989 4.4036 937.9 4.4780 34.9985 -0.0002 34.9967
27772 705.2 4.8903 33.7046 35.0007 4.8328 705.1 4.8877 34.9998 -0.0002 34.9974
29727 501.7 6.3098 34.9444 35.0416 6.2642 501.4 6.3160 35.0411 -0.0009 35.0411
31561 300.2 9.8096 38.2900 35.2177 9.7748 300.4 9.8223 35.2175 0.0019 35.2182
32351 222.9 11.5347 40.1434 35.4506 11.5061 223.4 11.5443 35.4523 0.0016 35.4466
33115 149.4 11.1913 39.5932 35.2673 11.1725 149.6 11.1871 35.2655 -0.0041 35.2698
33880 73.2 16.8604 46.2682 36.2616 16.8482 73.4 16.8355 36.2520 0.0106 36.2596
35080 1.8 21.0714 50.4429 36.1166 21.0710 2.8 21.0717 36.1180 0.0031 36.1207

Figure 4: Difference (SBE - OCEAN SEVEN 304 corrected) pressure, temperature, salinity
(red o are from bottle salinities) versus pressure and red * temperatures are SBE (T1-T2).

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 4/10

The raw deep-water OCEAN SEVEN 304 measurements at 8 hertz are displayed in figure 5 and show a
peak-to-peak salinity variation of 0.02 psu. Figure 6 shows histograms of the raw OCEAN SEVEN 304
measurements. Note the standard deviation of conductivity (0.0026 mS/cm) and salinity (0.0033 psu) and
contrast these values with histograms of the lag corrected value of conductivity (0.0011 mS/cm) and salinity
(0.0017 psu) shown in figure 7. The raw OCEAN SEVEN 304 conductivity contributes most of the noise to
the salinity measurement (temp. Std. = 0.0009 C). Applying a recursive filter to conductivity with a 2 scan lag
(see Appendix for Matlab function recur_fn.m) matches the conductivity to the temperature and reduces
both the conductivity and salinity noise as indicated in figure 7.

To provide a fair comparison of the OCEAN SEVEN 304 time series data at 8 hertz to the SBE 911 2-decibar
pressure averaged profile data, a roll filtered and 2-decibar pressure averaged OCEAN SEVEN 304 down
profile is presented in the deep water plots in figures 5 (black) and figure 9 (red). An estimate of the salinity
noise 2-decibar salinity noise is made for deepest 100 decibars by calculating the standard deviation (Std.)
after removing a salinity trend.

Idronaut 304 : Std. = 0.00060 psu

SBE 911: Std. = 0.00016 psu

It should be noted that the recorded OCEAN SEVEN 304 data resolution for cond. & temp. are 0.001 C &
mS/cm while the SBE 911 has a data resolution an order of magnitude greater (i.e. 0.0001).

Figure 5; Red & cyan are unlagged OCEAN SEVEN 304 salinity while black (OCEAN SEVEN 304) & green
(SBE911) 2-dbar pressure average with sensor lags applied.

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 5/10

 TABLE III - Salinity comparison
% Scan # OS304 SBE SaltWS WS-SBE WS-OS304*
% Numb. PSS psu psu psu psu *
% ************************************************
697 34.8880 34.8930 34.8908 -0.0022 0.0028
2159 34.8911 34.8960 34.8943 -0.0017 0.0032
4471 34.8972 34.9008 34.8997 -0.0011 0.0025
6253 34.9061 34.9115 34.9091 -0.0020 0.0029
8182 34.9137 34.9186 34.9166 -0.0020 0.0029
10079 34.9230 34.9277 34.9268 -0.0009 0.0038
11985 34.9307 34.9355 34.9347 -0.0008 0.0040
13848 34.9337 34.9380 34.9366 -0.0014 0.0029
15841 34.9312 34.9348 34.9317 -0.0031 0.0005
16950 34.9290 34.9320 34.9299 -0.0021 0.0009
18001 34.9324 34.9336 34.9309 -0.0027 -0.0015
19930 34.9512 34.9522 34.9502 -0.0020 -0.0010
21298 34.9583 34.9598 34.9577 -0.0021 -0.0006
23573 34.9639 34.9633 34.9614 -0.0019 -0.0025
25611 34.9989 34.9985 34.9967 -0.0018 -0.0022
27772 35.0007 34.9998 34.9974 -0.0024 -0.0033
29727 35.0416 35.0411 35.0411 0.0000 -0.0005
31561 35.2177 35.2175 35.2182 0.0007 0.0005
32351 35.4506 35.4523 35.4466 -0.0057 0.0015
33115 35.2673 35.2655 35.2698 0.0043 0.0025
33880 36.2616 36.2520 36.2596 0.0076 -0.0020
35080 36.1166 36.1180 36.1207 0.0027 0.0041

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 6/10

Figure 6: Raw data unlagged: note cond. & salinity noise levels are higher than figure 7 applying recursive
filter to conductivity

Figure 7: Deep observations with cond. recursively lagged 2 scans: note noise of cond. & salinity reduced to half
of unlagged (see figure 6).
The full down and up profiles of the OCEAN SEVEN 304 and SBE 911 salinity and temperature are plotted
versus pressure in figure 8. A temperature lag of 0.25 seconds (Tlag=2 scans) was applied to conductivity
using a recursive filter (i.e. c=recur_fn(cr,Tlag see appendix). A detailed down profile of time-series and 2-
decibar pressure averaged temperature and salinity for the deepest 200 decibars is shown in figure 9.

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 7/10

 Figure 8 Corrected OCEAN SEVEN 304 & SBE 911salinity and temperature versus pressure.

Figure 9: Expanded plot of T & S versus P for OCEAN SEVEN 304 and SBE 911 in deep water.

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 8/10

 Figure 10: Pot. Temp. versus Salinity Deeper waters; down (left); up (right)

The close match of salinity between the OCEAN SEVEN 304 and SBE 911 is demonstrated for both down
and up profiles, for the entire profile and for a detailed section of the deepest portion of the water column
shown in figures 10 a & b. Both CTD instruments have sensor lag corrections
rrections and pressure averaging
applied to the data. The OCEAN SEVEN 304 also has a conductivity slope of 0.99915 and Quartz glass cell
geometry corrections applied

Summary
A comparison was made of the IDRONAUT OCEAN SEVEN 304 and a Seabird 911 CTD over depths from
the surface to 3500 decibars in the Western North Atlantic. The temperature compares extremely well in all
regions of the profile except where strong temperature gradients preclude valid data checks. The average
temperature difference is nearly zero ( i.e. Average DT= Tsbe1 – Ti304= -0.00015 C). The OCEAN SEVEN
304 pressure sensor appears to have a systematic pressure calibration error making the deep observations
read high by roughly 0.3 percent compared to the SBE 911. The OCEAN SEVEN 304 conductivity measure
high compared to the bottle salinity measurements and is corrected by adjusting the conductivity slope
down to 0.99915. The conductivity sensor has a Quartz cell. Cell geometry corrections for dimensional
changes with pressure (bulk modulus = 5.3E+6 psi) and temperature (5.5E-7 cm/cm/C) obtained from the
CRC Handbook improved the conductivity measurements. Applying a recursive filter to lag conductivity
before computing salinity improves the noise of both by a factor of 2.5. The conversion of the OCEAN
SEVEN 304 data to a lagged, roll-filtered and pressure averaged data set at 2 decibar intervals further
improves the OCEAN SEVEN 304 noise.

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 9/10

Appendix
Matlab processing algorithms:

function [c1]=recur_f_2(c,w)
%function [c1]=recur_f_2(c,w)
% recur_f_2 (x,W)
% x is input data
% exp. filter to lag cond. to match temp.
% W is lag in scans;
% W = 1/(1+1/W)
% y=y(j-1)*W+x(j)*(1-W)
% jmt Feb 1999
%
w1=1/(1+1/w);
a(1)=1;
a(2)=-w1;
b(1)=(1-w1);
cb=c(1);
c1=filter(b,a,(c-cb));
c1=c1+cb;
return

function t68 = t90tot68(t90)

%function t68 = t90tot68(t90)
%
%-----------------------------------------------------------------------------
% Converts temperature from ITPS-90 to ITS-68
%
% Input:
% t90 - Temperature [ PTS-90]
%
% Output:
% t68 - Temperature [IPTS-68]
%
% References:
%
% Checkvalue: T90toT68(100) = 100.024
%
%---------------------------------------------------------------------

% S. Chiswell 1991

t68 = t90/0.99976;

IDRONAUT Srl – Brugherio OS304vsSBE911 Pag. 10/10