James Clarke Ross Cruise AMT12 CTD Processing

Sampling strategy

A total of 69 successful CTD casts were made during the cruise. Rosette bottles were fired throughout the water column on the upcast of most profiles. Data were measured at 24 Hz by a PC running SEASAVE, Sea-Bird's data acquisition software. The raw data files were supplied to BODC after the cruise.

Originator's processing

Only a subset of files had been partially processed on board during the cruise. The raw data were therefore reprocessed at BODC to produce and homogeneous set of CTD data files for this cruise. BODC used the latest version of the SeaBird Processing software available at the time to process the raw binary data files (DAT files) based on information held in the sensor configuration files (CON files), and bottle firing files (BL).

BODC post-processing and screening

Sea-Bird processing

The CON files were first checked for any changes which may have occurred during the cruise, with the exception of the transmissometer coefficients that changed on a cast by cast basis to account for source decay, none were made. The information was also cross checked against information held in the sensors' calibration reports.

The following SeaBird routines were then carried out using SBE Data Processing software version 5.30a: DATCNV, CELLTM, FILTER, LOOPEDIT, DERIVE, BINAVG, STRIP. After CELLTM was run, tests were carried out to check whether an alignment of the conductivity sensor was necessary. No lag was observed. Details of the routines and settings used were as follows:

DATCNV converts the raw data into engineering units. Both down and upcasts were selected. All channels were selected for transfer.

The manufacturer's calibration for the fluorometer was applied during Sea-Bird processing as follows:

Stainless Steel	Nominal chl-a conc (µg/l) = (0.009 * 10 ^voltage) - 0.016
Titanium	Nominal chl-a conc (µg/l) = (0.00864 * 10 ^voltage) - 0.0201

CELLTM was run on the DATCNV output using SeaBird's recommended settings of alpha= 0.03 and Tau=7.

FILTER was run on pressure using a low pass time constant of 0.15 seconds.

LOOPEDIT was run in order to minimise the marked wake effect linked to ship rolling observed on recent cruises.

DERIVE, BINAVG and STRIP were then run to derive the salinity and oxygen concentration, reduce the data to 2Hz and strip redundant channels from the final sets of ASCII files.

Conversion of transmissometer voltages to beam attenuation

There were air and blocked path readings along the cruise track available for this cruise. The transmissometer profiles were corrected for source decay through the calculation of coefficients M and B following SeaBird Application Note 7 .

M = (T _w / (W ₀ - Y ₀) * (A ₀ - Y ₀) / (A ₁ - Y ₁)

B = -M * Y ₁

where

		Stainless steel	Titanium
T _w =	% transmission for pure water	100%	100%
W ₀ =	voltage output in pure water	4.205 V	4.213 V
A ₀ =	manufacturer's air voltage	4.660 V	4.514 V
Y ₀ =	manufacturer's blocked path voltage	0.027 V	0.021 V
A ₁ =	current air voltage	cruise logsheet - cast specific
Y ₁ =	current blocked path voltage	cruise logsheet - cast specific

Reformatting

The data were converted from Sea-Bird ASCII format into BODC internal format (QXF) using BODC transfer function 357. The following table shows how the variables within the Sea-Bird files were mapped to appropriate BODC parameter codes:

Sea-Bird Parameter Name	Units	Description	BODC Parameter Code	Units	Comments
Pressure, Digiquartz	dbar	CTD pressure	PRESPR01	dbar	-
Temperature [ITS-90]	°C	Temperature of water column by CTD sensor 1	TEMPCU01	°C	-
Temperature, 2 [ITS-90]	°C	Temperature of water column by CTD sensor 2	TEMPCU02	°C	-
Salinity	-	Practical salinity of the water body by CTD sensor 1	PSALCU01	-	-
Salinity, 2	-	Practical salinity of the water body by CTD sensor 2	PSALCU02	-	-
Oxygen	µmol kg ^-1	Dissolved oxygen concentration	DOXYSU01	µmol l ^-1	Converted from µmol kg ^-1 to µmol l ^-1 using sigma-T during transfer
Fluorescence	mg m ^-3	Nominal chl-a concentration	CPHLPM01	mg m ^-3	Manufacturer's calibration applied during processing
Beam Attn	m ^-1	Beam attenuance	ATTNDR01	m ^-1	Generated using manufacturer's calibration corrected for source decay
-	-	Practical salinity of the water body by CTD sensor 1 - sample calibrated	PSALCC01	-	PSALCU01 calibrated against bench salinometer data
-	-	Practical salinity of the water body by CTD sensor 2 - sample calibrated	PSALCC02	-	PSALCU02 calibrated against bench salinometer data
-	-	Dissolved oxygen concentration - sample calibrated	DOXYSC01	µmol l ^-1	DOXYSU01 calibrated against Winkler titration data
-	-	Fluorometer - sample calibrated	CPHLPS01	mg m ^-3	CPHLPM01 calibrated against fluorometric chlorophyll-a data
-	-	Oxygen saturation	OXYSSC01	%	Generated by BODC using the Benson and Krause (1984) algorithm wioth parameters DOXYSC01, PSALCC01 and TEMPCU01
-	-	Potential temperature	POTMCV01	°C	Generated by BODC using UNESCO Report 38 (1981) algorithm with parameters PRESPR01, PSALCC01 and TEMPCU01
-	-	Sigma-theta	SIGTPR01	kg m ^-3	Generated by BODC using the Fofonoff and Millard (1982) algorithm with parameters PSALCC01 and POTMCV01

References

Benson, B.B. and Krause, D., 1984. The concentration and isotopic fractionation of oxygen dissolved in freshwater and seawater in equilibrium with the atmosphere. Limnol. Oceanogr., 29(3), 620-632

Fofonoff, N.P. and Millard, R.C., 1983. Algorithms for computations of fundamental properties of seawater. UNESCO Technical Papers in Marine Science No. 44, 53pp.

UNESCO, 1981. Background papers and supporting data on the International Equation of State of Seawater 1980. UNESCO Technical Papers in Marine Science No. 38, 192pp
Screening

Reformatted CTD data were transferred onto a graphics work station for visualisation using the in-house editor EDSERPLO. Downcasts and upcasts were differentiated and the limits manually flagged. No data values were edited or deleted. Flagging was achieved by modification of the associated BODC quality control flag for suspect or null values.

There were leaking connectors/corroding pins on the stainless steel CTD system due to damage caused on the previous cruise. This caused noise and loss of pumps on primary and secondary temperature and conductivity sensors for AMT12_40, 42, 43, 45 and 47.
Screening notes regarding specific casts:

Salinity: The secondary salinity channel (PSALST02) was suspect for casts AMT12_33, 34 and 36.

Dissolved oxygen: The dissolved oxygen signal for AMT12_02 was suspect between 170 - 270 db. For AMT12_24, the downcast concentration was high relative to the upcast below 2000m.

Beam attenuance: The Chelsea Alphatracka Mk II transmissometers are designed to operate in a temperature range of 1 to 25 °C. Where they encountered higher air and water temperatures, they suffered from hysteresis which affected the up- and downcasts. The severity of the problem depended on the degree of heating on the deck and in the water, and also on the pressure. Jeff Benson noted that the casts between AMT12_27 and AMT12_36 were affected. During screening of the CTD data, it appeared that the transmissometers were affected for more of the casts than originally thought. The worst affected casts were from AMT12_27 to AMT12_36 as identified by UKORS. However, the upcast and downcast signals were very different for casts from AMT12_20 to AMT12_26 and for AMT12_37 to AMT12_44. Minimum beam attenuation values were too low and require further calibration and correction. See the data quality section for more details.
Banking

Once quality control screening was complete, the CTD downcasts were banked. Finally, the data were binned against pressure at 1 dbar increments with flagged data excluded from the bin averaging. The primary salinity, temperature, density and potential temperature channels were retained as the best quality data channels from the two sensors.

Field Calibrations

Pressure

There were no casts where the CTD pressure was logging in air. No adjustments were made to the values resulting from application of manufacturer's coefficients during the initial processing.
Temperature

Temperature readings from the two temperature sensors were almost identical and no other independent measurements of better quality were available. No further correction was therefore applied to the data.

Salinity

There were 71 CTD bottle samples from the stainless steel casts and 73 samples from the titanium casts that were analysed on the bench salinometer. These have been compared with CTD values at the depth of bottle firing during the CTD upcast. The data were extracted from Sea-Bird bottle files created during processing. They were also compared with CTD values at the same depths taken from the CTD downcast. This additional information was used to remove outliers from the calibration data set.

Despite the large calibration set, there was still a high degree of variability in the relationship between the CTD sensor data and the salinometer data.

Stainless steel frame

Samples from casts AMT12_40, 42, 43, 45 and 47 were excluded from the data set due to the quality problems described above.

The mean offset between the primary and sceondary sensors showed a significant change over the cruise. Comparison with the bench salinometer data suggested that primary sensor was drifting over time. The casts have been split into two groups for the calibration of PSALCU01. The offset between bench salinometer data and secondary sensor values was consistent throughout the cruise. One calibration was, therefore, applied to produce PSALCC02 and this should be used in preference to PSALCC01.

Sensor	Stainless Steel Casts	Calibration	N	stddev	BODC cal ref
Primary	AMT12_01 to AMT12_27	PSALCC01 = PSALCU01 + 0.002113	35	0.002	2768
Primary	AMT12_29 to AMT12_67	PSALCC01 = PSALCU01 + 0.008350	25	0.004	2767
Secondary	All	PSALCC02 = PSALCU02 + 0.003734	62	0.003	2766

Titanium frame

The mean offset between the primary and sceondary sensors was -0.0055 with a standard deviation of 0.0024. However, it was noted that there was a trend of increasing offset over time. The CTD casts were split into two groups - from AMT12_09 to AMT12_39 and from AMT12_44 to AMT12_69. The offsets between the salinometer and CTD data for both sensors were calculated separately for the two groups of data. For the primary sensor, there was no significant change between the two groups, so one calibration has been applied to both sensors. The standard deviation was higher than the mean offset. However, this was due to a general noisiness of the data set throughout the cruise, rather than a changing offset over time. For the secondary sensor, there was a change in the offset, suggesting that this sensor showed a slight drift with time.

Sensor	Titanium Casts	Calibration	N	stddev	BODC cal ref
Primary	All	PSALCC01 = PSALCU01 + 0.00267	67	0.0055	2763
Secondary	AMT12_09 to AMT12_39	PSALCC02 = PSALCU01 - 0.00433	42	0.002	2764
Secondary	AMT12_44 to AMT12_69	PSALCC02 = PSALCU01 - 0.00716	29	0.001	2765

Dissolved oxygen

The dissolved oxygen channels have been calibrated against water bottle Winkler data from Nikki Gist and Carol Robinson, Plymouth Marine Laboratory.

Stainless steel frame

The oxygen sensor on the stainless steel CTD frame was calibrated in three groups. This was due to an apparent drift of the sensor after 02/06/2003.

Stainless Steel Casts	Calibration	N	R ²	BODC cal ref
AMT12_01 to AMT12_40	DOXYSC01 = 1.00232 * DOXYSU01 + 12.8242	120	0.998	3245
AMT12_42, 43, 45 (suspect data) and 47	DOXYSC01 = 1.06772 * DOXYSU01 + 7.826	16	0.999	3246
AMT12_49 to AMT12_67	DOXYSC01 = 1.08694 * DOXYSU01 + 7.289	70	0.992	3247

Titanium frame

The oxygen sensor on the titanium frame showed no drift throughout the cruise. However, cast AMT12_11 deviated from the rest of the data set, so a separate calibration was obtained.

Titanium Casts	Calibration	N	R ²	BODC cal ref
AMT12_03 to AMT12_69 (AMT12_11 excluded)	DOXYSC01 = 1.16848 * DOXYSU01 + 5.1	47	0.997	3243
AMT12_11	DOXYSC01 = 1.1626 * DOXYSU01 + 20.07	10	0.935	3244

Fluorescence

The nominal chlorophyll-a values have been calculated from the fluorometer data (with manufacturer's calibration applied) from the up-cast at bottle firing and the fluorometric chlorophyll-a data from sampled bottles. There was some drift in the sensor at the end of the transect and the calibrations have been split to take this into account. There were some casts without sample data at the start and end of the cruise and without these data to quantify the level of drift, these fluorometer profiles were not calibrated.

Stainless steel frame

There were no sample data for casts 1 and 3, so no calibrations were applied to these profiles.

Stainless Steel Casts	Calibration	N	R ²	BODC cal ref
AMT12_05 to AMT12_58	CPHLPS01 = 1.3531 * CPHLPM01 + 0.0112	107	0.375	6630
AMT12_60 to AMT12_67	CPHLPS01 = 0.6722 * CPHLPM01 + 0.0339	20	0.508	6631

Titanium frame

There were no sample data for casts 2, 4, 6, 9, 68 and 69, so no calibrations were applied to these profiles.

Titanium Casts	Calibration	N	R ²	BODC cal ref
AMT12_11 to AMT12_62	CPHLPS01 = 1.5561 * CPHLPM01 + 0.0184	82	0.640	6632
AMT12_64 and AMT12_66	CPHLPS01 = CPHLPM01 + 0.0309 (±0.0182)	5	-	6633

Search the data