The CTD/rosette package is based on a Sea-Bird Electronics (SBE) 911 plus, equipped with the following sensors:

• Pressure (Digiquartz Paroscientific, Inc.),
• Temperature (SBE 3),
• Conductivity (SBE 4),
• Dissolved oxygen (SBE 13 until January 2003, and then an SBE 43), and
• Altimeter (Datasonics PSA 900).

The sensors are mounted on a SBE 32 rosette with 11 12L Niskin bottles (the 12th being replaced by an AC-91 as discussed below).

The pressure sensor is not regularly calibrated (last calibration in 1995), it is known from past experience with this type of sensors that the uncertainty is on the order of 0.5 dbar. It is indeed acknowledged that this type of sensor only experience a drift of the zero, which is easy to control and to correct for when the sensor is at the surface.

The temperature sensor is sent once a year for calibration to SBE, which has the agreement from the NIST (National Institute of Standards and Technology) for this type of work. Fig. 11 provides an example of the calibrations performed in January 2002 and January 2003 (was about the same between January of 2001 and January of 2002). The drift is of the order of 0.0025^oC over one year, and it is linear. Therefore, any temperature value computed between two of these calibrations is provided with a ±0.0025^oC uncertainty.

The conductivity sensor is sent once a year to SBE for calibration. Additional calibration against values determined from water samples and subsequent lab analyses (with a lab salinometer) are not performed. Fig. 12 provides an example of the calibration performed in 2001. The change of the calibration slope is due to biological fouling that grow inside the sensor, modifying its geometry. The influence of this fouling is larger as the conductivity increases. Typical values of the BOUSSOLE site are of the order of 5Siemens m^-1. It is therefore estimated that the sensor drift over one year is of the order of 0.0005 in conductivity units, which is about 0.005 in salinity units. The drift was a bit higher in 2002, around 0.001 in conductivity units, which is about 0.01 in salinity units.

From July 2001 to December 2002, a SBE 13 sensor was used, with a Beckman membrane. This sensor showed a large drift. Starting in January of 2003, a new sensor is used (SBE 43), which is provided with a nominal 5% uncertainty and a less than 2% drift. No Winkler-based calibration are performed on these sensors.

At each cast, a first round trip down to 50m is performed with the CTD, in order to rinse all instruments and equilibrate them in temperature. Then the package is lowered down to 400m at a descent rate of about 0.5ms^-1. The data are collected at a 24Hz frequency, and then processed through the SeaSoft software until binned values are produced at a 1m resolution. To each data file is associated a header file that contains the necessary ancillary information (latitude, longitude, date and time of acquisition, meteorological information, etc.).

Some of the sampling depths are fixed (i.e., 5 and 10m for surface analyses) and others are determined at the end of the rosette descent, based on the phytoplankton fluorescence profile, in order to best represent any specific features of the vertical profile. No sampling is done deeper than 200m. Bottles are closed during the ascent of the rosette. When back on deck, 2.8L plastic bottles are used to collect water that is used for subsequent filtration. Filters are then stored in Petri slides, wrapped in aluminum and stored first in liquid nitrogen on board then in a -80^oC freezer in the laboratory. The material collected on filters is then used for HPLC-based analyses of the phytoplankton pigments (Sect. 7.1) and filter pad absorption (Sect. 7.2). For the surface samples, triplicate filtration are performed for the HPLC analyses.

Back to top

Calibration drift of the temperature sensor mounted onto the CTD.

 

Calibration drift of the conductivity sensor mounted onto the CTD.

Back to top

A Chelsea MK III Acquatracka fluorometer is installed onto one of the analog channels of the CTD SBE 911. This instrument is equipped with an excitation filter at 430nm and an emission filter at 680nm. The data are provided both as voltages and chlorophyll concentrations, the latter being computed with the initial calibration coefficients provided by the manufacturer. These concentrations are just indicators. Another calibration of the fluorometer signal can be performed against the HPLC-derived discrete chlorophyll concentration values.

Back to top

One of the rosette bottles is replaced by a 25cm path AC-9 plus sensor. This instrument measures the absorption and attenuation coefficient of the water at nine wavelengths (i.e., xxx). In order to visualize and collect the data in real time, this AC-9 is linked to the surface through a 19,200 baud serial connection that as well provides power to the instrument. Although the two sensors can be independently controlled, only one computer is used in order to provide the same time frame and depth measurements to both the CTD and the optics data. The connection is only working in the computer-to-AC-9 direction, although theoretically both ways are possible. This is due to the different transmission speeds (300 baud from the computer to the SBE 911 and 19,200 from the SBE 911 to the AC-9).

Figure 13 displays the cabling of the complete system, including the AC-9 and the other optical sensors.

A SBE 25 pump is installed at the entrance of the AC-9 water circuitry, and it is coupled with the pump that provides water to the conductivity sensor. Both are only functioning under water, when a certain conductivity threshold is reached.

The AC-9 is programmed as to start the following acquisition sequence immediately after being powered by the CTD sensor: 15s time lag before starting acquisition, then a 60s warming period, then a 60s period allowing for cleaning of the water circulation, and then a 35min data collection period. The latter is enough for the round-trip down to 400m, including the stops for bottle firing.

Note that the starting sequence of 135s is occurring during the initial round trip down to 50m. This procedure ensures a progressive and full temperature rise of the electronics and a correct water circulation into the instruments.

Data collection is performed by the WetView software v5.0a, as provided by the manufacturer.

Cabling of the AC9 – CTD system.

A WETLabs ECO-BB3 backscattering meter is connected onto one of the auxiliary ports of the AC-9 plus (serial port with a 19,200 baud speed). This sensor emits light at 440, 532, and 650nm, and measures the backscattered light at the same three wavelengths. Power is also provided by the CTD. The data cannot however be visualized in real time. They are stored into the AC-9 plus memory, and then collected once the rosette is back on the ship's deck and the 35min acquisition sequence of the AC-9 is closed.

A WETStar CDOM fluorometer is plugged into one of the analog ports of the CTD. This instrument has an excitation wavelength at 370nm and an emission filter at 460nm. It is supposed to provide a qualitative measure of the Colored Dissolved Organic Matter (CDOM). It is connected to the CTD water circulation; the water goes from the conductivity sensor to the dissolved oxygen sensor, then to the CDOM fluorometer and to the pump. The data are collected in parallel to the CTD data and processed in real time through the SeaSoft software. The raw signal is provided as well as a CDOM concentration computed with the calibration coefficients provided by WETLabs.

Back to top

The Satlantic SPMR (SeaWiFS Profiling Multichannel Radiometer) free-fall radiometer is used at LOV since 1995, and in many other laboratories around the World for now 8 years. The LOV version of this profiling instrument is equipped with two irradiance sensors, collecting the upwelling and downwelling plane irradiances at the following 13 wavelengths: 412, 443, 456, 490, 510, 532, 560, 620, 665, 683, 705, 779, and 865nm (the latter was replaced by 380nm after July 2003).

The SPMR Profiler is made of a long pressure case (1.2m, 9cm diameter) that contains the majority of the system electronics, while the optical sensors are located and separately housed at either end of the case and facing in opposite directions (i.e., up and down). The top end of the instrument has buoyant fins to stabilize the instrument's underwater free fall deployment, and the bottom end has a small annular lead ballast to further stabilize the orientation and provide for fine tuning of the free-fall velocity.

The light sensors used for measuring irradiance (in units of milliwatts per square centimeter per nanometer), have a black Delrin plate on the end. The plate contains 13 specially-designed, diffuser-based, cosine collectors. Tilt and pressure are recorded at the same frequency as the irradiance measurements (6Hz).

The SPMR is accompanied by a deck reference sensor, called the SeaWiFS Multichannel Surface Reference (SMSR). This sensor is equipped with the same 13 wavelengths as the SPMR, and is based on the same electronics. Data acquisition is synchronized between the SPMR and the SMSR and it is performed again at the same 6Hz frequency.

The absolute calibration (Sect. 6) of the SPMR and SMSR with respect to NIST-traceable standards is performed every six months in the Satlantic optics calibration laboratory, and it is tracked between these absolute calibrations using at the LOV the ultra-stable portable light source developed for that purpose by Stalantic, i.e., the SeaWiFS Quality Monitor, 1=SQM-II1 (Hooker and Aiken 1998). Combining these two elements allows a 3% maximum uncertainty to be maintained on the calibration of the SPMR and SMSR.

An SPMR profile starts when the instrument has reached a distance of 50m off the ship stern (the ship is 25m long and a mark on the cable indicates the 50 distance). The instrument is then released and falls at approximately 0.5ms in the water column, collecting data at a 6Hz frequency. The descent is stopped when the pressure sensor indicates a depth of about 150m, except in extremely clear waters where the profile is performed down to 200m. This technique allows to steer clear of the ship shadow, and to get measurements with tilt angles less than 2. The sun is usually on port side of the ship, which is anyway not so important precisely because the ship shadow is not affecting the measurements.

Back to top

The SIMBADA instrument is an above-water radiometer designed and manufactured by the University of Lille, France. It is an upgraded version of the SIMBAD (reference ?) above-water radiometer. It measures both water-leaving radiance and aerosol optical thickness in 11 spectral bands centered at 350, 380, 412, 443, 490, 510, 565, 620, 670, 750, and 870nm by viewing the ocean surface (ocean-viewing mode) and the sun (sun-viewing mode) sequentially. The same optics (field-of-view of 3^o), interference filters, and detectors are used in both ocean-viewing and sun-viewing mode. Different electronic gains are used for each mode. The optics are fitted with a vertical polarizer, to reduce reflected skylight when the instrument is operated in ocean-viewing mode. A small GPS is fixed in the front panel of the SIMBADA for automatic acquisition of geographic location at the time of measurement. Viewing angles are recorded automatically.

Viewing of the ocean must be made in clear sky conditions (3/4 of sky cloudless, and no clouds obscuring the sun), outside the sun glint region (relative angle between solar and viewing directions of 45-90^o), and at a nadir angle of about 45^o. For those angles, reflected skylight is minimized as well as residual ocean polarization effects. The measurements can be made on a vessel underway, so there is no need to stop the ship to make measurements. To normalize water-leaving radiance, incident solar irradiance is not measured, but computed using the aerosol optical thickness.

The operator can select, in addition to ocean-viewing and sun-viewing modes, dark current and calibration modes. Each series of measurements lasts 10s. Frequency of measurements is about 8Hz. Data is stored internally and downloaded onto diskette at the end of the day or a cruise. The instrument is powered by batteries that can be charged using a main supply of 110-240V, 50-60Hz. About 2h of charging should be enough for one day of measurements.

The SIMBADA has been used from the bow of the ship or from its upper superstructure when the weather was allowing to do so, which were the two more convenient locations where we were able to be sufficiently high above the sea surface. This is important, because any ship shadow effect or perturbation from reflection on the ship superstructure is minimized when increasing the vertical distance between the operator and the sea surface, and then the horizontal distance between the point of the sea surface which is aim at and the ship hull.

The full sequence of measurements was as follows:

1. Three dark current recordings, each 10s.
2. Three sun viewing, each of 10s (for derivation of the aerosol optical thickness).
3. Three or four sea viewing, each of 10s.
4. Three sun viewing, each of 10s (for derivation of the aerosol optical thickness).
5. Three dark current recordings, each 10s.

Back to top