Dataset: Biofloat_48 All Sensor Profiles
Deployment: Biofloat_48

Unless noted below, no filtering or de-spiking of float 48 sensor data has been performed.

Temperature, Potential Temperature, Salinity (parameter names temp, potemp, sal):  Float 48 had two CTD sensors: one on top of the float, the other on the bottom.  The CTD sensors on float 48, the R/V Knorr process cruise 193-03, the float deployment cruise on the R/V Bjarni Saemundsson and the pre- and post-cruise float sensor calibrations were compared. Formulae were found to bring them all into agreement to within 0.001 salinity using the Knorr CTD as the absolute standard.  A 3 point running median filter was applied to eliminate single spikes.  An algorithm to merge data from the two float CTD’s (top and bottom) were used to reduce the effect of other errors.  For more details, see CTD_float_Calibration-NAB08.pdf.

Particulate attenuation coefficient cp (parameter name beam_cp): Beam transmissometers from Floats 48 (CST-1063), the R/V Knorr cruise 193 (CST-284 and CST-1090), and the two cruises on the R/V Bjarni Saemundsson (CST-284, CST-1090) were compared and rectified to yield a consistent measurement of transmittance (%) and particulate attenuation coefficient due to particles cp at 652 nm.  This analysis allowed the application of a single correlation between cp and particulate organic carbon based on linear regression of cruise CTD measurements and bottle measurements of POC.  For more details, see C-Star_Calibration-NAB08.pdf.

Chlorophyll fluorescence (parameter name chl_raw_float) is reported as raw instrument counts without subtraction of dark counts.  Parameter name chl_raw is reported as instrument voltage output minus dark voltage (median in situ dark voltage = 0.083 volts); this parameter was intercalibrated with ship CTD, float and glider fluorometers and is reported as ship CTD volts.  

Chlorophyll a from fluorescence (parameter name chl_a_derived): Comparison of the chlorophyll fluorometers on the Knorr CTD and on float 48 with extracted chlorophyll from the R/V Knorr 193-03 water samples show clear depth and time dependences. Accordingly, a linear relationship between chlorophyll and fluorescence is abandoned in favor of a more complex, albeit empirical scheme. Counts from the float 48 fluorometer are converted to chlorophyll using a dark offset and a gain that has dependences on temperature, PAR, depth and yearday, with the gain and offset adjusted so that the float best matches the extracted chlorophyll at the R/V Knorr –float calibration casts (where nearly simultaneous profiles of the R/V Knorr CTD and float 48 were performed to 250 m). The resulting float 48 chlorophyll matches the bottles from both the R/V Knorr process cruise and R/V Bjarni deployment cruises with an error of 30-50%.  For more details see
Chlorophyll_Calibration-NAB08.pdf.

Dissolved oxygen (parameter names O2_cal): The Seabird SBE-43 oxygen sensor and the Aanderra Optode on float 48 both require calibration and removal of various sensor errors. The Optode is poorly calibrated in terms of dissolved oxygen, temperature and pressure.  During times of low flow and a closed drogue, it can have biases exceeding 10 μmol/kg, probably due to oxygen consumption by the float.  The SBE-43 exhibits mode-variable biases due to our attempts to reduce pumping energy.  By inter-comparing these sensors and comparing them with the Winkler bottle samples from the R/V Knorr and SBE-43 sensor taken on calibration casts during R/V Knorr cruise 193-03, adjustments were developed to bring both float 48 oxygen sensors and the R/V Knorr CTD SBE-43 sensor into agreement and into absolute calibration to an accuracy of better than 2 μmol/kg.  These corrections are applied to the float data. The best float 48 oxygen data was obtained from the SBE-43 in all float modes and the Optode in down and drift modes. The Optode data from settle and up mode still contains systematic errors.   For more details, see Oxygen_Calibration-NAB08.pdf.

Nitrate (parameter names nitrate): The Satlantic in situ ultraviolet spectrophometer ISUS V3 sensor was deployed on Float 48.  Post calibration of the ISUS instrument and subsequent processing of the spectrophotometric data acquired by the ISUS incorporating in situ temperature and salinity measurements to correct for the temperature-dependent, ultraviolet absorbance by the bromide ion results in a precision of duplicate nitrate measurements of 0.1μM. These nitrate concentrations were then calibrated against bottle measurements collected during the R/V Knorr process cruise (April 17 through May16) and subsequently analyzed for nitrate plus nitrite (NO3-+ NO2-) concentration.  A simple linear regression was sufficient to make the ISUS nitrate concentrations (NO3ISUS) fit the bottle for nitrate plus nitrite measurements with a standard deviation of 0.6μM:
      nitrate = 1.1536 x NO3ISUS + 2.6227    (r2 = 0.927, n = 53)
Comparisons with water samples taken on the R/V Bjarni Saemundsson deployment cruise are inconclusive due to high spatial variability.  No obvious drift of the ISUS response with time or depth was apparent.  For more details, see ISUS_Nitrate_Calibration-NAB08.pdf.

Particulate organic carbon (POC) derived from cp (parameter name POC_cp):  A relationship between Ship-based cp (R/V Knorr 193-03) and bottle POC data was used to produce a regression for application to CTD downcast cp.  The ship and float WET Labs C-Star beam transmissometers were intercalibrated (see above).  Therefore, the same cp-POC regression as used for the R/V Knorr 193-03 cruise can applied to float 48 cp data (parameter name cstar.beamc):
POC (mg C m-3) = cp (m-1) * 408 (mg C m-2) – 10 (mg C m-3); r2 = 0.80,
This relationship is empirical and it should not be applied to the other cruises without independent verification. For more details, see the datasets associated with KN193-03 and POC_cp_bbp_Calibration-NAB08.pdf.

Particulate backscattering coefficient, bbp (parameter name bbp700):  Backscattering voltage was converted to Β at 140° by subtracting a factory-determined dark voltage and multiplying by a factory calibration scale factor, modified based on measurements and calculations of Sullivan et al. (subm.), parameter name flntu.beta700.  The calibration factor did not change significantly between pre-cruise and post-cruise factory calibration.  Β  at 140° was converted to bbp (m-1) by subtracting Β  of seawater (Zhang et al., 2009) and multiplying by 2π  (where  Β = 1.132) to produce parameter name flntu.bbp700cal0.  For the float 48, the resulting bbp700cal0 value was adjusted to fit the R/V Knorr FLNTU, producing flntu.bbp700 using the following intercalibration equation:
      flntu.bbp700 = 1.040 * flntu.bbp700cal0 – 0.000790
See calibration report for more details: Backscatter_Calibration-NAB08.pdf.

Particulate organic carbon (POC) derived from bbp (parameter name POC_bbp):  A relationship between Ship-based bbp (R/V Knorr 193-03) and bottle POC data was used to produce a regression for application to CTD downcast bbp.  The ship and float WET Labs FLNTU backscattering meters were intercalibrated (see above).  Therefore, the same bbp-POC regression equation as used for the R/V Knorr can be applied to Float 48 bbp(700) data (parameter name flntu.bbp700).
      POC (mg C m-3) = bbp(700) (m-1) * 35800 (mg C m-2) – 16.2 (mg C m-3); r2 = 0.81.
This relationship is empirical and it should not be applied to the other cruises without independent verification. For more details, see the datasets associated with KN193-03 and POC_cp_bbp_Calibration-NAB08.pdf.
In situ photosynthetically active radiation (PAR; parameter name PAR):  Using factory-supplied calibration factors, the LI-COR 192SA PAR sensor has a stated NIST-traceable accuracy of +/- 5% in air, stability of less than +/- 2%. Reported values for in situ PAR from the float 48 LI-COR 192-SA sensor do not match those computed by spectral integration from the TriOS hyperspectral downwelling irradiance (ACC) sensor (parameter name acc.par); a systematic bias exists between the two PAR values.  Further, using matchups when the float was at the surface, neither LI-COR PAR or TriOS PAR match PAR values computed from various R/V Knorr 193-03 PAR sensors.  Shading of the LI-COR sensor is possible by the top-mounted CTD, GPS/Argos antennas, and float ring supports; effects of this shading has not been quantified.  The effects of self-shading due to the 1 m square drogue deployed during drift mode have not yet been quantified.  For more details, see Radiometry_and_PAR_Calibration-NAB08.pdf.

In situ hyperspectral downwelling plane irradiance Ed(λ ) and hyperspectral upwelling radiance Lu(λ ) , parameter names Ed and Lu, respectively): The TriOS ACC Ed sensor measured hyperspectral downwelling plane irradiance from 320-950 nm.  The TriOS ARC Lu sensor measured upwelling radiance from 320-950 nm.  Both sensors sampled at ~3.3 nm spectral resolution and 0.3 nm spectral accuracy.  From the raw 12 bit digital values, calibrated irradiance Ed(λ ) (W m-2 nm-1) and radiance Lu(λ ) (W m-2 sr-1nm-1) was computed from the ACC and ARC sensors, respectively, using pre-cruise factory calibration coefficients and the equations provided by TriOS (TriOS 2004).  ACC Ed and ARC Lu sensor measurement wavelengths are not identical; no spectral interpolation has been performed.  Two data quality parameters are available:  tilt and qflag.  The parameter “tilt” indicates the tilt of the sensor (and float) in degrees from vertical, while qflag indicates data quality: 1 = acceptable, 0 = biofouling suspect. The ACC Ed sensor appears to be biofouled during days 120 – 142; while no biofouling correction has been applied to the radiometric data, qflag is set to zero during this time.  Shading of the ACC (Ed(λ )) by the top-mounted  CTD, GPS and Argos antennas, ring supports is possible; effects of this shading have not been quantified but will appear as variability in the measurements rather than bias.  The effects of self-shading due to the 1 m square drogue deployed during drift mode have not yet been quantified.  For more details, see Radiometry_and_PAR_Calibration-NAB08.pdf.
 

* The float suffered a software failure on day 104.  From approximately yearday 104.124 to 104.685, float 48 drifted at the surface; no data was collected during this time.  Prior to yearday 104.124, only reduced temporal resolution (“Quicklook”) sensor data is available for float 48 sensors, except the CTD where full resolution data is available.  After yearday 104.685, full resolution data is available for all sensors.During the first 2 days of the mission (yeardays 95, 96) Float 48 completed two cycles per day.
* Biofouling considerations:  Pumped seawater output from the lower conductivity cell, containing a biocide, periodically flushed the up-facing C-Star transmitter pressure window before each measurement.  Pumped seawater output from the lower conductivity cell, containing a biocide, was used for ISUS nitrate measurement.  The FLNTU chlorophyll fluorometer & backscattering sensor included a copper shutter with rubber wiper which was used to protect against biofouling.  The ACC Ed sensor was biofouled during days 120 – 142; no biofouling correction has been applied to Ed data. 
* The MLFII floats used in NAB08 will almost certainly spin about their vertical axis. The drogues have a pinwheel aspect to them when retracted during profiling.  In another experiment where an MLFII float was equipped with an accelerometer, the float spun with a rotation period during profiling of about 80s.  During drift mode when the drogue is deployed, the float will also spin, with the spin rate probably more irregular and depending strongly on the turbulence levels.  A spin rate of 1 RPM would not be unusual.

Related files and references: 
[Include any useful supporting documents; e.g. separate files or published papers with description of sampling and analytical methodology]

gps-float48-2010-10-14-v8.mat
ctd-float48-2010-10-14-v8.mat
ed-float48-2010-10-14-v8.mat
lu-float48-2010-10-14-v8.mat
licor-float48-2010-10-14-v8.mat
ctsar-float48-2010-10-14-v8.mat
flntu-float48-2010-10-14-v8.mat
optodeO2-float48-2010-10-14-v8.mat
seabirdO2-float48-2010-10-14-v8.mat
isus-float48-2010-10-14-v8.mat

NAB08Float48SamplingMethodologyV1.pdf

Calibration reports:
CTD_float_Calibration-NAB08.pdf
C-Star Calibration-NAB08.pdf
Chlorophyll_Calibration-NAB08.pdf
Oxygen_Calibration-NAB08.pdf
ISUS_Nitrate _Calibration-NAB08.pdf
POC_cp_bbp_Calibration-NAB08.pdf
Backscatter_Calibration-NAB08.pdf
Phytoplankton_Carbon-NAB08.pdf Radiometry_and_PAR_Calibration-NAB08.pdf

References:

Garcia, H. E. and L. I. Gordon, 1992.  Oxygen solubility in seawater: Better fitting equations.  Limnology & Oceanography, Vol. 37 (6), 1307-1312.

TriOS, 2004.  RAMSES hyperspectral radiometers manual, release: 2004-08-16.

Sullivan, J. M., M. S. Twardowski, J. R. V. Zaneveld and C. C. Moore. In-water measurement of optical backscattering at fixed angles.  Submitted to Applied Optics.

Zhang, X., L. Hu and M. He.  2009.  Scattering by pure seawater: Effect of salinity. Optics Express 17 (7), 5698-5710.
Version: 15 October 2008