3.c. Data Descriptions, Corrections and Estimated Accuracy/Precision

a. Pressure: When on deck, the SBE-911 pressure sensor recorded values of less than 0.30 dbar. As the estimated "sensitivity" error in the pressure measurements is less than 0.15% of reading (SeaBird Application Note 27), the implied offset was deemed insignificant and no pressure corrections were made.

b. Temperature (IPTS-68): Comparisons between the primary and secondary temperature sensors (SeaBird SBE-3plus) resulted in a residual variance (Tpri-Tsec) of less than 0.0005 degC in selected profiles (excluding rapid change regions like the thermocline). Since the units were calibrated October 1996 and these inaccuracies are less than the manufacturer's specified temperature sensor "span errors" of plus or minus 0.005 degC (SeaBird Specification Sheet), the CTD temperatures needed no further correction.

c. Salinity (PSS-78): CTD salinity was computed by the SEASOFT program from measured conductivity (Sea Bird SBE-4C) and temperature using the practical salinity scale of 1980 (Fofonoff and Millard, 1983). The computed CTD salinities were corrected by comparison with Nisken bottle sample salinities collected during each cast. These bottle sample salinities were analyzed using an Autosal Analyzer. The standard deviation of all samples was 0.0057 psu.

Salinity ranged between 31 and 35 psu. Bottle salinities, plotted versus CTD-derived minus bottle salinity differences (Figure 3), provides us with the basis to reject two spurious data points (probably related to leaking sample bottles, strong salinity gradient, incorrect sampling, etc.). Plots and regressions of salinity differences versus pressure and bottle salinities suggested no apparent depth bias. The CTD salinities were an average of 0.0272 psu higher than bottle salinities. A linear regression of bottle salinity on CTD salinity yielded the corrected salinity (Scorr) equation:

Scorr = 0.9998 * Sctd + 0.0210 ,

where Sctd is the CTD-measured salinity. The standard deviation of discrepancies was 0.0059 psu. This equation was used to correct all of the SBE-911 CTD salinity data.

d. Dissolved Oxygen: An algorithm using oxygen sensor current and sensor temperature measured by a SeaBird SBE-23Y, along with CTD water temperature, salinity, and pressure, calculates dissolved oxygen concentration in milliliters / liter (SeaBird Application Note 13-1B). Published accuracy and resolution specifications are 0.10 mL/L and 0.01 mL/L, respectively.

e. Transmissivity: A Sea Tech unit measured transmission loss of a 25 cm beam of red light as it passed through the water column, and presented it as a percentage relative to the transmission loss of pure water.

f. Fluorescence: A Sea Tech unit measures flourescence emission excited by a beam of blue light on a sample of sea water. An algorithm is used to infer the sample's chlorophyll-alpha (Chl-a) concentration in micrograms / liter.

g. Irradiance (PAR): A Biospherical QSP-200PD measures omnidirectional photosynthetically available radiation (PAR) in the 400-700 nm range. An algorithm converts the instrument's photodetector current to microeinsteins / second / square meter to within 5% of available light.

These CTD data were post-processed using MATLAB algorithms (P. Morgan, 1994) which are based on Fofonoff and Millard (1983). The following additional variable profiles were computed:

a. Potential Temperature (Theta), a function of temperature, salinity, and pressure and relative to the surface, was computed using SW_PTMP.M.

b. Potential Density (Sigma Theta), a function of temperature, salinity, and pressure and relative to the surface, was computed using SW_PDEN.M.

c. Brunt-Vaisaila Frequency (N-squared), a function of temperature, salinity, depth and latitude, was computed using SW_BFRQ.M. This measure of vertical stability is large when vertical motion is suppressed and negative under unstable conditions.

d. Sound Velocity, a function of temperature, salinity, and pressure, was computed using SW_PTMP.M