A. Cruise narrative

In 1995, the NOAA Ship MALCOLM BALDRIGE carried out three WOCE Indian Ocean cruises. As its primary WOCE objective, the BALDRIGE revisited lines occupied for the one-time survey, also in 1995. The reoccupations were designed to capture the large seasonal variability in the basins related to the reversing monsoon winds. Together, the repeat and one-time WHP surveys are used to estimate monsoon variations of the thermohaline overturning, meridional and zonal flows, and heat and freshwater fluxes. These processes have been selected as fundamental to increasing the understanding of the Indian Ocean's role in climate.

B. Description of Measurement Techniques and Calibrations

The pressure sensor was calibrated by using the pre cruise laboratory calibration with a linear offset drift of approximately 0.5 db/year. The linear offset drift was determined by analyzing CTD pressure measurements at the sea surface.

Pre and post cruise laboratory calibrations were obtained for the temperature sensors. The temperature sensors were calibrated using both pre and post cruise laboratory calibrations with a linear offset drift over time determined from the laboratory calibrations. The reported temperature is an average of the two independently calibrated temperature sensors used on each cast.

Pre and post cruise laboratory calibrations were obtained for the conductivity sensors. The conductivity sensors were calibrated using both pre and post cruise laboratory calibrations, with slope and offset drifts determined from the rosette bottle salinity measurements and the uptrace conductivity sensor measurements. To determine slope and offset drifts, all good bottles below 300 db and within + and - 5 days of each station were used. The nominal Seabird temperature and pressure corrections for the conductivity sensors were used. The calculated drifts were smoothed by a 5 station running mean (Figures 1a and 1b). The reported salinity is an average of the salinities calculated from the two calibrated conductivity-temperature sensor pairs used on each cast. A small temperature dependency in the surface values and a small pressure dependency in the deep values remain in the final data. However, the above procedure produced the best overall fit to the rosette bottle salinity measurements. In a minority of stations there was a problem with one of the conductivity sensors. In these cases, the reported salinity values are only determined from the optimally performing conductivity sensor. The most significant case was the failure of the "c1" conductivity sensor between stations 268 and 291.

The oxygen sensor was calibrated by using the pre cruise laboratory calibration, with slope and offset drifts determined from the rosette oxygen measurements and the uptrace oxygen sensor measurements. A better overall fit was obtained when using the uptrace oxygen sensor measurements, rather than the downtrace measurements as is often the procedure for the oxygen calibration. To determine slope and offset drifts, all good bottles within + and - 5 days (usually) of each station were used. Rather than using the Seabird nominal temperature and pressure corrections for the oxygen sensor, the values were adjusted slightly for each sensor. The calculated slopes and drifts were smoothed by a 5 station (usually) running mean (Figure 2).

Pressure plots and histograms of the differences between the calibrated CTD/O2 sensors and the rosette bottle measurements are shown for all stations for salinity (Figure 3) and for oxygen (Figure 4).

The 2 db reported temperature, salinity, and oxygen profiles are uptrace values, not the more typically reported downtrace values. This resulted in a better overall calibration for the oxygen profiles, and the uptrace surface values are not affected by a slow pump turn on as is the case for the downtrace surface values. In addition, a few of the stations had intervals of spurious values on the downtraces, but not on the uptraces. However, the profiles and bottle sensor values of station 292 are all downtrace values, as the uptrace data was lost. Finally, station 222 is filtered by a 25 m running mean, as there seems to have been inadequate water flow past the sensors during this cast.

On-station velocity profiles were obtained using a RDI 150 kHz Broadband ADCP (Lowered or LADCP) mounted looking downward from the CTD frame. This technique measures and records velocity shear profiles extending 150 to 350 meters below the instrument approximately once per second. In post-processing, the individual shear profiles are averaged by depth to produce a full-depth shear profile, which is integrated to estimate the depth dependent (baroclinic) component of the velocity field. The depth-independent (barotropic) component of velocity can be recovered if positions at the start and end of the cast are known; positions were logged on this cruise using a Trimble Centurion P-code GPS receiver, accurate to 5 - 10 meters. Readers are advised to refer to Fischer and Visbeck (1993) for a full explanation of methods and standard processing procedures. Errors appear to be somewhat dependent on deep ocean acoustic scattering conditions, but past comparisons with concurrent Pegasus dropsonde profiles [P. Hacker, et al., unpublished] have shown that LADCP estimates of barotropic velocity are commonly accurate to better than 1 cm/s. Based on the same comparisons, baroclinic velocity is estimated have rms (calculated over an entire profile) errors of less than 5 cm/s.

A more complete description of the ADCP and LADCP data is reported separately.

Fischer, J., and M. Visbeck, Deep Velocity Profiling with Self-contained ADCP's, Journal of Atmospheric and Oceanic Technology, 10, 764-773, 1993.

Hacker, P., E. Firing, W. D. Wilson, and R. Molinari, Direct observations of the current structure east of the Bahamas, Geophysical Research Letters, 23, 1127-1130, 1996.

The following are the abbreviations used in the water sample table: PR = pressure (db), TE = temperature (oC), SA = salinity, RS = rosette salinity, OX = oxygen (umol/kg), RO = rosette oxygen (umol/kg), QC = quality control, NK = niskin bottle.

The first QC flag is for the water sample: 1 = bottle information unavailable, 2 = no problems noted, 3 = leaking, 4 = did not trip correctly, 5 = not reported, 6 is not applicable here, 7 = unknown problem, 8 is not applicable here, and 9 = samples not drawn.

The second QC flag is for the rosette salinity, and the third QC flag is for the rosette oxygen: 1 = sample drawn, but analysis not received, 2 = acceptable measurement, 3 = questionable measurement, 4 = bad measurement, 5 = not reported, 6 = mean of replicate measurements, 7 is not applicable here, 8 is not applicable here, 9 = sample not drawn. Rosette salinity and oxygen measurements falling within 3 standard deviations of the mean CTD/O2-bottle differences are flagged with a 2. Rosette salinity and oxygen measurements falling between 3 and 6 standard deviations from the mean CTD/O2-bottle differences are flagged with a 3. Rosette salinity and oxygen measurements falling outside 6 standard deviations of the mean CTD/O2-bottle differences are flagged with a 4. Note that this standard deviation criteria assigns all error flags to the rosette measurements, and never to the CTD/O2 measurements.

C. Reported Data