Hydrographic data from the GEF Patagonia cruises

Instruments Data Provenance & Structure


Introduction
The Argentine continental shelf is one of the largest shelf areas in the world ocean and comprises the Patagonian Shelf Large Marine Ecosystem (PLME; Heileman, 2009).The Atlantic Patagonia continental shelf extends from 55 • S at the tip of Tierra del Fuego to approximately 39 • S. The shelf is a shallow submerged plateau, which is very wide in the south (∼ 850 km) and narrows toward the north.The offshore edge is marked by a sharp change in bottom slope located at 115-240 m depth (Parker et al., 1997).This region is one of the most productive in the Southern Hemisphere and supports a wide variety of marine life (Falabella et al., 2009).In situ estimates of primary production in austral spring range between ∼ 200 mg C m −2 d −1 and > 3000 mg C m −2 d −1 near frontal regions (Lutz et al., 2010).The high biological productivity of the PLME sustains intense fishing activity, mostly by Argentine fleets but also by other international fleets (Heileman, 2009).In addition, this large primary production leads to the absorption of large quantities of carbon dioxide from the atmosphere (Bianchi et al., 2005) accounting for about 1 % of the global ocean's net annual CO 2 uptake, almost four times the mean rate of CO 2 uptake of the global ocean (Bianchi et al., 2009).
The high production is mostly associated with various shelf and shelf-break fronts generated by strong winds, largeamplitude tides, large buoyant discharges and the proximity of the nutrient-rich Malvinas Current (e.g., Acha et al., 2004;Saraceno et al., 2005;Palma et al., 2008;Matano and Palma, 2008;Matano et al., 2010).To determinate the seasonal variability of physical, chemical and biological properties and improve the knowledge base of the Patagonia marine environment and its biodiversity, three oceanographic cruises were carried out on board R/V ARA Puerto Deseado as part of the Global Environment Facility (GEF) Patagonia project (Fig. 1).The cruises were carried out in October 2005 (GEFPAT-1) and March (GEFPAT-2) and September 2006 (GEFPAT-3).Each survey consisted of the occupation of seven to nine cross-shelf sections from the nearshore area to the upper slope of the western Argentine Basin close to the 2000 m isobath.The cruise design provided quasisynoptic observations of the nearshore tidal fronts, the midshelf region, the shelf-break front and the western edge of the Malvinas Current (e.g., Romero et al., 2006).We briefly describe procedures of acquisition and processing of vertical conductivity-temperature-depth (CTD) profiles and underway surface temperature and salinity data.

CTD profiles
At each station a vertical quasi-continuous CTD profile was collected with a Sea-Bird Electronics model 911plus unit, equipped with fluorescence and turbidity sensors in GEFPAT-1, an oxygen sensor in GEFPAT-2, and oxygen, fluorescence and turbidity sensors in GEFPAT-3.Additional redundant temperature and conductivity sensors were used in some stations during GEFPAT-3.Table 1 summarizes the CTD sensors used in each cruise.Most vertical profiles reached to within ∼ 5 m from the bottom within the continental shelf and 10 m from the bottom at stations deeper than 200 m, except under adverse weather conditions or when the distance of the package from the bottom was uncertain, such as over regions with a steep bottom slope.The CTD was mounted with a rosette sampler and the package was deployed on a conducting cable, which allowed for real-time data acquisition and display on board.A General Oceanics (model GO 1015) 12-bottle water sampler was employed in GEFPAT-1 whereas a Sea-Bird Carousel (model SBE 32) 24-bottle water sampler was employed in GEFPAT-2 and GEFPAT-3.Both models held 5 L Niskin bottles.Duplicate CTD casts were carried out in GEFPAT-1 to collect water samples for ancillary biological programs.Duplicate casts were identified by station file names with the suffix b.Downcast profile data were reported because during downcast the CTD sensors sample the water column with minimal interference from the underwater package.However, in some stations that presented noisy data during the down-cast, up-cast data were reported.Down-cast (up-cast) file names were prefixed by d (u).Station date and times are reported in UTC.

CTD data processing
CTD data were post-processed according to common standards, using Sea-Bird Data Processing software routines (Seasoft-Win32, http://www.seabird.com/software/sswin. htm, SBE, 2005).The nominal calibrations were used for data acquisition.Final conductivity calibration was determined empirically by comparison with the salinities of discrete water samples taken during each up-cast.Conductivity and dissolved oxygen sensors were calibrated as described in the following sections.Fluorescence and turbidity data are reported based on factory calibrations only.
Data were subsequently averaged at 1 dbar pressure intervals.The data for each cast were inspected and any remaining density spikes removed by linear interpolation of the original temperature and conductivity data and all derived parameters recalculated at that level.

Temperature
Temperature sensor SN031691 used in GEFPAT-1 was factory calibrated in 2010.The pre-and post-cruise calibration data were used to determine the sensor drift, which was used for sensor calibration (SBE, 2010).The laboratory calibrations showed a drift correction of −0.00037 • C year −1 , corresponding to a temperature offset correction of 0.004 • C for GEFPAT-1 data.
The temperature sensor SN032951 used in GEFPAT-2 and GEFPAT-3 was factory post-calibrated in 2009.The resulting temperature drift correction was −0.00006 • C year −1 corresponding to an offset correction of 0.00004 • C and 0.00007 • C for each cruise, respectively.These temperatures offset corrections are lower than sensor resolution (0.0002 • C).Thus, the GEFPAT-2 and GEFPAT-3 temperature data do not require calibration for drift.
Temperature sensor SN031689 (secondary sensor for GEFPAT-3) was not factory calibrated after the cruise.

Conductivity
On board calibration of the conductivity sensor was performed empirically by comparing its nominally calibrated output against the calculated conductivity values obtained from water sample salinities using the pressure and temperature of the CTD at the time of bottle closure.An estimate of bias (offset) and slope corrections to the nominal factory calibration were determined using a linear least squares fit.For each cruise, bottle conductivities were fitted to CTD conductivities and conductivity differences greater than 2 standard deviations from the fits were rejected.
Vertical profiles of salinity differences between CTD and bottle salinity were plotted to identify suspect water samples caused by leaky Niskin bottles or drawn from regions of relatively large vertical salinity gradients.Also, potential temperature-salinity (θ-S) diagrams of historical hydrographic data collected in the same region were overlaid to check for consistency.All suspect bottle data were discarded and not used in the CTD calibration process described above.CTD observations of GEFPAT-2 were collected during a late austral summer, when mid-shelf waters present a strong vertical stratification associated with vertical temperature gradients on the order of 1 • C m −1 .Across these intense temperature gradients, spurious CTD salinity spikes were frequently observed.Salinity spikes were removed based on the comparison with bottle salinities obtained at selected stations where water samples across the thermocline were obtained at ∼ 1 m resolution.In stations without high-resolution bottle sampling, both down-and up-casts and water sample salinities were combined to reconstruct salinity profiles.The reconstructed data were inspected to check for density inversions, which were removed and filled in by linear interpolation and all derived parameters were recalculated at the interpolated level.Station file names of reconstructed profiles were renamed as the station file name with the suffix re.
To illustrate the quality of the conductivity calibration, Fig. 2 displays the salinity residuals before and after calibration and Table 2 summarizes the comparisons between CTD and water sample salinity for each cruise.

Dissolved oxygen
Calibration of the oxygen sensor was performed using a statistical method estimating calibration coefficients for calculating dissolved oxygen in milliliters per liter (mL L −1 ) from SBE 43 output voltage.The technique requires dissolved oxygen concentrations reported in mL L −1 determined from a range of Winkler-titrated water samples and SBE 43 oxygen voltage outputs measured at the times the water samples were collected (SBE, 2002).Though the sensor manufacturer recommends advancing the oxygen voltage data relative to the CTD pressure (SBE, 2005), we carried out several tests and concluded that this alignment led to a larger dissolved oxygen mismatch between CTD and water samples across the thermocline.Thus, no alignment corrections were applied.
The oxygen from water samples was compared with historical data collected in the region to check for consistency and to identify suspicious data.Historical data were obtained from the Argentine Oceanographic Data Center (CEADO, http://www.hidro.gob.ar/ceado/Fq/extrnac. asp#nacionales).CEADO archive data came from Argentine and international research institutions.The standard deviation of the residuals was approximately 1 µmol kg −1 .Figure 3 presents the differences between SBE 43 dissolved oxygen before and after calibration and Winkler-titration dissolved oxygen.

Water sample analysis
Water samples at selected levels were taken from 5 L Niskin bottles for the determination of salinity and dissolved oxygen.Salinity and dissolved oxygen were determined on board.Salinity samples were collected in 200 mL glass flasks and salinity was determined with a Guildline Autosal 8400B salinometer.The Autosal standardization was carried out with Ocean Scientific International Ltd. (OSIL) standard seawater (SSW) batches P130 (1996) and P141 (2002) for GEFPAT-1, P141 (2002) andP146 (2005) for GEFPAT-2 and P146 (2005) for GEFPAT-3, according to the procedure described in the salinometer technical manual (Guildline, 2004).To test the possible effects of the aging of standard seawater we standardized the instrument with more recent batches (P146, 2005) and then ran samples of P130 (1996).The tests indicated that the conductivity ratios of these batches were within 3 × 10 −6 of the value indicated by the manufacturer.This indicates that our salinity determinations were not affected by the age of the standard seawater.Salinity values were calculated and reported based in practical salinity units (PSS78, UNESCO, 1981).The double conductivity ratio of SSW during GEFPAT-3, showed a positive trend with time determined from the difference between the beginning and the completion of each run of samples.In order to determine whether the observed drift was due to alteration of the SSW or due to instrument drift, the instrument was standardized with a new vial of SSW at the beginning and at the end of one run.This test revealed an estimated Autosal drift of double conductivity ratio of 0.00013 over a period of 5 h, which is approximately equivalent to a rate of change in salinity of 0.00048 h −1 .In addition this test indicates an SSW alteration of 0.0001 in the double conductivity ratio, equivalent to a salinity change of 0.002 during the 5 h period.These estimates can be considered upper error limits as other salinity sample runs took about the same or shorter time.To account for the salinometer drift for each run a linear trend correction was estimated by a least squares fit and salinity from seawater samples corrected by removing the spurious trend.In GEFPAT-1 and GEFPAT-2 cruises the double conductivity ratio of SSW presented no trends.
Dissolved oxygen was determined with a modified Winkler method (Carpenter, 1965) using an amperometric endpoint detection technique.In the three GEF Patagonia cruises, the dissolved oxygen concentrations were determined with a Mettler DL 21 automatic titration system.Samples were collected in ∼ 125 mL volume-calibrated borosilicate glass flasks and whole-bottle titration was carried out in the analysis.Standardizations were performed with commercial potassium iodate solutions (0.01 N) provided by OSIL and also prepared ashore.The thiosulfate solution was standardized at a temperature within ±1 • C of the temperature at which the oxygen samples were analyzed.Duplicate thiosulfate standardizations were run requiring endpoints to be within (±0.3 %) of each other.

Underway measurements
Throughout the three cruises underway near-surface (∼ 3 m) quasi-continuous temperature and salinity data were collected using two Sea-Bird Electronics Seacat 21 thermosalinographs (SN 2310 in GEFPAT-1 and SN 3265 in GEFPAT-2 and GEFPAT-3).The data were recorded every 30 s and occasionally every 60 s along some tracks of GEFPAT-2.Bottle salinity samples were taken periodically from the thermosalinograph water intake to verify the thermosalinograph calibration.These water sample salinities were determined on board following the same procedures described above.Pump malfunctioning, continental near-shore discharges and intense phytoplankton blooms tend to clog the thermosalinograph filter and alter the flow rate.Flow rate disturbances can result in spurious temperature and conductivity fluctuations.At different stages of the Patagonia GEF cruises the thermosalinograph data presented indications of such malfunctioning, which required flow rate readjustment and filter replacement.The calibrated thermosalinograph and surface CTD data were overlaid to identify suspicious data and bad data were removed.To smooth the noise in the thermosalinograph data caused by flow rate disturbances, temperature and conductivity were filtered using a cosine filter with an 11-point window length.Filtering was carried out employing the Window Filter routine available in the Sea-Bird Data Processing software (SBE, 2005)

Sensors calibration
Thermosalinograph temperature and conductivity were compared with corrected CTD temperature and conductivity data extracted from the 3 dbar level during down-and upcasts for each station.Similar to the CTD calibration procedure, bias (offset) and slope corrections to the nominal calibration were determined from a linear least squares fit to the CTD versus thermosalinograph data of each variable.Values greater than 2 standard deviations from the fits were rejected.In addition, the corrected thermosalinograph salinities were compared with the salinity from bottle samples collected underway to provide independent verification of the calibration described above.Differences between the corrected thermosalinograph salinities and the bottle salinities for each cruise are shown in Fig. 4. Table 3 presents the thermosalinographbottle-salinity comparisons after temperature and conductivity sensors were calibrated.

Data access
The hydrographic data sets from GEF Patagonia cruises are reported in standard Sea-Bird Converted Data File (cnv) format.Converted files consist of a descriptive header followed by data converted to engineering units.The header contains station time and position information; the name of the raw input data file; the number of data rows and columns; a description of observed and derived variables in each column; interval between rows, scan rate or bin size; and records of all processing steps.The header information is followed by data records with a flag field in the last column, indicating whether the data record was interpolated in the last column.Data from individual stations are presented in separate ASCII character files consisting of 1 dbar data records in physical units.For each cruise, thermosalinograph data are reported in various ASCII files in the original sampling frequency with records in physical units.The thermosalinograph filenames are indicated with prefix "TSG".The data are available at the US National Oceanographic Data Center, NOAA; under doi:10.7289/V5RN35S0.

Figure 1 .
Figure 1.Location of hydrographic stations occupied during the GEF Patagonia cruises (symbols) and cruise tracks along which surface observations were collected (lines).Selected station numbers for each cruise are shown in the same colors.The background shading and contours indicate bottom topography in meters.

Table 1 .
Summary of CTD sensors used in GEF Patagonia cruises.

Table 2 .
Comparison of CTD vs. water sample salinity.Pre-and post-calibration (bold) for each cruise calculated for the whole water column (0 < p < p max ) and below 200 dbar (200 < p < p max ).Number of samples (N) is indicated for each set.

Table 3 .
Calibrated thermosalinograph versus water sample salinity comparison for GEF Patagonia cruises.Number of samples (N ) is indicated for each cruise.

.
M. Charo and A. R. Piola: Hydrographic data from the GEF Patagonia cruises