Methodology:

All data were collected from the RRS John Biscoe (cruises JB3, JB4, JB5, JB6, JB8, JB10) between November 1981 and February 1990, or the RRS James Clark Ross (cruises JR2, JR6, JR11, JR17, JR25, JR28, JR38, JR57, JR70, JR82, JR116, JR161, JR177, JR200) between November 1992 and April 2009. Chemical analysis was undertaken aboard ship using a Segmented-Flow Analyser and data analysis was performed post-cruise. All methods (including statistical analysis) are detailed in the following papers:

Korb RE, Whitehouse M (2004) Contrasting primary production regimes around South Georgia, Southern Ocean: large blooms versus high nutrient, low chlorophyll waters. Deep-Sea Res I 51: 721-738. https://doi.org/10.1016/j.dsr.2004.02.006.

Korb RE, Whitehouse MJ, Ward P, Gordon M, Venables HJ, Poulton AJ (2012) Regional and seasonal differences in microplankton biomass, productivity. And structure across the Scotia Sea: implications for the export of Biogenic carbon. Deep-Sea Res II 59-60: 67-77. https://doi.org/10.1016/j.dsr2.2011.06.006.

Morris AW, Howland RJM, Bale AJ (1978) A filtration unit for use with continuous autoanalytical systems applied to highly turbid waters. Estuarine and Coastal Marine Science, 6: 105-109.

Whitehouse MJ (1997) Automated seawater nutrient chemistry. British Antarctic Survey, Cambridge, 14 pp.

Whitehouse MJ, M Preston (1997) A flexible computer-based technique for the analysis of data from a sea-going nutrient autoanalyser. Analytica Chimica Acta, 345: 197-202. https://doi.org/10.1016/S0003-2670(96 )00630-7.

Whitehouse MJ, VR Woodley (1987) Automated Seawater Nutrient Analysis. British Antarctic Survey, Cambridge, 41 pp.

Woodley VR (1989) A User-Friendly Software Interface for a Program Using a Digitizer in Automated Seawater Nutrient Analysis. University of Cambridge Diploma in Computer Science.

Coincident temperature and salinity data were collected either through the ship's underway Oceanlogger system or from vertical profiles using a Conductivity-Temperature-Depth (CTD) sensor.

Samples for vertical profiles were collected from water bottles attached to a CTD rosette while samples for horizontal transects were collected from the ship's non-toxic seawater supply with an intake depth about 6 - 7 m. Chemical analyses were undertaken using a custom-built Segmented-Flow Analyser initially based on Chemlab colorimeters and Ismatec proportioning pumps until 1993 when Technicon colorimeters were adopted. Initially, data were logged to paper chart and processed manually but in the mid-1980s, data extraction was automated and processing was managed with a digitising tablet and associated PC. In 1993 as part of a comprehensive re-build, data were logged to a PC. Data acquisition and subsequent processing was achieved with custom-built software (BAS in-house and Cambridge Beacon Ltd) and this was updated further in the mid-1990s when National Instrument data acquisition programmes were used to collect data, and calibration along with data organisation were switched to commercially available spreadsheets. Data collected during transecting were logged along with a time-stamp acquired from the central ship's clock to allow synchronisation with concurrent measurements such as temperature and salinity from the ship's Oceanlogger system.

A conductivity-temperature-depth (CTD) sensor was used to determine vertical profiles of temperature and salinity at each station. Further details of the data processing and calibration of the salinity data, where undertaken, are available in the cruise reports. Where available, temperature and salinity from the respective water bottle firings on the upcast of the profile have been matched to the nutrient data. Where these data were not available, measurements from the depth of the bottle firing (± 2 m) on the downcast of the CTD were used instead.

Cruise reports are ava...(52)

Data quality:

Archived underway nutrient data were compiled with ship, cruise number, geographic location, and, for cruises JR161, JR177 and JR200, temperature and salinity. Each timestamp was reformatted as a date vector, a serial date number, and a datetime string (DD-MMM-YYYY hh:mm:ss). The serial date numbers were ordered temporally, converted into a datetime string (DD-MMM-YYYYThh:mm:ssZ), and converted to a table with minute intervals and corresponding arithmetic means of concentrations of each nutrient, temperature and salinity. Profile nutrient data were compiled with ship, cruise number, event number, geographic location and depth. Temperature and salinity data were extracted and matched with profile nutrient data from each cruise using latitude and depth as primary selector variables, with event numbers and timestamps used as secondary checks. Timestamps associated with the start and end of the CTD casts were saved in the format DD-MMM-YYYYThh:mm:ssZ.