Interactive comment on “ The Irminger Sea and the Iceland Sea time series measurements of sea water carbon and nutrient chemistry 1983 – 2006 ”

The Irminger Sea and the Iceland Sea time series measurements of sea water carbon and nutrient chemistry 1983–2006 J. Olafsson, S. R. Olafsdottir, A. Benoit-Cattin, and T. Takahashi Marine Research Institute, Skulagata 4, IS 121 Reykjavik, Iceland Institute of Earth Sciences, University of Iceland, Sturlugata 7, IS 101 Reykjavik, Iceland Lamont-Doherty Earth Institute, Palisades, NY 10964, USA Received: 23 September 2009 – Accepted: 28 September 2009 – Published: 16 October 2009 Correspondence to: J. Olafsson (jon@hafro.is) Published by Copernicus Publications.

For a complete list of all parameters available in CARINA see Key et al. (2009).Note the different names for the parameters in the Exchange files (the individual cruise files) and the merged data product.

Introduction
In 1983 a study of the seasonal variability of carbon-nutrient chemistry was initiated off the Iceland shelf in two hydrographically different regions of the northern North Atlantic (Takahashi et al., 1985;Peng et al., 1987).One station was in the northern Irminger Sea (IRM-TS) with relatively warm and saline (S > 35) Modified North Atlantic Water derived from the North Atlantic Drift.This location may also be described as representing the sub-polar gyre (Hátún et al., 2005b).The other station was in the Iceland Sea (IS-TS) where cold Arctic Intermediate Water, formed from Atlantic Water and low salinity Polar Water usually predominates but the Polar Water influence in the surface layers is variable (Stefánsson, 1962;Hansen and Østerhus, 2000).Both stations are thus in regions important as sources for North Atlantic Deep Water (NADW).The original seasonal variability study was expanded in time and the sampling incorporated into the quarterly cruises of the repeat hydrography network of the Marine Research Institute (MRI) in Reykjavik, Iceland.Quarterly sampling is insufficient to adequately describe annual biochemical processes in these waters.It has, however, been estimated, with respect to hydrographic variability in sub-arctic waters of the N-Atlantic, that 4 observations/year are sufficient to record decadal variability (Hátún et al., 2005a).The Iceland Sea time series data has recently been evaluated to describe the rate of seawater acidification, at surface and deep levels (Olafsson et al., 2009).The time series observations have been carried out under the EC projects ESOP-2, TRACTOR and currently CARBOOCEAN and EPOCA.

Data provenance
The CARINA database includes data and metadata from 188 oceanographic cruises/campaigns, of which five entries consist of multiple cruises (Key et al., 2009).The IS and IRM time series contribute to the CARINA data.However, since these stations are in relatively shallow locations compared to the CARINA secondary quality control criteria, depth > 1500 m and considering also the high temporal variability, IRM-TS could not be included in the secondary QC.Data from the IS-TS is however included in the secondary QC for TCO 2 and nutrients as described in Olsen (2009) and Olafsson and Olsen (2010).Here we describe the methods and quality control procedures applied in gathering the two time series data sets.The IRM-TS is included in the CARINA-ATL region (Tanhua et al., 2010) and the IS-TS is included in the CARINA-AMS region (Olsen et al., 2009).
The repeat hydrography network of the Marine Research Institute (MRI) is carried out in quarterly cruises conducted generally in February, May, August and November each year.The time series stations are located at 64.33 • N, 28.0 • W (IRM-TS) where the depth is 1000 m and the IS-TS station is at 68.0 • N, 12.67 • W where the depth is 1850 m (Fig. 1).When it has not been possible to reach the time series location but observations have been available at a nearby station in the region, these have been incorporated into the time series.The nearby stations are generally within 70 km from the time series location except in 1983-1984 when the initial Iceland Sea sampling was 250 km to the west of the fixed location.Samples from all collection depths have been taken for salinity, dissolved oxygen and inorganic nutrients.From 1983 to 1991 only surface samples for pCO 2 and TCO 2 were collected.Water column sampling for TCO 2 started in 1991 and for pCO 2 in 1993.

Hydrography
From 1983 to the end of 1989 the station water sampling was conducted with TPN-Nansen water bottles, from HYDROBIOS GmbH, on a hydrowire.
They were fitted with reversing mercury thermometers.From the beginning of 1990 the station work has been conducted using SEA-BIRD Conductivity-Temperature-Depth (CTD) profiling instruments and water bottles on a rosette.Sample salinity measurements were carried out using Guildline Autosal Model 8400 salinometers.

Dissolved oxygen
Dissolved oxygen has been determined throughout the time series by an in-bottle microburette Winkler titration and visual end point detection (Carpenter, 1965).The sample bottles are volume calibrated Quickfit brand Erlenmeyer type of 50 ml nominal volume.

Inorganic nutrients
Samples for the determinations of the phosphate, nitrate (nitrate+nitrite) and silicate concentrations have been collected in 250 ml soft low density polyethylene bottles washed with dilute hydrochloric acid prior to each cruise.They are kept refrigerated if the analysis is carried onboard, as is common in the spring, but frozen for analysis ashore as is common in the other seasons.In spring and summer, samples from the surface layer, 0-60 m, are syringe filtered through a 0.45 µm Whatman PURADISC syringe filter to avoid turbidity blank effects, particularly on phosphate.Samples from deeper water are not filtered.Prior to 1987 a single channel Technicon AutoAnalyzer II was used for nitrate and silicate and a manual method for phosphate (Murphy and Riley, 1962).A Chemlab three channel autoanalyzer has been used since 1987, set up for determinations of dissolved phosphate, nitrate and silicate.The methods were those described by Grasshoff (Grasshoff, 1970) except for phosphate were a modified version of the Murphy and Riley method was automated (Murphy and Riley, 1962).A series of 5 working standards is prepared with each batch of samples and the response fitted to concentration with a 3rd order polynomial regression.

DATE
In addition, a laboratory reference material, LRM (Aminot and Kérouel, 1998), is regularly produced, assessed and compared to QUASIMEME materials.The LRM has nutrient concentrations within the range found for the regional seawater and samples of the LRM are run with each sample batch.Results from these measurements are indicative of the precision and accuracy of the daily procedures.The LRM results are generally within accepted limits, ±0.2 µmol/l for nitrate and silicate, ±0.03 µmol/l for phosphate.Corrections based on the LRM results need rarely to be applied but the results do occasionally give cause for re-running samples.

Partial pressure of carbon dioxide in seawater
From 1983 to spring 1993, samples for pCO 2 were collected onboard by recirculating 500 ml marine air in a closed system through a bubbler immersed in a 4 l surface seawater sample.The equilibrated gas was isolated and sealed in a 300 ml glass flask equipped with a stopcock at each end and shipped for analysis at Lamont-Doherty Earth Observatory (LDEO) by means of gas chromatography.Air samples   were also collected in 300 ml glass flasks by suction using a hand pump and were analysed the same way.The gas chromatograph was calibrated using air-CO 2 gas mixtures which had been analysed by C. D. Keeling of the Scripps Institution of Oceanography.The procedure changed after mid year 1993.Then 500 ml seawater samples for pCO 2 were brought back to MRI in screw caped Pyrex bottles, inoculated with saturated HgCl 2 solution (400 microliters) and kept in dark cold storage until analyzed.Their pCO 2 values were determined at a known temperature and pressure using the bubble-type equilibrator system coupled with a gas chromatograph (Chipman et al., 1993), which was calibrated with three air-CO 2 mixtures tied to the Keeling standards.The pCO 2 was measured in the stored sea water samples generally within 14 days after the samples were collected at sea.The pCO 2 data are not included in the merged CARINA data products, but are included in the individual cruise file available at http://cdiac.ornl.gov/oceans/CARINA/Carina inv.html.

Storage experiment of pCO 2 samples
The effects of sample storage were evaluated at MRI by collection of two sets of 12 paired samples during spring bloom, 2 June 1998 in the Iceland Sea at 68 • N, 18.8 • W. One set was from the surface (t = 1.83 • C) and the other from 200 m (t = 1.05 • C), well below the euphotic layer.Duplicate samples were analysed ashore over the period 3 to 23 June (Fig. 3).The average concentration for the storage experiment was 163.8 µatm at 5 m depth (s.d.0.6 µatm, n = 6) and 383.4 µatm at 200 m depth (s.d.1.7 µatm, n = 6).The change with time of pCO 2 in the surface samples and 200 m samples were −0.006 ± 0.05 µatm d −1 and −0.06 ± 0.08 µatm d −1 respectively, and was found to be very Earth Syst.Sci.Data, 2, 99-104, 2010 www.earth-syst-sci-data.net/2/99/2010/ small.This indicates that sample storage of up to three weeks has insignificant influence on pCO 2 and that the overall precision of the pCO 2 determinations is better than ± 2 µatm.

Dissolved inorganic carbon
Total dissolved inorganic carbon (TCO 2 ) has been determined in HgCl 2 -preserved samples by coulometry (Chipman et al., 1993) using at LDEO the Coulometrics CM-5010 instrument.HgCl 2 was found to be an effective preservation agent for TCO 2 in seawater samples in tightly sealed bottles stored up to 7 months (Takahashi et al., 1970).Prior to 1991 the samples were analysed at LDEO where the coulometer was calibrated using three different methods: a) with weighed quantities of clear Iceland spar (CaCO 3 ), b) with weighed quantities of heat treated Na 2 CO 3 ,and c) volumetrically determined CO 2 gas (Takahashi et al., 1985).These calibrations were consistent with Keeling's manometric values as demonstrated by Wong (1970) using Keeling's manometric system.The overall precision in the period 1983 to 1990 is estimated as ± 4 µmol kg −1 (Takahashi et al., 1985).As demonstrated extensively through the WOCE Program, the gas loop calibration method yielded TCO 2 values consistent with those in the Dickson reference waters, which were tied to the Keeling's manometric system (Rubin et al., 1998;Takahashi et al., 2006).From 1991 the coulometric determinations (using an improved Coulometrics Model CM-5011) were performed at MRI where the coulometer was calibrated using 99.998% CO 2 gas in fixed volume gas loops at known pressure and temperature (Chipman et al., 1993).
The Iceland Sea time series TCO 2 data were assesed as a part of the Nordic Seas CARINA data (Olsen, 2009).They were not included in the crossover and inversion analyses but it was concluded on the basis of nitrate-TCO 2 relations that the TCO 2 data "appeares reasonable" (Olsen, 2009).

TCO 2 quality control
Accuracy of the TCO 2 determinations at MRI is maintained since 1992 by comparison of results with sea water Certified Reference Material (CRM) calibrated and supplied by Dr. Andrew G. Dickson of the Scripps Institution of Oceanography.The evaluation of the CRM results indicated a systematic error of about −4.7 ± 2.0 µmol C kg −1 in 212 CRM determinations.In 1999 the coulometer system gas loop volumes were redetermined by gravimetry and the systematic error ascribed to errors in the gas loop volumes.A correction multiplier of 1.0029 was hence applied to all results from analyses in the 1991-1998 period.The CRM results representing this period are shown in Fig. 4. Due to instrument problems there were no TCO 2 analyses of samples or CRMs in 1999 and 2000.Samples collected in this period were analysed in 2001 and 2002.The samples were all spiked with HgCl 2 , stored refrigerated at TCO2 differance (µmol/kg) 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 ) 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008  The precision of the TCO 2 determinations at MRI is estimated from the standard deviation of the analysis of individual CRM batches.
It ranges from 1.4 to 2.3 µmol kg −1 for the period 1993 to 2008 based on 382 CRM determinations in 13 batches.The average is 1.85 µmol kg −1 which rounds to ± 2 µmol kg −1 .Since all the time series TCO 2 data are tied to CRMs, as described above, we estimate the overall accuracy as ± 2 µmol kg −1 , the uncertainty in the CRM determinations.

Summary
The ways and means of assembling and quality controling the Irminger Sea and Iceland Sea time-series biogeochemical data are described.These stations are in relatively shallow, but oceanographically important locations.Compared to the CARINA criteria, depth > 1500 m, IRM-TS could not be included in secondary QC and the IS-TS only in a limited way.However, with the information provided here, the quality of the data can be assessed e.g. on the basis of the results obtained with the use of reference materials.

Data access
The whole CARINA database set is published at http://cdiac.ornl.gov/oceans/CARINA/Carinainv.html.It contains 188 individual cruise files in comma-separated, WHPO exchange www.earth-syst-sci-data.net/2/99/2010/ Earth Syst.Sci.Data, 2, 99-104, 2010 format.Condensed metadata are contained in the header of each data file.In addition, the CARINA database contains three merged, comma-separated data files with the data products.These files are divided into the three geographical regions of CARINA.No special software is needed to access the data, but software for MATLAB users is offered to facilitate reading of the data.

Figure 1 .
Figure 1.Locations of the Irminger Sea and the Iceland Sea time series stations inserted on the N-Atlantic surface current chart of Hansen and Østerhus (2000).

Figure 2 .
Figure 2. Average MRI-IS differences of (a) nitrate, (b) phosphate and (c) silicate concentrations (in µmol/l) from assigned values in QUASIMEME sea water test materials 1993-2008 (silicate from 1996).The overall long term mean deviations are shown with black slashed lines and their standard deviations are shown with a red dotted line.

Figure 3 .
Figure 3. Results of pCO 2 sample storage experiment.Average pCO 2 (µatm at 4 • C) in (a) samples from 5 m depth, and (b) samples from 200 m depth.The average concentrations for the storage experiment are shown with a black slashed line and their standard deviations are shown with red dotted lines.

Figure 4 .
Figure 4. Results of MRI TCO 2 determinations in Dickson's sea water reference materials (CRMs) in two periods 1992-1998 and 2001-2008.The average differences (MRI-CRM) are shown as black slashed lines and their standard deviations are shown with red dotted lines.