Holocene Mg/Ca,alkenones,and light stable isotope measurements on the outer North Iceland shelf (MD99-2269): A comparison with other multi-proxy data and sub-division of the Holocene

Abstract

To evaluate whether proxies that record surface, near-surface, and bottom water conditions from the North Iceland shelf have similar trends and periodicities, we examine Holocene century-scale paleoceanographic records from core MD99-2269. This core site lies close to the boundary between Atlantic and Arctic/Polar waters, and in an area frequently influenced by drift ice. The proxies are stable δ¹³C and δ¹⁸O values on planktonic and benthic foraminifera, alkenone-based sea-surface temperatures (SST°C), and foraminiferal Mg/Ca SST°C and bottom water temperature (BWT°C) estimates. These data were converted to equi-spaced 60-year time-series; significant trends were extracted using Singular Spectrum Analysis, which accounted for between 50% and 70% of the variance. In order to evaluate within-site ocean climate variability, a comparison between these data and previously published proxies from MD99-2269 was carried out on a standardized data set of 14 proxies covering the interval 400–9200 cal. yr BP. Principal component (PC) analysis indicated that the first two PC axes accounted for 57% of the variability with high loadings primarily defining ‘nutrient’ and ‘temperature’ proxies. Fuzzy k-mean clustering of the 14 climate proxies indicated major environmental changes at ~6350 and ~3450 cal. yr BP, which define local early-, middle-, and late-Holocene climatic shifts. Our results indicate that the major control on the combined proxy signal is the Holocene decrease in June insolation, but regional changes in such factors as sea-ice extent and salinity are required to explain the threefold division of the Holocene.

Keywords

alkenone SST°C Mg/Ca SST°C and BWT°Cs North Iceland shelf stable O and C isotope records on foraminifera statistical analysis

Introduction

The steep oceanographic gradients across the North Iceland shelf (Belkin et al., 2009; Stefansson, 1962, 1969) (Figure 1) have made it a target for the study of North Atlantic ocean climate change since the last deglaciation (Andersen et al., 2004; Moros et al., 2006; Sicre et al., 2008). We define ‘ocean climate’ as the average sea-surface and seafloor temperature, salinity, and geochemistry through a seasonal cycle, and the extent and duration of drift ice. The area of interest (Figures 1 and 2) lies along the North Iceland Front (Belkin et al., 2009), the west–east marine frontal system, which demarcates the boundary between sea-ice-bearing Arctic and Polar waters to the north, and the Atlantic waters carried with the warm North Iceland Irminger Current (NIIC) on the inner and mid-shelf (Figure 1). Changes in the ocean climate around Iceland influence the climate of Iceland (Olafsson, 1999) and are representative of changes in the wider North Atlantic (Dickson et al., 1988). On the North Iceland margin, studies of cores B997-321, B997-330, MD99-2263, MD99-2264, MD99-2266, MD99-2269, MD99-2275, and JR51GC35 (Figure 2) (e.g. Bendle and Rosell-Mele, 2007; Eiriksson et al., 2004; Knudsen et al., 2004; Moossen et al., 2015; Olafsdottir et al., 2010; Sicre et al., 2008) have also increased our knowledge of changes in ocean climate during the Holocene. Published records from MD99-2269 include information on coccoliths (Giraudeau et al., 2004, 2010), diatoms (Andersen et al., 2004), dinocysts (Solignac et al., 2006), foraminifera (Giraudeau et al., 2004; Justwan et al., 2008; Kristjánsdóttir et al., 2007a), quartz and calcite percentage data (Moros et al., 2006; Stoner et al., 2007) (Table 1, Figure 3), and sediments (Andrews et al., 2003). Thus, a large number of proxies have been utilized to develop time-series of oceanographic change at the sea surface/near-surface as well as at the seafloor for this critical core site (Table 1). Several of these records are available from the NOAA Paleoclimate Data base (http://www.ncdc.noaa.gov/data-access/paleoclimatology-data/datasets). Furthermore, all our new data have been added to the Pangaea Data Base (Kristjánsdóttir et al., in preparation: doi:10.1594/PANGAEA.847657, see Appendix).

Figure 1.

(a) February temperature (2008) at 50 m in Iceland waters (01 February 2008–14 February 2008). The red dashed lines indicate the hydrographic sections surveyed by the Iceland Marine Institute; also shown are 2014 February temperatures along the (b) Hornbanki and (c) Siglunes hydrographic sections (http://www.hafro.is/Sjora/) showing NIIC waters overlying cold Arctic/Polar water in the deep troughs. Filled square in A and B represents the site of MD99-2269 (Figure 2).

Figure 2.

Plot of the location of MD99-2269 and other cores mentioned in this paper. The average April location of the sea-ice margin AD 1870–1920 is shown as a heavy dashed black line (http://nsidc.org/data/gis/data.html). Location of hydrographic sections Siglunes (S) and Hornbanki (H) are shown with a dashed pink line (see Figure 1).

Table 1.

The 14 proxies and references for core MD99-2269. Arranged from bottom waters to surface waters.

Figures 3 and 8	Proxy	Number*	Ave yr/sample	Responds to	References
1	Mg/Ca benthic forams	180	57	T°C	Kristjánsdóttir (2005, this paper)
2	δ¹⁸O benthic forams	180	57	T°C and S‰	Kristjánsdóttir (2005, this paper)
3	δ¹⁸O planktonic forams	220	54	T°C and S‰	Kristjánsdóttir (2005, this paper)
4	Mg/Ca planktic forams	148	63	T°C	Kristjánsdóttir (2005, this paper)
5	δ¹³C planktic forams	220	54	T°C and S‰	Kristjánsdóttir (2005, this paper)
6	Quartz wt%	414	28	Sea ice, T°C	Moros et al. (2006)
7	Dinoflagellates	209	37	Sea ice, T°C	Solignac et al. (2006)
8	Calcite wt%	409	28	Nutrients, T°C	Moros et al. (2006)
9	Coccoliths	282	37	Nutrients, T°C	Giraudeau et al. (2004, 2010)
10	Dinoflagellates	209	37	S‰	Solignac et al. (2006)
11	C37	182	62	S‰	Moros (this paper)
12	UK37	182	62	T°C	Moros (this paper)
13	Dinoflagellates	209	37	August SST°C	Solignac et al. (2006)
14	Diatoms	131	72	August SST°C	Justwan et al. (2008)

SST°C: sea-surface temperature.

Number of samples.

Figure 3.

Suite of Holocene published proxy records and our new data for MD99-2269 (Table 1). The boundaries shown for the early-, mid-, and late-Holocene are defined later in this paper (see Figures 8 and 9).

Objectives and organization

Regional comparisons tend to ignore the question of the variability between proxies that record processes active at different seasons and locations with the water column. Our paper is only concerned with the Holocene records from MD99-2269 (Table 1, Figures 1 and 2), and our goal is to investigate whether there are indeed any common trends and periodicities in proxies that respond to changes in surface, near-surface, and seafloor ocean climate. As Li et al. (2010) noted, the analysis of multi-proxy data from a single core has the advantage that temporal synchronization of the records is assured (e.g. Blaauw, 2012; Torbenson et al., 2015).

We first present our Holocene geochemical proxy data from surface, near-surface, and benthic organisms (Figure 3); we then combine this information with existing published proxies (Table 1, Figure 3) to answer the following questions: (1) are there significant trends and periodicities in the new proxy records, (2) are these trends and periodicities similar to those from previously published proxies for this one core site (e.g. Figure 3), and (3) do the combined data show any consistent patterns in terms of identifying major regime shifts in Holocene?

Core site and background

Core MD99-2269 was retrieved in Húnaflóadjúp in 365 m of water (66°37.53′N and −20°51.15′W; Figure 1), and recovered 25.4 m of sediment (Labeyrie et al., 2003). The site of this giant Calypso core MD99-2269 (Figure 2) (Labeyrie et al., 2003) was chosen based on the B997-seismic and coring program (Helgadottir, 1997). Sub-bottom profiles indicated that this trough contained a thick, drift-like sediment unit (Andrews, 2007; Helgadottir, 1997), with several distinct parallel reflectors that record major tephra events, including the Saksunarvatn tephra at ~22-m core depth (Andrews et al., 2002; Kristjánsdóttir et al., 2007b).

Oceanography of the area

MD99-2269 lies between two of the hydrographic sections that the Iceland Marine Institute occupies (http://www.hafro.is/Sjora/) – the Hornbanki (H) section to the west and the Siglunes (S) section to the east (Figure 2). Figure 1a illustrates the potential February temperature at 50-m depth, and Figure 1b and c shows the temperature profiles for the two sections in February 2014. The NIIC is a surface to intermediate depth Atlantic Water current, which overlies cold upper Arctic Intermediate Water (UAIW) (potential temperature < 0°C) in the troughs (Figure 1b and c). This cold water mass (<0°C) is present in all of the deep North Iceland troughs and the hydrographic sections indicate that it is generally present at water depths of ⩾350–400 m. MD99-2269 lies close to the boundary between UAIW and the overlying cooled Atlantic Water (Figure 1b and c). In winter (Feb) there is little vertical variation in potential density, but the water column becomes more stratified during the summer months.

The sensitivity of the ocean climate of North Iceland is readily seen in the composite temperature and salinity records from the section (Figure 2) (Olafsson, 1999), which illustrates the impact of salinity excursions (i.e. the Great Salinity Anomaly of 1969 (Dickson et al., 1988; Malmberg, 1969, 1985) and the 1980s salinity anomaly (Belkin et al., 1998)). This sensitivity is further illustrated by the location of the MD99-2269, close to, but south of the average, April 15% sea-ice concentration AD 1870–1920, which is used as a measure of the position of the sea-ice edge (Divine and Dick, 2006) (http://nsidc.org/data/gis/data.html) (Figure 2). Icebergs and sea ice extend southward of this average sea-ice edge and in extreme years strand on the coast. Variations in the extent of drift ice from the time of the Settlement AD 870 have been reconstructed from historical data (Koch, 1945; Ogilvie, 1996; Ogilvie et al., 2000; Wallevik and Sigurjonsson, 1998). Periodic transport of sediment-bearing drift ice to the north Iceland shelf is shown by the presence of allochtonous quartz (Moros et al., 2006) (Figure 3), the sea-ice biomarker IP25 (in MD99-2275, Figure 2) (Masse et al., 2008), and FeO grains sourced to sites within the Arctic Basin in MD99-2263 (same site as MD99-2264, Figure 2) (Darby et al., 2015).

Chronology

The depth/age model for MD99-2269 (Stoner et al., 2007; Dates in Dunhill et al., 2004; Smith and Licht, 2000) is based on co-mingling of calibrated ¹⁴C dates from MD99-2269 (27 dates) and an East Greenland Shelf core (MD99-2322) (20 dates) based on depth-correlated paleomagnetic secular variations (PSVs). All but one of the radiocarbon dates for the age model were on mollusks or benthic foraminifera. The depth/age model (Stoner et al., 2007) has a nearly monotonic sediment accumulation rate (SAR) averaging 21.7 cm/100 yr (or 4.6 ± 2 yr/cm but with variations between 1.3 and 9 yr/cm). In comparison, the best-fit linear depth/age model has r² = 0.99 with an average departure from the Stoner et al. (2007) data of only ±45 year. Our compilation of proxy records published prior to 2007 (Figure 3) keeps their original age models because of the near constant rate of sediment accumulation noted above. However, we use the Stoner et al. (2007) depth/age model in all of our subsequent analyses and figures (Figures 4 –9). We restrict our data analysis here to the last 11,000 cal. yr BP, although the full data to 11.7 cal. ka BP is provided in the online data archive (Kristjánsdóttir et al., 2015; see Appendix).

Figure 4.

(a) All δ¹³C data on the planktonic foram Neogloboquadrina pachyderma (sin) (Np s) in core MD99-2269 (line with filled squares) and the trend (solid line); (b) original δ¹⁸O Neogloboquadrina pachyderma sin (Np s) data (filled squares), 60-year integrated estimates after sea-level correction solid line and filled circles, and SSA trend (black line) (Kristjánsdóttir, 2005); (c) original (i.e. neither sea level nor vital effect corrected) δ¹⁸O for Islandiella norcrossi/helenae and the trend (see text) solid line on the data.

Figure 5.

(a) Alkenone derived temperature (blue line with infilled squares), trend (black line), and %C₃₇ (red solid line); (b) plot of the bottom water temperatures (red line) and trend (black line) based on the Mg/Ca ratios on Islandiella norcrossi/helenae; (c) Mg/Ca ratios for Islandiella norcrossi/helenae (blue) and Neogloboquadrina pachyderma (sin) (red) and the linear trends, slopes, and values for the correlation coefficient r.

Figure 6.

Plot of Mg/Ca and δ¹⁸O-based (sea-level corrected) temperature estimates based on results from three benthic foraminifera (Kristjánsdóttir, 2005): (a) Islandiella norcrossi/helenae (Is n/h), (b) Melonis barleeanu (Mb), and (c) Cassidulina neoteretis (Cn).

Figure 7.

Box plot of 11 proxy temperature estimates for MD99-2269. The box encloses 50% of the data. The small open circles are the outliers (Velleman and Hoaglin, 1981) that were replaced in the standardized data set (see text). Outliers are points whose values are either greater than UQ + 1.5 * IQD or less than LQ – 1.5 * IQD, where UQ and LQ are the upper and lower quartiles and IQD is the Interquartile Distance.

Figure 8.

(a) Plot of the normalized scores on 14 proxy data sets between 500 to 9200 cal. yr BP. Sign on the scores adjusted to reflect a cold (blue) to warm (red) gradient. Proxies are arranged from bottom (1) to top (14) with respect to bottom waters to surface water; (b) bar plot of the loadings on the first and second PCAs. Solid red bar = PC 1 scores; gray = PC 2 scores (see also Supplementary Figure 1, available online). The shaded rectangle includes ±0.2 of the PC loading value; (c) PCA scores on the first and second PCAs – the shaded gray area represents the mid-Holocene zone as defined by fuzzy k-mean clustering (see text and Figure 9d).

Figure 9.

(a) Plot of the scores on the first PC and the insolation receipts at 60°N (Berger and Loutre, 1991) – the mid-Holocene interval (based on Figure 9d) is shaded gray; (b) polynomial trends and the explained variances for the Uk₃₇ and coccolith data (Giraudeau et al., 2004, 2010); (c) interpretation of the data on 9B and other figures (see Figures 3 and 8a); (d) fraction of each sample (y axis) assigned to the three-part cluster membership.

Methods

We next outline the laboratory procedures used to obtain the new proxy data (Figures 3 –6).

Stable isotopes (new data)

Stable isotope measurements of benthic and planktonic foraminifera were obtained from core MD99-2269 (Kristjánsdóttir, 2005). Isotopic analyses were performed at the Leibniz-Laboratory for radiometric dating and stable isotope research at Christian-Albrechts-University Kiel, Germany, using a Finnigan MAT 251 mass spectrometer with an online coupled Kiel Carbon device. Kristjánsdóttir (2005) obtained δ¹⁸Ο data for several benthic foraminiferal species (Cibicides lobatulus (Cl), Cassidulina neoteretis (Cn), Islandiella norcrossi/helenae (Is n/h), Melonis barleeanus (Mb)), but here we concentrate on the data from Islandiella norcrossi/helenae, which has the most complete series. Cibicides lobatulus is an epifaunal species, but the other three are infaunal (Jennings et al., 2004; Kristjánsdóttir et al., 2007a; Rytter et al., 2002). We note that the records from the other infaunal species parallel each other, with a small offset probably associated with their specific vital effect (Kristjánsdóttir et al., 2007a; Smith et al., 2005). δ¹⁸O data ratios for the near-surface planktonic foraminiferal species Neogloboquadrina pachyderma (sin) (Np s), and Turborotalita quinqueloba (Tq) were also obtained. The δ¹⁸O data were corrected for the change in sea level associated with the melting of the late glacial ice sheets (Fairbanks, 1989).

Castaneda et al. (2004) and Smith et al. (2005) presented a regional analysis of Neogloboquadrina pachyderma (sin) δ¹⁸O data; Smith et al. (2005: 1730) argued that the δ¹⁸O records were primarily a function of temperature. A comparison of temperatures at 50 m and the δ¹⁸O of Neogloboquadrina pachyderma (sin) (equation #6, Smith et al., 2005: 1730) indicated that salinity had little impact on the δ¹⁸O values, but this interpretation was not borne out for the early-Holocene (Jennings et al., 2015; Quillmann et al., 2012), and is not supported by our data.

Alkenone sea-surface temperatures (new data)

Alkenone measurements followed the procedure described in Moros et al. (2004) and Risebrobakken et al. (2010). We calculated the sea-surface temperatures (SST°C) from $U_{37}^{k^{'}}$ according to Müller et al. (1998): $U_{37}^{k^{'}}$ = 0.033T + 0.044. The analytical precision based on multiple extractions of sediment samples was better than 0.6°C for the $U_{37}^{k^{'}}$ index. The alkenone %C_37:4 was calculated according to Bendle et al. (2005): %C_37:4 = (C_37:4/C_37:2 + C_37:3 + C_37:4) * 100. The value %C_37:4 of mostly less than 4% and summer SST°C of >8°C allow SST°C calculation using $U_{37}^{k^{'}}$ in the study area (Bendle and Rosell-Melé, 2004). Alkenones were analyzed starting at 70-cm depth in the core or an age of ~400 cal. yr BP, with a temporal resolution of about ~60 years per sample (Table 1).

Mg/Ca ratios (new data)

Sample preparation, species calibrations, and Mg/Ca paleotemperature estimates for the last 4000 years for Neogloboquadrina pachyderma (sin) and three benthic foraminiferal species, Cassidulina neoteretis, Melonis barleeanus, and Islandiella norcrossi/helenae, are provided in Kristjánsdóttir et al. (2007a). The same calibrations are applied to the longer time-series that we present in this paper (Kristjánsdóttir, 2005; Kristjánsdóttir et al., 2015). We focus on the data from Islandiella norcrossi/helenae in order to compare sea-level corrected δ¹⁸O and Mg/Ca data.

Data analysis

In preparation for statistical analyses, the new proxy data sets (Table 1, Figures 4 –6) were converted into equi-spaced 60-year time steps (the highest common resolution) using the integration option in AnalySeries (Paillard et al., 1996). Outliers, as defined by Exploratory Data Analysis (EDA) (Velleman and Hoaglin, 1981), were eliminated (e.g. Figure 7) and their values were adjusted to be the average between adjacent intervals. This procedure results in some loss of information, but the calculated correlation values between the original and converted time-series are >0.8 and allow power spectral analyses to be undertaken. Time-series data potentially consist of a long-term trend, one or more statistically significant periodicities, and noise (Weedon, 2005). We used the commercial version of the UCLA toolkit (kspectra) (Ghil et al., 2002) to evaluate whether significant trends and periodicities exist in our proxies (e.g. Figure 4). We define ‘trend’ as a long-term, low frequency (i.e. multi-millennial) tendency, which is usually removed prior to the investigation of shorter-term variations. We used Singular Spectrum Analysis (SSA) (Kondrashov et al., 2005; Weedon, 2005) to evaluate whether there is a statistically significant trend in the data, and the Multi-Taper Method (MTM) (Mann and Lees, 1996) to ascertain whether residuals from the trend contained any significant century to millenial-scale periodicities.

In the ‘Discussion’ section, we compile a total of 14 new and published proxy time-series between 400 and 9200 cal. yr BP and more intermittently from 100 to 11,000 cal. yr BP (Table 1, Figure 3). We used AnalySeries (Paillard et al., 1996) to convert the data matrix to time-series with samples spaced every 100 years from 400 to 9200 cal. yr BP. Furthermore, we changed the sign on some proxies (i.e. δ¹⁸O, quartz wt%, and sea-ice duration) to reflect temperature. Mg/Ca ratios are used directly rather than converted to estimated temperatures. To allow for statistical analysis, the data were standardized with mean values of 0 ± 1. Principal component analysis (PCA) was employed to compare 14 proxies (Table 1) (e.g. Axford et al., 2011) and to determine the strength of association between them. Furthermore, to test whether there were coherent temporal changes in the proxies between 400 and 9200 cal. yr BP we subjected the normalized 14 variables and 100-year estimates (Kristjánsdóttir et al., 2015) to fuzzy k-mean cluster analysis (program ‘FuzMe’; Ding and He, 2004; Ghosh and Dubey, 2013; Granath, 1984; Minasny and McBratney, 2002). This approach allows individual samples to have a degree of membership in more than one cluster, the extent of which is measured by the ‘Confusion Index’ (0 = no confusion; 1 = equal fractions of all clusters), a measure of cluster mixing (Minasny and McBratney, 2002).

Results on new proxy data

In order to assess the broad-scale evolution of Holocene ocean climate on the North Iceland shelf, we focus on whether there are significant trends in our proxies and search for significant periodicities in the residuals from the trends (Table 2).

Table 2.

Significant Multi-Taper Method (MTM) periodicities on detrended proxy data.

Periodicity year	100	200	300	400	500	800	1000
Nps ¹³C	188*	212	312	435*	555
Nps ¹⁸O	187	227,* 287		446*	–
¹⁸O benthic				434		820
UK37		208			529*
Is Mg/Ca		256					1000*

CI: confidence interval.

99% CI.

Stable isotopes

The Neogloboquadrina pachyderma (sin) δ¹³C data show a broad, threefold pattern with depleted δ¹³C values from 11,000 to ~8500 cal. yr BP, enriched values peaking ~5000 cal. yr BP, and then values falling below 0‰ over the last 2000 cal. yr BP. A total of 69% of the variance is explained by the first two SSAs (Figure 4a), which define the trend. Variations around the trend are of higher amplitude over the last 3000 years.

The Neogloboquadrina pachyderma (sin) δ¹⁸O series had a significant trend defined by the first SSA, which accounted for 60% of the variance (Figure 4b) but the residuals from the trend showed no evidence for multi-century significant periodicities (Table 2), hence can be considered ‘noise’. Of particular note is the abrupt decrease in δ¹⁸O values from ~3.5‰ to ~2.5‰ between 11,000 and 10,000 cal. yr BP (Figure 4b). The Islandiella norcrossi/helenae data, uncorrected for sea-level rise in the Holocene, show a prolonged decrease in δ¹⁸Ο between 10,000 to 6000 cal. ka BP. After 4000 cal. yr BP, these values vary around ~4‰. This change lags the rise in the planktonic δ¹⁸Ο values (Figure 4b) and implies water column stratification during the early-Holocene. The benthic series was detrended by removal of the first SSA, which accounted for 70% of the variance and is largely associated with the global rise in sea level (Figure 4c). Residuals from the trend had significant periodicities of 430 and 820 years (Table 2).

Alkenones

The data from MD99-2269 have average estimated $U_{37}^{k^{'}}$ based on SST°C of 10.1 ± 0.5°C with a range between 9.2 and 11.7°C (Kristjánsdóttir et al., 2015; Table 2). Present day summer temperatures near MD99-2269 range between 6 and 10°C (http://www.hafro.is/Sjora/). The 60 yr/sample data show several sharp oscillations in the estimated SST°C superimposed on a long-term reduction in SST°C. SST°C maxima occur ~1000 and 8000 cal. yr BP and there is a pronounced low between ~4200 and 200 cal. yr BP. A total of 79% of the variance in the alkenone SST°C estimates are associated with the SSA trend (Figure 5a). The variations from the trend (Figure 5a) show a series of oscillations with amplitudes of ±0.5°C. MTM analysis of the residuals indicates the presence of a highly significant ~530-year period plus a higher frequency 208-year cycle (Table 2). The overall trend of the C_37:4 data (Figure 5a) follows the $U_{37}^{k^{'}}$ SST°C and has a marked peak ~8200 cal. yr BP matching a sharp, short-lived SST°C peak. This is approximately coeval with the so-called 8200 cal. yr BP cold event (Alley and Agustsdottir, 2005), which has been associated with a cooling and freshening of surface and bottom Irminger Current temperatures from SW to NW Iceland (Jennings et al., 2015; Quillmann et al., 2012).

Studies from northern latitudes highlight an empirical relationship between the amount of C_37:4 and surface water salinity (Harada et al., 2003; Rosell-Melé, 1998; Rosell-Melé et al., 2002; Sicre et al., 2002). However, the application of C_37:4% in the reconstruction of past salinities is still controversial (e.g. Sikes and Sicre, 2002). According to Sikes and Sicre (2002), the observed correlation between C_37:4% and salinity in the North Atlantic is an artifact of the strong link between salinity and temperature in these areas. Other studies, however, argue that the relative proportion of C_37:4 is a valid indicator for documenting qualitative changes in the influence of freshwater in surface waters in the North Atlantic and Pacific oceans (e.g. Bendle et al., 2005; McClymont et al., 2008). In addition, a coherence of changing proportions of C_37:4% in surface sediment and water samples related to salinity gradients has been observed in the Skagerrak area (Blanz et al., 2005; Schulz et al., 2000) and C_37:4% has been applied to study Baltic Sea freshwater outflow changes (Rohde Krossa et al., 2014).

Mg/Ca ratios

The temperature reconstructions in Húnaflóadjúp over the last 4000 cal. yr BP (Kristjánsdóttir et al., 2007a) show an incursion of Arctic Waters ca. 500 cal. yr BP; this is also seen in the re-appearance of the Arctic benthic foraminifera Elphidium excavatum forma clavata (Giraudeau et al., 2004).

The modern calibration data for Islandiella norcrossi/helenae (Kristjánsdóttir et al., 2007a) indicate that the linear transfer function for the temperature for the yⁱ sample is as follows: T°C = (Mg/Ca – 1.04)/0.075 with r² = 0.93 and a Root MSE of 0.56. Although there are theoretical reasons for a non-linear relationship (Kristjánsdóttir et al., 2007a), in this particular case there is no statistical advantage in the use of other expressions. The results indicate an average infaunal calcification temperature of 4.1 ± 1.5°C (4.1 ± 1.4°C for the exponential equation) with a maximum estimate of 9.1°C and a minimum of ~1.0°C. The results show a trend for bottom water temperatures (BWT°Cs) to decrease throughout the length of the record (~360–1,300 cal. yr BP). The trend in the data explained 49% of the variance, and it is marked by a broad temperature maximum at about 10,000 cal. yr BP and an abrupt decrease in BWT°C ~4000 cal. yr BP (Figure 5b). There is a significant 1000-year periodicity in the detrended time-series (Table 2).

Mg/Ca ratios were obtained on the near-surface planktic foraminifera Neogloboquadrina pachyderma (sin) (Kristjánsdóttir, 2005; Kristjánsdóttir et al., 2015) (Figure 5c). Questions exist on the temperature calibration of this species (Meland et al., 2006), but application of the Elderfield and Ganssen (2000) equation resulted in an average calcification temperature estimate of 5.5 ± 0.8°C. We use the Mg/Ca Neogloboquadrina pachyderma (sin) ratios to derive an overall trend and to compare that record with the Mg/Ca ratios from Islandiella norcrossi/helenae. An important question is whether the near-surface and infaunal BWT°Cs follow similar trends throughout the period of our records. Figure 5c shows that both near-surface and infaunal species show a systematic and highly significant (prob. F > 0.0000) decrease in Mg/Ca ratios throughout the last 11,000 years requiring an overall decrease in temperature of the entire water column.

For the detrended data (Figures 4 –6), the most frequent periodicities of residuals are in the 400- to 500-year interval (Table 2), but it is unclear whether these have any temporal coherence or association with solar forcing (Clemens, 2005; Turney et al., 2005).

Estimate of changes in temperature and salinity

We compare temperature estimates based on Mg/Ca ratios and on δ¹⁸O from foraminifera (Figure 6). Temperature estimates (not shown) were calculated from Neogloboquadrina pachyderma (sin) δ¹⁸O ratios using the Shackleton (1974) equation (Kristjánsdóttir, 2005; Kristjánsdóttir et al., 2007a). For the benthic foraminifera, the δ¹⁸O values require correction for changes in sea level and vital effects. Vital effect/disequilibrium corrections are Islandiella norcrossi/helenae δ¹⁸O_calcite VPDB –0.298 (n = 9), Melonis barleeanus δ¹⁸O_calcite VPDB 0.276 (n = 18), and Cassidulina neoteretis δ¹⁸O_calcite VPDB – 0.047 (n = 18) (Kristjánsdóttir, 2005; Kristjánsdóttir et al., 2007a: 18/27). The δ¹⁸O_sw was taken to be 0.28‰ (Kristjánsdóttir et al., 2007a; Smith et al., 2005). Temperature estimates based on benthic foraminifera δ¹⁸O are higher than those derived from their Mg/Ca ratios (Figure 6). When paired data are plotted, it is noticeable that there is a divergence close to the middle-/late-Holocene boundary with the Mg/Ca estimates trending more sharply downward than the δ¹⁸O-based estimates implying a decrease in salinity. If the Mg/Ca T°C estimates are substituted into the Shackleton (1974) equation, the equation can be re-arranged to solve for δ¹⁸O_sw. We did this for our most complete series Islandiella norcrossi/helenae. The results indicate an average δ¹⁸O_sw of 0.28‰, and there is a decrease in δ¹⁸O_sw over the last 5000 cal. yr BP implying that the salinity of the bottom waters has decreased. This is illustrated in the divergence of the T°C estimates, especially for Islandiella norcrossi/helenae (Figure 6a).

Discussion: Structure of Holocene ocean climate variability at MD99-2269

Our analysis is next focused on ascertaining the broad structure of changes in Holocene conditions on the North Iceland shelf based on (1) our new ocean climate proxies and (2) other published records (Table 1). However, the area was also affected by other factors whose impacts may be contained in the sediment record. Two specific examples are changes in relative sea level (Quillmann et al., 2010; Rundgren et al., 1997), and possible impact of the Storegga submarine slide (Bondevik et al., 1996; Dawson et al., 2011). The combination of a lower relative sea level (0 to −20 m) and the tsunami generated by the slide could have potentially impacted the sediment record (e.g. the >63-µm grain-size data; Figure 3). However, there is no evidence that deposition at MD99-2269 was disrupted (Kristjánsdóttir, 2005).

We ask, should we expect to see significant levels of correlation between surface or near-surface proxies and those being imprinted at the seafloor? (e.g. Figure 3). We compare the proxies in MD99-2269, based on their location within the water column (surface, near-surface, bottom). Figure 7 shows the estimated Holocene range of temperature estimates on 11 proxies for surface, near-surface, and bottom temperatures. The results show an overall decrease in T°C averaging around 10°C at the sea surface to ~5°C at the seafloor with temperature ranges of 3–7°C (Figure 7). However, this figure also clearly demonstrates that different proxies result in considerably different median temperature estimates and ranges, but it is difficult to explain exactly what are the causes of the differences, and they may be associated with calibration approaches.

Figure 8a plots the standardized data matrix for the 14 ocean climate variables (see section ‘Methods’ and Table 1) between 400 and 9200 cal. yr BP. PCA (Davis, 1986) indicated that the first PC explained 39% of the variance and the second PC added a further 18% explanation (Figure 8b). The first PC is positively loaded with temperature-associated variables (e.g. Mg/Ca of benthic foraminifera, quartz wt%, and SST°C estimates from diatoms), and negatively associated with nutrient associated variables such as coccoliths and calcite wt% (Figure 8b; Supplementary Figure 1, available online). The PC 1 and 2 scores define the basic structure of the 14 proxies and portray a long-term trend (PC 1) and an inverted U-shape (PC 2) pattern (Figure 8c).

In Figure 9, we present an interpretation of the data. The dominant trend is consistent with decreasing summer insolation during the Holocene (Figure 9a) (e.g. Jiang et al., 2015; Thornalley et al., 2013). However, we have to explain the low levels of calcite in the early-Holocene versus the high SST°C estimates (Figure 9b) and the low δ¹⁸O values of benthic foraminifera. We hypothesize that the lag in the timing of the early-/middle-Holocene boundary and the strikingly different trends in SST°C and coccolith records (Figure 9b and c) reflect a long-lived freshwater cap on the Iceland Shelf that must reflect the export of freshwater from the final retreat phase of Northern Hemisphere Ice Sheets, and in particular the Laurentide Ice Sheet (Dyke et al., 2002; Jennings et al., 2015; Renssen et al., 2002). This episode is also marked by high wt% of quartz (Figure 3), which reflects significant advection of drift ice (Moros et al., 2006), probably via Fram Strait (e.g. Fisher et al., 2002), which in turn suggests an increase in freshwater transport. The increase in freshwater during the late-Holocene (Figure 6) may well reflect the increase in the export of sea ice through Fram Strait associated with the flooding of the shallow Arctic shelves (Bauch et al., 2001; Blaschek and Renssen, 2013). The importance of changes in the salinity of surface waters on nutrient availability is well documented for the Iceland Shelf (Stefansson and Olafsson, 1991; Thordardottir, 1984, 1986).

Division of the Holocene

We used the performance indicators calculated in FuzMe (Minasny and McBratney, 2002) to determine that three discrete clusters could be identified in the 14 variable matrix. We tested the robustness of the threefold cluster assignments using discriminant function analysis (Davis, 1986). The null hypothesis was rejected at the p > 0.0001 level and only one sample was misclassified. A plot of the cluster membership versus age clearly demarcated early-, middle-, and late-Holocene intervals with boundaries of ~6350 and ~3450 cal. yr BP (Figure 9a and d – see also Figure 3). The Confusion/Mixing Index reaches maximum values during the transitions. The early- to middle-Holocene transition has two strong precursors prior to the abrupt transition ca 6350 cal. yr BP, whereas the middle- to late-Holocene transition is abrupt, with the suggestion toward an initial move to cooler conditions ca 4000 cal. yr BP (Figure 9d). In a high-resolution biomarker study at site MD99-2266, to the west of our site (Figure 2), Moossen et al. (2015) identified major ocean climate boundaries at 7.8 and 3.2 cal. ka BP, whereas Olafsdottir et al. (2010) noted two regime shifts at 4 and 8 cal. ka BP in an adjacent site, MD99-2264 (Figure 2). An evaluation of climate proxy records from Icelandic lake sediments (Geirsdottir et al., 2013) concluded that peak summer warmth occurred by 7.9 cal. ka BP and that evidence for cooler conditions was evident by 5.5 cal. ka BP, with all proxies recording ‘cold perturbation’ 4.3–4.0 cal. ka BP (p. 48). Our early/middle-Holocene boundary occurs 1400 years later than noted in these marine and lake sites, and ~1600 years later than the suggestions of Walker et al. (2012), whereas the middle-/late-Holocene boundary is similar in age. Not surprisingly, this comparison reveals that regional climate conditions can lead or lag globally defined boundaries, especially in areas near climate frontal boundaries. It also suggests that the North Iceland shelf (MD99-2269) has a more complex Holocene history than sites in NW Iceland, reflecting its more complex oceanographic setting.

Conclusion

Our paper uses proxy ocean climate records from a single core, located in a climatically highly sensitive area of the northern North Atlantic (Figures 1 and 2), to ascertain whether they portray a common signal:

The common signal at MD99-2269 is for both surface and BWT°Cs to decrease throughout the Holocene.

Nutrient availability and surface productivity was highest in the middle-Holocene, here defined as the interval between 6350 and ~3450 cal. yr BP.

Many proxies had a significant trend in their values that explained a large fraction (>40%) of the variance in the signal – this trend is associated with decreasing Holocene summer insolation in the Northern Hemisphere.

A variety of multi-century to millennial periodicities can be detected in the detrended data at the ⩾95% confidence level, but there is no obvious common periodicity other than the ~430-year interval.

The early-Holocene is marked by contrasting proxy signals that we propose are the result of the impact of a freshwater cap on both SST°C and on nutrient depletion.

The increase in quartz and the decrease in salinity in the late-Holocene probably reflect an increase in the advection of sea ice from the Arctic Ocean to northern Iceland.

Footnotes

Acknowledgements

We thank the authors of the previous papers dealing with data from MD99-2269 for their considerable contributions. We appreciate the comments and suggestions of Professor Simon Belt on an earlier draft of this paper, and the comments of two reviewers and the Editor, which resulted in substantial changes in organization of the manuscript. The paper is a contribution to the Abrupt North Atlantic Transitions and Ice, Lakes, and Sea and Arctic Holocene Transition Projects.

Funding

Cruise B997 was supported by NSF-ATM-9531397 and in large part by Icelandic resources. NSF OCE-9809001 and subsequent research by NSF OPP-0004233 supported our participation in the IMAGES IV cruise. Kristjánsdóttir research was supported by an NSF-ESH grant ATM-0317832 to JTA and AEJ and along with student grants to, from RANNIS, Fulbright, and AAUW.

References

Alley

Agustsdottir

(2005) The 8k event: Cause and consequences of major Holocene abrupt climate change. Quaternary Science Reviews 24: 1123–1149.

Andersen

Koc

Jennings

. (2004) Nonuniform response of the major surface currents in the Nordic Seas to insolation forcing: Implications for the Holocene climate variability. Paleoceanography 19: 1–16.

Andrews

(2007) Holocene denudation of Iceland as determined from accumulation of sediments on the continental margin. Boreas 36: 240–252.

Andrews

Geirsdottir

Hardardottir

. (2002) Distribution sediment magnetism and geochemistry of the Saksunarvatn (10180 60 ± cal yr BP) tephra in marine lake and terrestrial sediments NW Iceland. Journal of Quaternary Science 17: 731–745.

Andrews

Hardardottir

Stoner

. (2003) Decadal to millennial-scale periodicities in North Iceland shelf sediments over the last 12000 cal yrs: Long-term North Atlantic oceanographic variability and solar forcing. Earth and Planetary Science Letters 210: 453–465.

Axford

Andresen

Andrews

. (2011) Do paleoclimate proxies agree? Statistical comparison of climate and sea-ice reconstructions from Icelandic marine and lake sediments 300–1900 AD. Journal of Quaternary Science 26: 645–656.

Bauch

Mueller-Lupp

Taldenkova

. (2001) Chronology of the Holocene transgression at the North Siberian margin. Global and Planetary Change 31: 125–139.

Belkin

Cornillon

Sherman

(2009) Fronts in large marine ecosystems. Progress in Oceanography 81: 223–236.

Belkin

Levitus

Antonov

. (1998) ‘Great salinity anomalies’ in the North Atlantic. Progress in Oceanography 41: 1–68.

10.

Bendle

JAP

Rosell-Melé

(2004) Distributions of Uk37 and Uk’37 in the surface waters and sediments of the Nordic Seas: Implications for paleoceanography. Geochemistry, Geophysics, Geosystems 5: Q11013.

11.

Bendle

JAP

Rosell-Mele

(2007) High-resolution alkenone sea surface temperature variability on the North Icelandic Shelf: Implications for Nordic Seas paleoclimatic development during the Holocene. The Holocene 17: 9–24.

12.

Bendle

JAP

Rosell-Melé

Ziveri

(2005) Variability of unusual distribution of alkenones in the surface waters of the Nordic seas. Paleoceanography 20: PA2001.

13.

Berger

Loutre

(1991) Insolation values for the climate of the last 10 million years. Quaternary Science Reviews 10: 297–318.

14.

Blaauw

(2012) Out of tune: The dangers of aligning proxy archives. Quaternary Science Reviews 36: 38–49.

15.

Blanz

Emeis

Siegel

(2005) Controls on alkenone unsaturation ratios along the salinity gradient between the open ocean and the Baltic Sea. Geochimica et Cosmochimica Acta 69: 3589–3600.

16.

Blaschek

Renssen

(2013) The impact of early Holocene Arctic shelf flooding on climate in an atmosphere-ocean-sea-ice model. Climate of the past 9: 2651–2667.

17.

Bondevik

Svendsen

Johnsen

. (1996) The Storegga tsunami along the Norwegian coast, its age and runup. Boreas 26: 29–53.

18.

Castaneda

Smith

Kristjánsdóttir

. (2004) Temporal changes in Holocene del18O records from the northwest and central North Iceland Shelf. Journal of Quaternary Science 19: 1–14.

19.

Clemens

(2005) Millennial-band climate spectrum resolved and linked to centennial-scale solar cycles. Quaternary Science Reviews 24: 521–531.

20.

Darby

Myers

Herman

. (2015) Chemical fingerprinting, a precise and efficient method to determine sediment sources. Journal of Sedimentary Research 85: 247–253.

21.

Davis

(1986) Statistics and Data Analysis in Geology. New York: John Wiley & Sons, 646 pp.

22.

Dawson

Bondevik

Teller

(2011) Relative timing of the Storegga submarine slide, methane release, and climate change during the 8.2 ka cold event. The Holocene 21: 1167–1171.

23.

Dickson

Meincke

Malmberg

. (1988) The ‘great salinity anomaly’ in the Northern North Atlantic 1968–1982. Progress in Oceanography 20: 103–151.

24.

Ding

(2004) Cluster structure of K-means clustering via principal component analysis. In: Dai

Srikant

Zhang

(eds) Advances in Knowledge Discovery and Data Mining: Proceedings of the 13th Pacific-Asia Conference, PAKDD 2009, Bangkok, Thailand, April 27-30, 2009. Berlin: Springer, pp. 414–418.

25.

Divine

Dick

(2006) Historical variability of the sea ice edge position in the Nordic Seas. Journal of Geophysical Research. DOI: 10.1029/2004JC002851.

26.

Dunhill

Andrews

Kristjánsdóttir

(2004) Radiocarbon Date List X: Baffin Bay Baffin Island Iceland Labrador and the Northern North Atlantic. Occasional Paper no 56. Boulder, CO: Institute of Arctic and Alpine Research, University of Colorado, 77 pp.

27.

Dyke

Andrews

Clark

. (2002) The Laurentide and Innuitian ice sheets during the Last Glacial Maximum. Quaternary Science Reviews 2: 9–31.

28.

Eiriksson

Larsen

Knudsen

K-L

. (2004) Marine reservoir age variability and water mass distribution in the Iceland Sea. Quaternary Science Reviews 23: 2247–2268.

29.

Elderfield

Ganssen

(2000) Past temperature and d18O of surface ocean waters inferred from foraminiferal Mg/Ca ratios. Nature 405: 442–445.

30.

Fairbanks

(1989) A 17000-year glacio-eustatic sea level record: Influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature 342: 637–642.

31.

Fisher

Smith

Andrews

(2002) Preboreal oscillation caused by a glacial Lake Agassiz flood. Quaternary Science Reviews 21: 873–878.

32.

Geirsdottir

Miller

Larsen

. (2013) Abrupt Holocene climate transitions in the northern North Atlantic region recorded by synchronized lacustrine records in Iceland. Quaternary Science Reviews 70: 48–62.

33.

Ghil

Allen

Dettinger

. (2002) Advanced spectral methods for climatic time series. Reviews of Geophysics 40: 3–41.

34.

Ghosh

Dubey

(2013) Comparative analysis of K-means and Fuzzy C-means algorithms. International Journal of Advanced Computer Science and Technology 4: 35–39.

35.

Giraudeau

Jennings

Andrews

(2004) Timing and mechanisms of surface and intermediate water circulation changes in the Nordic Sea over the last 10000 cal years: A view from the North Iceland shelf. Quaternary Science Reviews 23: 2127–2139.

36.

Giraudeau

Solignac

Moros

. (2010) Millennial-scale variability in Atlantic water advection to the Nordic Seas derived from coccolith concentration records. Quaternary Science Reviews 29: 1276–1287.

37.

Granath

(1984) Application of fuzzy clustering and fuzzy classification to evaluate the provenance of glacial till. Mathematical Geology 16: 283–303.

38.

Harada

Shin

Murata

. (2003) Characteristics of alkenones synthesized by a bloom of Emiliania huxleyi in the Bering Sea. Geochimica et Cosmochimica Acta 67: 1507–1519.

39.

Helgadottir

(1997) Paleoclimate (0 to >14 ka) of W and NW Iceland: An Iceland/USA Contribution to PALE. Cruise Report B9-97, July. Reykjavik: Marine Research Institute of Iceland.

40.

Jennings

Andrews

Wilson

(2015) Detrital carbonate events on the Labrador shelf, a 13 to 7 kyr template for freshwater forcing from the Laurentide Ice Sheet. Quaternary Science Reviews 107: 62–80.

41.

Jennings

Weiner

Helgadottir

. (2004) Modern foraminiferal faunas of the Southwest to Northern Iceland shelf: Oceanographic and environmental controls. Journal of Foraminiferal Research 34: 180–207.

42.

Jiang

Muscheler

Björck

. (2015) Solar forcing of Holocene summer sea-surface temperatures in the northern North Atlantic. Geology 43: 203–206.

43.

Justwan

Koc

(2008) A diatom based transfer function for reconstructing sea ice concentrations in the North Atlantic. Marine Micropaleontology 66: 264–278.

44.

Knudsen

Eirikson

Jansen

. (2004) Paleoceanographic changes off North Iceland through the last 1200 yrs: Foraminifera stable isotopes diatomes and ice rafted debris. Quaternary Science Reviews 23: 2231–2246.

45.

Koch

(1945) The East Greenland Ice. Meddelelser om Gronland 130(3): 346.

46.

Kondrashov

Feliks

Ghil

(2005) Oscillatory modes of extended Nile River records (AD 622–1922). Geophysical Research Letters. Epub ahead of print 24 May. DOI: 10.1029/2004GL022156.

47.

Kristjánsdóttir

(2005) Holocene changes in climate environment and ocean reservoir age on the Iceland shelf: Mg/Ca d18O and tephrachronology of core MD99-2269. PhD Dissertation, Department of Geological Sciences, University of Colorado, 423 pp.

48.

Kristjánsdóttir

Lea

Jennings

. (2007a) New spatial Mg/Ca-Temperature Calibrations for three Arctic, benthic foraminifera and reconstruction of North Iceland shelf temperature for the past 4000 years. Geochemistry, Geophysics, Geosystems 8: QO3P21.

49.

Kristjánsdóttir

Stoner

Gronvold

. (2007b) Geochemistry of Holocene cryptotephras from the North Iceland Shelf (MD99-2269): Intercalibration with radiocarbon and paleomagnetic chronostratigraphies. The Holocene 17: 155–176.

50.

Kristjánsdóttir

. (in preparation) Holocene Mg/Ca, alkenones, and light stable isotope measurements on the outer North Iceland shelf, sediment core MD99-2269. DOI: 10.1594/PANGAEA.847657, Supplement to: Kristjánsdóttir, Greta B; Moros, Matthias; Andrews, John T; Jennings, Anne E: Holocene Mg/Ca, alkenones, and light stable isotope measurements on the outer North Iceland shelf (MD99-2269): A comparison with other multi-proxy data and sub-division of the Holocene. The Holocene.

51.

Labeyrie

Jansen

Cortijo

(2003) Les rapports de campagnes a la mer MD114/IMAGES V. Brest: Institut Polaire Francais Paul-Emile Victor.

52.

Nychka

Ammann

(2010) The value of multiproxy reconstruction of past climate. Journal of the American Statistical Association 105: 883–895.

53.

McClymont

Rosell-Melé

Haug

. (2008) Expansion of subarctic water masses in the North Atlantic and Pacific oceans and implications for mid-Pleistocene ice sheet growth. Paleoceanography 23: PA4214.

54.

Malmberg

S-A

(1969) Hydrographic changes in the waters between Iceland and Jan Mayen in the last decade. Jokull Journal 19: 30–43.

55.

Malmberg

S-A

(1985) The water masses between Iceland and Greenland. Journal of the Marine Research Institute 9: 127–140.

56.

Mann

Lees

(1996) Robust estimation of background noise and signal detection climatic time series. Climatic Change 33: 409–445.

57.

Masse

Rowland

Sicre

M-A

. (2008) Abrupt climate changes for Iceland during the last millennium: Evidence from high resolution sea ice reconstructions. Earth and Planetary Science Letters 269: 565–569.

58.

Meland

Jansen

Elderfield

. (2006) Mg/Ca ratios in the planktonic foraminifer Neogloboquadrina pachyderma (sinistral) in the northern North Atlantic/Nordic Seas. Geochemistry, Geophysics, Geosystems. DOI: 10.1029/2005gc001078.

59.

Minasny

McBratney

(2002) FuzMe Version 3.0. Eveleigh: Australian Center for Precision Agriculture, The University of Sydney.

60.

Moossen

Bendle

JAP

Seki

. (2015) North Atlantic Holocene climate evolution recorded by high-resolution terrestrial and marine biomarker records. Quaternary Science Reviews 129: 111–127.

61.

Moros

Andrews

Eberl

. (2006) The Holocene history of drift ice in the northern North Atlantic: Evidence for different spatial and temporal modes. Paleoceanography. DOI: 10.1029/2005PA001214.

62.

Moros

Emeis

Risebrobakken

. (2004) Sea surface temperatures and ice rafting in the Holocene North Atlantic: Climate influences on northern European and Greenland. Quaternary Science Reviews 23: 2113–2126.

63.

Müller

Kirst

Ruhland

. (1998) Calibration of alkenone paleotemperature index UK’37 based on core tops from the eastern South Atlantic and the global ocean (60°N–60°S). Geochimica et Cosmochimica Acta 62: 1757–1772.

64.

Ogilvie

AEJ

(1996) Sea-ice conditions off the coasts of Iceland AD 1601-1850 with special reference to part of the Maunder Minimum period (1675–1715). In: Pedersen

(ed.) North European Climate Data in the Latter Part of the Maunder Minimum Period AD 1675–1715. Stavanger: Archaeological Museum of Stavanger, pp. 9–12.

65.

Ogilvie

AEJ

Barlow

Jennings

(2000) North Atlantic climate c AD 1000: Millennial reflections on the Viking discoveries of Iceland Greenland and North America. Weather 55: 34–45.

66.

Olafsdottir

Jennings

Geirsdottir

. (2010) Holocene variability of the North Atlantic Irminger current on the south- and northwest shelf of Iceland. Marine Micropaleontology 77: 101–118.

67.

Olafsson

(1999) Connections between oceanic conditions off N-Iceland, Lake Myvatn temperature, regional wind direction variability and the North Atlantic Oscillation. Rit Fiskideildar 16: 41–57.

68.

Paillard

Labeyrie

Yiou

(1996) Macintosh program performs time-series analysis. EOS Transactions 77: 379.

69.

Quillmann

Jennings

Andrews

(2010) Reconstructing Holocene paleoclimate and paleoceanography in Isafjaradrdjup, Northwest Iceland, from two fjord cores overprinted by relative sea level and local hydrographic changes. Journal of Quaternary Science 25: 1144–1159.

70.

Quillmann

Marchitto

Jennings

. (2012) Cooling and freshening at 82 ka on the NW Iceland shelf recorded in paired d18O and Mg/Ca measurements of the benthic foraminifera Cibicides lobatulus. Quaternary Research 78: 528–539.

71.

Renssen

Goose

Fichefet

(2002) Modeling the effect of freshwater pulses on the early Holocene climate: The influence of high-frequency climate variability. Paleoceanography 17: 1–18.

72.

Risebrobakken

Moros

Ivanova

. (2010) Climate and oceanographic variability in the SW Barents Sea during the Holocene. The Holocene 20: 609–621.

73.

Rohde Krossa

Moros

Blanz

. (2014) Late-Holocene Baltic Sea outflow changes reconstructed using C37:4 content from marine cores. Boreas 44: 81–93.

74.

Rosell-Melé

(1998) Interhemispheric appraisal of the value of alkenone indices as temperature and salinity proxies in high-latitude locations. Paleoceanography 13: 694–703.

75.

Rosell-Melé

Jansen

Weinelt

(2002) Appraisal of a molecular approach to infer variations in surface ocean freshwater inputs into the North Atlantic during the last glacial. Global and Planetary Change 34: 143–152.

76.

Rundgren

Ingolfsson

Bjorck

. (1997) Dynamic sea-level change during the last deglaciation of northern Iceland. Boreas 26: 201–215.

77.

Rytter

Knudsen

K-L

Seidenkrantz

M-S

. (2002) Modern distribution on benthic foraminifera on the North Icelandic shelf and slope. Journal of Foraminiferal Research 32: 217–244.

78.

Schulz

H-M

Schöner

Emeis

K-C

(2000) Long-chain alkenone patterns in the Baltic Sea – An ocean-freshwater transition. Geochimica et Cosmochimica Acta 64: 469–477.

79.

Shackleton

(1974) Attainment of isotopic equilibrium between ocean water and the benthic foraminifera genus Uvigerina: Isotopic changes in the ocean during the last glacial. CNRS Colloques Internationaux 219: 203–209.

80.

Sicre

M-A

Bard

Ezat

. (2002) Alkenone distributions in the North Atlantic and Nordic sea surface waters. Geochemistry, Geophysics, Geosystems 3: 1–13.

81.

Sicre

M-A

Jacob

Ezat

. (2008) Decadal variability of sea surface temperatures off North Iceland over the last 2000 years. Earth and Planetary Science Letters 268: 137–142.

82.

Sikes

Sicre

(2002) Relationship of the tetra-unsaturated C-37 alkenone to salinity and temperature: Implications for paleoproxy applications. Geochemistry Geophysics Geosystems 3: 1–11.

83.

Smith

Licht

(2000) Radiocarbon Date List IX: Antarctica Arctic Ocean and the Northern North Atlantic. INSTAAR Occasional Paper no 54. Boulder, CO: University of Colorado, 138 pp.

84.

Smith

Andrews

Castaneda

. (2005) Temperature reconstructions for SW and N Iceland waters over the last 10000 cal yr BP based on d18O records from planktic and benthic Foraminifera. Quaternary Science Reviews 24: 1723–1740.

85.

Solignac

Giraudeau

de Vernal

(2006) Holocene sea surface conditions in the western North Atlantic: Spatial and temporal heterogeneities. Paleoceanography 21: 1–16.

86.

Stefansson

(1962) North Icelandic waters. Rit Fiskideildar 3: 269.

87.

Stefansson

(1969) Temperature variations in the North Icelandic coastal area during recent decades. Jokull Journal 19: 18–28.

88.

Stefansson

Olafsson

(1991) Nutrients and fertility of Icelandic waters. Rit Fiskideildar XII: 1–56.

89.

Stoner

Jennings

Kristjansdottir

. (2007) A paleomagnetic approach toward refining Holocene radiocarbon based chronostratigraphies: Paleoceanographic records from North Iceland (MD99-2269) and East Greenland (MD99-2322) margins. Paleoceanography 22: PA1209.

90.

Thordardottir

(1984) Primary Production North of Iceland in Relation to Water Masses in May-June 1970–1980. Copenhagen: Council for the Exploration of the Sea CM 1984/L20, pp. 1–17.

91.

Thordardottir

(1986) Timing and Duration of Spring Blooming South and Southwest of Iceland (NATO ASI Series G-7). Berlin: Springer, pp. 345–360.

92.

Thornalley

DJR

Blaschek

Davies

. (2013) Long-term variations in Iceland-Scotland overflow strength during the Holocene. Climate of the past 9: 2073–2084.

93.

Torbenson

MCA

Plunkett

Brown

. (2015) Asynchrony in key Holocene chronologies: Evidence from Irish bog pines. Geology 43: 799–802.

94.

Turney

CSM

Baillie

Clemens

. (2005) Testing of solar forcing of pervasive Holocene climate cycles. Journal of Quaternary Science 20: 511–518.

95.

Velleman

Hoaglin

(1981) Applications, Basics, and Computing of Exploratory Data Analysis. Boston, MA: Duxbury, 354 pp.

96.

Walker

MJC

Berkelhammer

Bjorck

. (2012) Formal subdivision of the Holocene Series/Epoch: A discussion paper by a working group of INTIMATE (Integration of ice-core marine and terrestrial records) and the Subcommission on Quaternary Stratigraphy (International Commission on Stratigraphy). Journal of Quaternary Science 27: 649–659.

97.

Wallevik

Sigurjonsson

(1998) The Koch Index: Formulation, Corrections and Extension. Reykjavik: Icelandic Meteorological Office Report.

98.

Weedon

(2005) Time-Series Analysis and Cyclostratigraphy: Examining Stratigraphic Records of Environmental Cycles. Cambridge: Cambridge University Press, 259 pp.