Diatom (Bacillariophyceae) flora of early Holocene freshwater sediments from Skalafjord, Faeroe Islands

Relative abundance data of diatom (Bacillariophyceae) species were generated for sediment core SKPC-01B from the Skalafjord, Faeroe Islands. The record shows distinct temporal changes in species composition. In the lowermost 65 cm of the 230 cm long core a species-rich freshwater diatom assemblage was found. Most of the taxa observed in this section are typical of oligotrophic to dystrophic lakes in northern Europe (Scandinavia, Iceland and Spitsbergen). Above this interval the diatom flora is dominated by marine taxa. The change from a freshwater to a marine flora is inferred to be caused by rising sea-level that took place about 7700–6400 years BP. Drastic changes in the diatom species composition within the transitional core section show that environmental change in the Skalafjord took place in several pulses. The first stage included strong inflow (possibly catastrophic) of marine waters. As a possible trigger of this phenomenon the tsunami released by the Storegga Slide is proposed. Before the final flooding by marine waters, freshwater conditions were re-established within the Skalafjord. These results have important implications for the interpretation of the palaeogeographical development of the Eysturoy area. Hence, it is suggested that the Storegga Slide led to inflow of marine waters at a distinctly lower water level in the area of the Skalafjord than proposed in recent publications and that the inundation of the threshold in the fjord happened after the tsunami.


INTRODUCTION
Fossil diatom floras of freshwater and marine origin may be used for reconstructing environmental changes. Freshwater diatom floras from limnic sediments are useful for reconstructions of the climate changes that took place at high latitudes in the Northern Hemisphere since the last deglaciation. Recently, studies of lacustrine sediments from the land areas surrounding the North Sea and Norwegian Sea have been shown to contain a record of past catastrophic events that took place in the area. One such record is a tsunami caused by the Storegga Slide dated at c. 7500 14 C years BP (Dawson & Smith, 2000). Diatoms represent one of the best indicators of the impact of this tsunami on the sedimentary record. This phenomenon has been shown to occur in sediment cores from lakes from the Faeroe Islands (Grauert et al., 2001). Abrupt changes in diatom species composition were interpreted as indicators of this catastrophic event. However, the altitude of the sediment section analysed here and its significance for the ocean level at which the Storegga Slide took part is not in agreement with palaeogeographical interpretations given by Bennike et al. (1998) and Grauert et al. (2001).
Although the first publications on freshwater diatoms from the northern part of the North Atlantic are from the nineteenth and twentieth centuries, knowledge of the freshwater diatom flora of the Faeroe Islands is rather poor. Early publications (Lyngbye, 1819;Cleve, 1873Cleve, , 1896Cleve, , 1898Cleve, , 1900Lagerstedt, 1873;Cleve & Grunow, 1880;Østrup, 1897;Brun, 1901) dealt with the high latitude North Atlantic in general and usually concerned both marine and freshwater floras. Later, Hustedt (1937), Krasske (1938) and Foged (1964Foged ( , 1974 published results on their studies of the freshwater diatom flora from some North Atlantic islands (for example, the Faeroe Islands and Spitsbergen). Only Hustedt (1937) dealt with diatoms from Iceland, the Faeroe Islands and Spitsbergen. The first report focusing on freshwater diatoms from the Faeroe Islands was by Lyngbye (1819). The next study specifically dealing with the freshwater diatom flora from the Faeroe Islands was published by Østrup (1901). Somewhat later, Østrup (1903) published a report on marine diatoms from this area. Since then, no papers on freshwater diatoms from the Faeroes have been published, to the best of our knowledge.
Recently, an effort was directed towards studies of the marine diatom flora of the North Atlantic and the results were used for studies of climate change following the last deglaciation. The major objective of these studies was to decipher palaeoceanographical changes (Koc & Schrader, 1990;Koc & Jansen, 1992;Schrader et al., 1993a, b;Kohly, 1998;Wachnicka, 1999;Jozkow, 2000;Jiang et al., 2001;Witak et al., 2004).
Core SKPC-01B from the Skalafjord, Faeroe Islands has been analysed for diatoms. The Skalafjord penetrates into Eysturoy, the biggest of the Faeroe Islands (Fig. 1). Diatoms are well preserved and dominated by freshwater forms in the lowermost part and marine forms in the upper part. The focus is on the freshwater diatom flora from the lower part of the core. The sediments are of early Holocene age, and the flora is typical of high latitude nutrient-poor (oligotrophic to dystrophic) lakes (e.g. Cleve-Euler, 1951-1955. The position of the freshwater deposits within the section and the weak representation of marine elements suggest that deposition took place before the threshold in the Skalafjord was inundated. the North Atlantic basalt province, and the whole area was influenced by Tertiary volcanism (Boldreel & Andersen, 1995;van Weering et al., 1998). During glacial periods, part of the North Faeroe shelf was exposed subaerially and may have been glaciated (Jørgensen & Rasmussen, 1986). Lakes and bogs are common throughout the islands, and the deposits within them provide Holocene palaeoenvironmental records (Grauert et al., 2001). Skalafjord is 13 km long with a greatest depth of around 70 m. At the entrance of the fjord a sill with a water depth of 25 m is present. The fjord is surrounded by 500-600 m high mountains. Post-glacial sediments in the fjord were deposited in two separate basins and have a maximum thickness of about 20 m (Juul, 1992).

MATERIAL AND METHODS
Core SKPC-01B was one of nine sediment cores retrieved from the Faeroe Islands during the September-October 1995 R/V Skagerak cruise organized by Göteborg University in collaboration with the Geological Survey of Denmark and Greenland (Fig. 1). The coring site was situated in the central part of the Skalafjord (62(10#70$N and 6(47#87$W) in a water depth of 50 m.
Samples were prepared in the manner of Håkansson & Ross (1984). Samples for diatom analyses were collected at 5 cm intervals. One gram of sediment was dried at 60( for 24 hours. The sediment was treated with 10% HCl to dissolve carbonates and then washed several times with distilled water. The siliceous material was gently boiled in concentrated (37%) H 2 O 2 and washed several times with distilled water. The supernatant was decanted off after 20 hours. An aliquot of the shaken suspension was transferred by pipette to an 18 18 mm square coverslip. The coverslips were left to dry at room temperature. After evaporation, the coverslips were placed onto labelled slides. Permanent diatom preparations were mounted with Naphrax (refractive index=1.78) and briefly heated to 200(C. Diatom analyses were performed with a LEICA DMLB light microscope, using 100/1.25 planapochromatic oil-immersion objective. Scanning electron microscope analysis was performed by means of a Zeiss DSM 940 at 25 kV. In each sample more than 300 valves were counted. Diatoms were counted by the Schrader & Gersonde (1978) method.

Abundance and concentration
The core length studied is 230 cm, and it is characterized by predominantly homogeneous olive-grey clayey mud (Fig. 2). More or less corroded shell fragments were observed along the whole profile. Their quantity distinctly increased at 120-130 cm depth. At 108-118 cm the sediment was distinctly laminated and somewhat darker.
Within the whole sediment profile the diatoms represent two completely different environments. At a depth of 230-165 cm, taxa typical of limnic environments predominated (Fig. 3). Above 165 cm the flora is almost exclusively marine. The sediments representing these two different environments are connected by an apparently transitional section between 180-165 cm. In this part of the sediment profile a transition from limnic to marine conditions is recorded. First, in the section from 180-170 cm, a strong peak in marine diatoms occurs followed by a dominance of freshwater taxa in the sediment interval from 170-165 cm. These abrupt environmental changes took place during the period 7700-6400 years BP.
A total of 166 diatom taxa have been identified. In general, the preservation state was satisfactory but, at some levels, the valves were fragmented. Freshwater diatoms were represented by 121 taxa, brackish-water forms by 16 taxa and marine forms by 28 taxa. The freshwater flora was dominated by benthic species (126 species), while the planktonic flora consisted of 39 species.
In this paper the freshwater diatoms that occurred in the lowermost part of the core are described. Two diatom assemblage zones (DAZ) and several subzones are distinguished (Fig. 3). The first zone (DAZ-1) corresponds to the lower part of the core (230-165 cm). Two subzones were distinguished,  and   (Fig. 4). The following criteria were applied to distinguish the diatom assemblage zones: + changes in the ratio between marine and freshwater taxa; + habitat characteristics, i.e. planktonic versus benthic forms; + diatom concentration in number of valves per 1 g of sediment.
DAZ-1a. The age of the boundary between subzone DAZ-1a and DAZ-1b sediments was estimated to be about 7700 14 C years BP. Diatom zone DAZ-1 is characterized by abundant freshwater taxa and less abundant marine ones (Fig. 3). Diatom valves are usually very well preserved.

DAZ-1b.
Diatom assemblage subzone DAZ-1b encompasses the sediment interval 195-165 cm. The age of the transition from subzone DAZ-1b to DAZ-2 was estimated to be about 6400 years BP. The bottom of the zone is marked by a drastic decrease in diatom valve concentration to c. 50 10 5 valves/g (Fig. 3). Freshwater taxa dominate, but in the interval 180-170 cm a peak of marine taxa is recorded. Amongst the predominant species are: Paralia sulcata (up to 30%), Thalassiosira nordenskioeldii (up to 17%), Thalassiosira hyalina (up to 9%) and Odontella aurita (up to 5%). Among the freshwater forms the following taxa were the most abundant: C. rossii (up to 53%), A. distans (up to 15%), A. subarctica (up. to 10%), Fragilaria ulna (up to 10%) and T. flocculosa (up to 10%). With respect to habitat, planktonic forms dominate with 50-80% ( Fig. 4). Only in the section rich in marine forms is a peak in benthic forms (up to 40%) seen.

DISCUSSION
The development of the Skalafjord since the last deglaciation was studied by Bennike et al. (1998;macrofossils in core SKPC-18), Jozkow (2000; diatoms in core SKPC-18) and Wachnicka (1999; diatoms in core DAPC-01). The diatom record in core SKPC-01B provides excellent documentation of the Holocene development of the Faeroe Islands area -the best record of Early Holocene changes known so far. Well-preserved lacustrine deposits at a similar altitude within the Skalafjord were also studied by Jozkow (2000). However, the upper part of core SKPC-18 studied by Jozkow (2000) apparently did not contain the complete record of the diatom flora.
The sediments of core SKPC-01B show two distinct developmental stages. The first one recorded in the lowermost part of the core, subzone DAZ-1a, encompasses lacustrine sediments with a very low content of marine-and brackish-water diatoms. As the proportion between fully marine taxa and brackish-water ones is similar, it is assumed that no permanent connection existed between the central part of the fjord and the ocean. Apparently, diatoms of a marine origin were transported into the Skalafjord during storms. It appears that relative sea-level was below the threshold.
The species composition recorded in subzone DAZ-1b indicates the existence of a lake with abundant planktonic diatoms dominated by C. rossii and F. ulna (Fig. 4). Their vertical distributions do not show any dramatic changes, implying rather stable conditions. Sporadic inflow of marine waters did not cause any spectacular changes during this developmental stage. Environmental conditions of diatoms in subzone DAZ-1a indicate oligotrophic to mesotrophic waters (Krammer & Lange-Bertalot, 1986, 1997, 1991a, b, 2000Håkansson, 1990;Denys, 1992;Hoffman, 1994;Van Dam et al., 1994). It is likely that the dominating taxon, C. rossii, formed blooms. Most of the taxa recorded in DAZ-1a are typical for oligotrophic to dystrophic waters (e.g.  accompanied by, for example, A. distans, F. exigua and the T. flocculosa complex (Fig. 4). Generally, within DAZ-1a, a continuous distribution of freshwater taxa is observed. Several of these taxa have been recently described (e.g. Lange-Bertalot & Krammer, 2000Krammer, , 2002 or are known only from very few localities. e.g. Fragilaria opacolineata Lange-Bertalot, Stauroneis neohyalina Lange-Bertalot, Gomphonema subtile Ehrenberg, Pinnularia ovata Krammer, Pinnularia platycephala Krammer and Pinnularia turbulenta Krammer. The Skalafjord lake formed after the last deglaciation of the area at c. 10 000 years BP (core SKPC-18; Bennike et al., 1998). The diatom record of the early stages of lake development is recorded by Jozkow (2000). The chronology of core SKPC-18 was established from 14 C dating and tephra chronology based on the Saksunarvatn ash (Bennike et al., 1998). Diatom analysis of this core revealed a species composition very similar to that in core SKPC-01B, with C. rossii as the dominant species. As core SKPC-01B did not penetrate the Saksunarvatn ash and the uppermost part of the former core is disturbed, these two cores complement each other. Core SKPC-18 provides a record of the early stages of lake development, while core SKPC-01B records the later stage including the transition from lacustrine to marine conditions. Bennike et al. (1998) determined the inundation of the threshold and the change from lacustrine to marine conditions in core SKPC-18 at c. 7800 14 C years BP. They also determined the relative sea-level which at that time was c. 25 m below the present sea-level. However, in core SKPC-01B, the transition between lacustrine and marine conditions is dated between 7700 years BP and 6400 years BP. Previously, studies of the benthic foraminiferal fauna of the fjord (Juul, 1992) indicated that the marine transgression occurred around 7500 14 C years BP. In addition the diatom species composition prior to the change from lacustrine to marine conditions shows a rise in marine diatoms at 185-180 cm. In the overlying section (180-165 cm) lacustrine taxa with C. rossii dominate again, with a distinct decrease in marine forms.
The effect of a tsunami, triggered by the Storegga Slide, has been documented in Norway and Britain (e.g. Dawson et al., 1988;Bondevik et al., 1997). Deposits resulting from this event were recognized by Grauert et al. (2001) in Lake Vagur on Suduroy Island, which is located south of Skalafjord. The lithology (redeposited organic material and marine microfossils) marks the tsunami section. The age of the tsunami event was estimated at c. 7200 years BP. In core SKPC-01B the rapid increase in marine taxa between 180 cm and 170 cm (Fig. 3) may signal an inflow of marine waters. As there is no simultaneous change in lithology it appears that this event did not significantly affect sedimentation processes in the lake. Therefore, this event may instead have been caused by a storm surge. The change in diatom flora at 180-170 cm in core SKPC-01B may be evidence of one of the first large-scale inflows of marine waters prior to the inundation of the threshold. It may be assumed that the relative sea-level at that time was a few metres lower than 25 m, as limnic conditions were re-established in the basin after termination of the marine inflow. The beginning of environmental change from lacustrine conditions, which resulted in the establishment of the marine environment, is dated between 7000 14 C years BP and 6400 14 C years BP, implying that environmental change took place rather rapidly.
Diatom analysis of subzone DAZ-1b shows that the lake in Skalafjord was affected by a strong inflow (possibly catastrophic) of marine waters. This phenomenon happened somewhat later than 7700 14 C years BP. However, towards the top of this subzone lacustrine conditions were re-established. The development of the lake in the Skalafjord implies that the Storegga Slide led to inflow of marine waters at a distinctly lower water level in the area of Eysturoy and that the inundation of the threshold in the fjord happened after the tsunami. The lithology of this part of the core indicates that the tsunami impact in this area was relatively weak.

ACKNOWLEDGEMENTS
We wish to express our gratitude to Björn Malmgren, University of Göteborg, Ole Bennike and Antoon Kuijpers, Geological Survey of Denmark and Greenland, Copenhagen, for their critical comments, support and improvement of the text. Special thanks are due to Ditmar Metzeltin and Horst Lange-Bertalot for providing detailed comments on the diatom taxonomy.